PROJECT2_PART2_322691718_206684011.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "fAVvQ6i6NDBZ"
      },
      "source": [
        "# Pyspark Installation And Imports"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "F_HoQHVHG9eB",
        "outputId": "9592e142-7835-4a86-cf26-08f305071fce"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/\n",
            "Collecting findspark\n",
            "  Downloading findspark-2.0.1-py2.py3-none-any.whl (4.4 kB)\n",
            "Installing collected packages: findspark\n",
            "Successfully installed findspark-2.0.1\n",
            "Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/\n",
            "Collecting pyspark\n",
            "  Downloading pyspark-3.3.0.tar.gz (281.3 MB)\n",
            "\u001b[K     |████████████████████████████████| 281.3 MB 37 kB/s \n",
            "\u001b[?25hCollecting py4j==0.10.9.5\n",
            "  Downloading py4j-0.10.9.5-py2.py3-none-any.whl (199 kB)\n",
            "\u001b[K     |████████████████████████████████| 199 kB 45.1 MB/s \n",
            "\u001b[?25hBuilding wheels for collected packages: pyspark\n",
            "  Building wheel for pyspark (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for pyspark: filename=pyspark-3.3.0-py2.py3-none-any.whl size=281764026 sha256=55a379f61127f3a5ac01bdacb85ac148ce76c954ae4cb743d57000df5325e020\n",
            "  Stored in directory: /root/.cache/pip/wheels/7a/8e/1b/f73a52650d2e5f337708d9f6a1750d451a7349a867f928b885\n",
            "Successfully built pyspark\n",
            "Installing collected packages: py4j, pyspark\n",
            "Successfully installed py4j-0.10.9.5 pyspark-3.3.0\n"
          ]
        }
      ],
      "source": [
        "!pip install findspark\n",
        "!pip install pyspark"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 196
        },
        "id": "BCPQk5voJZGo",
        "outputId": "0e60234f-6077-438a-dcb7-c9e53179bfee"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<SparkContext master=local[*] appName=Distributed Human Health Association>"
            ],
            "text/html": [
              "\n",
              "        <div>\n",
              "            <p><b>SparkContext</b></p>\n",
              "\n",
              "            <p><a href=\"http://be22fa5af8a4:4040\">Spark UI</a></p>\n",
              "\n",
              "            <dl>\n",
              "              <dt>Version</dt>\n",
              "                <dd><code>v3.3.0</code></dd>\n",
              "              <dt>Master</dt>\n",
              "                <dd><code>local[*]</code></dd>\n",
              "              <dt>AppName</dt>\n",
              "                <dd><code>Distributed Human Health Association</code></dd>\n",
              "            </dl>\n",
              "        </div>\n",
              "        "
            ]
          },
          "metadata": {},
          "execution_count": 2
        }
      ],
      "source": [
        "import pyspark\n",
        "import findspark\n",
        "findspark.init()\n",
        "from pyspark.sql import SparkSession\n",
        "\n",
        "def init_spark(app_name: str):\n",
        "    spark = SparkSession.builder.appName(app_name).getOrCreate()\n",
        "    sc = spark.sparkContext\n",
        "    return spark, sc\n",
        "\n",
        "spark, sc = init_spark('Distributed Human Health Association')\n",
        "sc"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "7JVhc0e6Tra6"
      },
      "outputs": [],
      "source": [
        "import pyspark.sql.functions as F\n",
        "from pyspark.ml.feature import VectorAssembler, VectorSizeHint, OneHotEncoder, StringIndexer, IndexToString, SQLTransformer, Interaction, UnivariateFeatureSelector\n",
        "from pyspark.ml import Pipeline\n",
        "from pyspark.ml.classification import RandomForestClassifier, LogisticRegression\n",
        "from pyspark.ml.evaluation import MulticlassClassificationEvaluator\n",
        "from pyspark.ml.tuning import CrossValidator, ParamGridBuilder"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "K-tlPjlCKvhW"
      },
      "source": [
        "# Loading The Static Data"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "LAXtArUoKhrs",
        "outputId": "f66a5d47-9a25-4093-f2a7-79f9d3ea1ce5"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "root\n",
            " |-- Arrival_Time: long (nullable = true)\n",
            " |-- Creation_Time: long (nullable = true)\n",
            " |-- Device: string (nullable = true)\n",
            " |-- Index: long (nullable = true)\n",
            " |-- Model: string (nullable = true)\n",
            " |-- User: string (nullable = true)\n",
            " |-- gt: string (nullable = true)\n",
            " |-- x: double (nullable = true)\n",
            " |-- y: double (nullable = true)\n",
            " |-- z: double (nullable = true)\n",
            "\n",
            "+-------------+-------------------+--------+-----+------+----+-----+-------------+------------+-------------+\n",
            "|Arrival_Time |Creation_Time      |Device  |Index|Model |User|gt   |x            |y           |z            |\n",
            "+-------------+-------------------+--------+-----+------+----+-----+-------------+------------+-------------+\n",
            "|1424686735175|1424686733176178965|nexus4_1|35   |nexus4|g   |stand|0.0014038086 |5.0354E-4   |-0.0124053955|\n",
            "|1424686735378|1424686733382813486|nexus4_1|76   |nexus4|g   |stand|-0.0039367676|0.026138306 |-0.01133728  |\n",
            "|1424686735577|1424686733579072031|nexus4_1|115  |nexus4|g   |stand|0.003540039  |-0.034744263|-0.019882202 |\n",
            "|1424686735779|1424688581834321412|nexus4_2|163  |nexus4|g   |stand|0.002822876  |0.005584717 |0.017318726  |\n",
            "|1424686735982|1424688582035859498|nexus4_2|203  |nexus4|g   |stand|0.0017547607 |-0.018981934|-0.022201538 |\n",
            "+-------------+-------------------+--------+-----+------+----+-----+-------------+------------+-------------+\n",
            "only showing top 5 rows\n",
            "\n",
            "Loaded 329194 Rows and 10 Columns.\n"
          ]
        }
      ],
      "source": [
        "data = spark.read.json('/content/data.json')\n",
        "data.printSchema()\n",
        "data.show(5, truncate=False)\n",
        "print(f\"Loaded {data.count()} Rows and {len(data.columns)} Columns.\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "1tNnH8zuYve1",
        "outputId": "eee3b3f1-c833-4641-c6cd-85a0c4f6a595"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+------------+-------------+------+-----+-----+----+---+---+---+---+\n",
            "|Arrival_Time|Creation_Time|Device|Index|Model|User| gt|  x|  y|  z|\n",
            "+------------+-------------+------+-----+-----+----+---+---+---+---+\n",
            "|           0|            0|     0|    0|    0|   0|  0|  0|  0|  0|\n",
            "+------------+-------------+------+-----+-----+----+---+---+---+---+\n",
            "\n"
          ]
        }
      ],
      "source": [
        "data.select([F.count(F.when(F.isnan(c) | F.col(c).isNull(), c)).alias(c) for c in data.columns]).show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hUP_gnykbT8h"
      },
      "source": [
        "# Building Preprocessing Pipeline"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "sItS4Myi5jYt"
      },
      "outputs": [],
      "source": [
        "userIndexer = StringIndexer(inputCol=\"User\", \n",
        "                            outputCol=\"userIndex\",\n",
        "                            handleInvalid=\"keep\")\n",
        "userIndexerModel = userIndexer.fit(data)\n",
        "userToString = IndexToString(inputCol=\"userIndex\", outputCol=\"User\", labels=userIndexerModel.labels)\n",
        "\n",
        "deviceIndexer = StringIndexer(inputCol=\"Device\", \n",
        "                              outputCol=\"DeviceIndex\",\n",
        "                              handleInvalid=\"keep\")\n",
        "deviceIndexerModel = deviceIndexer.fit(data)\n",
        "deviceToString = IndexToString(inputCol=\"DeviceIndex\", outputCol=\"Device\", labels=deviceIndexerModel.labels)\n",
        "\n",
        "gtIndexer = StringIndexer(inputCol=\"gt\", \n",
        "                          outputCol=\"label\",\n",
        "                          handleInvalid=\"keep\")\n",
        "gtIndexerModel = gtIndexer.fit(data)\n",
        "gtToString = IndexToString(inputCol=\"label\", outputCol=\"gt\", labels=gtIndexerModel.labels)\n",
        "PredToString = IndexToString(inputCol=\"prediction\", outputCol=\"pred\", labels=gtIndexerModel.labels)\n",
        "\n",
        "continuousFeatureSelector = UnivariateFeatureSelector(featuresCol=\"continuous_features\", \n",
        "                                                      outputCol=\"selected_continuous_features\") \\\n",
        "                                                      .setFeatureType(\"continuous\") \\\n",
        "                                                      .setLabelType(\"categorical\") \\\n",
        "                                                      .setSelectionMode(\"numTopFeatures\")\n",
        "\n",
        "categoricalFeatureSelector = UnivariateFeatureSelector(featuresCol=\"categorical_features\", \n",
        "                                                       outputCol=\"selected_categorical_features\") \\\n",
        "                                                      .setFeatureType(\"categorical\") \\\n",
        "                                                      .setLabelType(\"categorical\") \\\n",
        "                                                      .setSelectionMode(\"numTopFeatures\")\n",
        "\n",
        "featureSelector = UnivariateFeatureSelector(featuresCol=\"features\") \\\n",
        "                                          .setFeatureType(\"continuous\") \\\n",
        "                                          .setLabelType(\"categorical\") \\\n",
        "                                          .setSelectionMode(\"numTopFeatures\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "_m5YJLa4ES3o",
        "outputId": "2899a1ca-d79c-4235-acca-52f05cdab741"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-----+-----+\n",
            "|features                                                                                                                                                                                                                                                                                                                                                                                                                                                             |label|gt   |\n",
            "+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-----+-----+\n",
            "|[1.424686735175E12,1.42468673317617894E18,2.0,4.0,0.0014038086,5.0354E-4,-0.0124053955,2.0297322905359066E30,2.84937347035E12,5.6987469407E12,1.9999874911445875E9,7.173867586300195E8,-1.7673802413449635E10,2.8493734663523579E18,5.6987469327047158E18,1.9999874883386255E15,7.173867576235331E14,-1.767380238865347E16,8.0,0.0028076172,0.00100708,-0.024810791,0.0056152344,0.00201416,-0.049621582,7.068737824440001E-7,-1.74148008893013E-5,-6.24661285007E-6]|3.0  |stand|\n",
            "+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-----+-----+\n",
            "only showing top 1 row\n",
            "\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "546083"
            ]
          },
          "metadata": {},
          "execution_count": 67
        }
      ],
      "source": [
        "features = ['Arrival_Time', 'Creation_Time', 'DeviceIndex', 'userIndex', 'x', 'y', 'z']\n",
        "f_statement = str(['Arrival_Time', 'Creation_Time', 'x', 'y', 'z', 'gt']).strip(\"[]\").replace(\"'\", \"\")\n",
        "\n",
        "feature_interactions = [Interaction(inputCols=[f1, f2], \n",
        "                                    outputCol=str(f1+' '+f2))\n",
        "                        for i, f1 in enumerate(features) \n",
        "                        for j, f2 in enumerate(features) \n",
        "                        if i < j]\n",
        "feature_interactions_names = [str(f1+' '+f2)\n",
        "                              for i, f1 in enumerate(features) \n",
        "                              for j, f2 in enumerate(features) \n",
        "                              if i < j]\n",
        "\n",
        "final_features = [*features, *feature_interactions_names]\n",
        "features_to_remove = [\"Arrival_Time Creation_Time\",]\n",
        "for feature in features_to_remove:\n",
        "    final_features.remove(feature)\n",
        "\n",
        "preprocessing_pipeline = Pipeline(stages=[\n",
        "    StringIndexer(\n",
        "                inputCol=\"gt\", \n",
        "                outputCol=\"label\",\n",
        "                handleInvalid=\"keep\"),\n",
        "    StringIndexer(\n",
        "                inputCol=\"Device\", \n",
        "                outputCol=\"DeviceIndex\",\n",
        "                handleInvalid=\"keep\"),\n",
        "    StringIndexer(\n",
        "                inputCol=\"User\", \n",
        "                outputCol=\"userIndex\",\n",
        "                handleInvalid=\"keep\"),\n",
        "    SQLTransformer(\n",
        "        statement=f\"SELECT {f_statement}, (DeviceIndex + 1) AS DeviceIndex, (userIndex + 1) AS userIndex, (label + 1) AS label FROM __THIS__\"),\n",
        "    *feature_interactions,\n",
        "    VectorAssembler(inputCols=final_features,\n",
        "                    outputCol='features'),\n",
        "    VectorSizeHint(inputCol='features', \n",
        "                    size=len(feature_interactions_names) + len(features), \n",
        "                    handleInvalid='skip'),\n",
        "    SQLTransformer(statement=\"SELECT features, label, gt FROM __THIS__\")])\n",
        "\n",
        "preprocessed_data = preprocessing_pipeline.fit(data).transform(data).cache()\n",
        "preprocessed_data.show(1, truncate=False)\n",
        "preprocessed_data.count()"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "features = ['Arrival_Time', 'Creation_Time', 'DeviceIndex', 'userIndex', 'x', 'y', 'z']\n",
        "f_statement = str(['Arrival_Time', 'Creation_Time', 'x', 'y', 'z', 'gt']).strip(\"[]\").replace(\"'\", \"\")\n",
        "\n",
        "feature_interactions = [Interaction(inputCols=[f1, f2], \n",
        "                                    outputCol=str(f1+' '+f2))\n",
        "                        for i, f1 in enumerate(features) \n",
        "                        for j, f2 in enumerate(features) \n",
        "                        if i < j]\n",
        "feature_interactions_names = [str(f1+' '+f2)\n",
        "                              for i, f1 in enumerate(features) \n",
        "                              for j, f2 in enumerate(features) \n",
        "                              if i < j]\n",
        "\n",
        "final_features = [*features, *feature_interactions_names]\n",
        "features_to_remove = [\"Arrival_Time Creation_Time\",]\n",
        "for feature in features_to_remove:\n",
        "    final_features.remove(feature)\n",
        "\n",
        "preprocessing_pipeline = Pipeline(stages=[\n",
        "    gtIndexerModel,\n",
        "    deviceIndexerModel,\n",
        "    userIndexerModel,\n",
        "    SQLTransformer(\n",
        "        statement=f\"SELECT {f_statement}, (DeviceIndex + 1) AS DeviceIndex, (userIndex + 1) AS userIndex, (label + 1) AS label FROM __THIS__\"),\n",
        "    *feature_interactions,\n",
        "    VectorAssembler(inputCols=final_features,\n",
        "                    outputCol='features'),\n",
        "    VectorSizeHint(inputCol='features', \n",
        "                    size=len(final_features), \n",
        "                    handleInvalid='skip'),\n",
        "    SQLTransformer(statement=\"SELECT features, label, gt FROM __THIS__\")])\n",
        "\n",
        "preprocessed_data = preprocessing_pipeline.fit(data).transform(data).cache()"
      ],
      "metadata": {
        "id": "-1LfX07XSot1"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "y6ECdIS1b5K_"
      },
      "source": [
        "# Building Basic Model"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "nXracVcOO-XG"
      },
      "source": [
        "## RandomForest"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ctR2BfOZNpcc",
        "outputId": "c3557acf-07dd-4de3-fade-db35264b69d4"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "bootstrap: Whether bootstrap samples are used when building trees. (default: True)\n",
            "cacheNodeIds: If false, the algorithm will pass trees to executors to match instances with nodes. If true, the algorithm will cache node IDs for each instance. Caching can speed up training of deeper trees. Users can set how often should the cache be checkpointed or disable it by setting checkpointInterval. (default: False)\n",
            "checkpointInterval: set checkpoint interval (>= 1) or disable checkpoint (-1). E.g. 10 means that the cache will get checkpointed every 10 iterations. Note: this setting will be ignored if the checkpoint directory is not set in the SparkContext. (default: 10)\n",
            "featureSubsetStrategy: The number of features to consider for splits at each tree node. Supported options: 'auto' (choose automatically for task: If numTrees == 1, set to 'all'. If numTrees > 1 (forest), set to 'sqrt' for classification and to 'onethird' for regression), 'all' (use all features), 'onethird' (use 1/3 of the features), 'sqrt' (use sqrt(number of features)), 'log2' (use log2(number of features)), 'n' (when n is in the range (0, 1.0], use n * number of features. When n is in the range (1, number of features), use n features). default = 'auto' (default: auto)\n",
            "featuresCol: features column name. (default: features)\n",
            "impurity: Criterion used for information gain calculation (case-insensitive). Supported options: entropy, gini (default: gini)\n",
            "labelCol: label column name. (default: label)\n",
            "leafCol: Leaf indices column name. Predicted leaf index of each instance in each tree by preorder. (default: )\n",
            "maxBins: Max number of bins for discretizing continuous features.  Must be >=2 and >= number of categories for any categorical feature. (default: 32)\n",
            "maxDepth: Maximum depth of the tree. (>= 0) E.g., depth 0 means 1 leaf node; depth 1 means 1 internal node + 2 leaf nodes. Must be in range [0, 30]. (default: 5)\n",
            "maxMemoryInMB: Maximum memory in MB allocated to histogram aggregation. If too small, then 1 node will be split per iteration, and its aggregates may exceed this size. (default: 256)\n",
            "minInfoGain: Minimum information gain for a split to be considered at a tree node. (default: 0.0)\n",
            "minInstancesPerNode: Minimum number of instances each child must have after split. If a split causes the left or right child to have fewer than minInstancesPerNode, the split will be discarded as invalid. Should be >= 1. (default: 1)\n",
            "minWeightFractionPerNode: Minimum fraction of the weighted sample count that each child must have after split. If a split causes the fraction of the total weight in the left or right child to be less than minWeightFractionPerNode, the split will be discarded as invalid. Should be in interval [0.0, 0.5). (default: 0.0)\n",
            "numTrees: Number of trees to train (>= 1). (default: 20)\n",
            "predictionCol: prediction column name. (default: prediction)\n",
            "probabilityCol: Column name for predicted class conditional probabilities. Note: Not all models output well-calibrated probability estimates! These probabilities should be treated as confidences, not precise probabilities. (default: probability)\n",
            "rawPredictionCol: raw prediction (a.k.a. confidence) column name. (default: rawPrediction)\n",
            "seed: random seed. (default: 8161727117636271880)\n",
            "subsamplingRate: Fraction of the training data used for learning each decision tree, in range (0, 1]. (default: 1.0)\n",
            "thresholds: Thresholds in multi-class classification to adjust the probability of predicting each class. Array must have length equal to the number of classes, with values > 0, excepting that at most one value may be 0. The class with largest value p/t is predicted, where p is the original probability of that class and t is the class's threshold. (undefined)\n",
            "weightCol: weight column name. If this is not set or empty, we treat all instance weights as 1.0. (undefined)\n"
          ]
        }
      ],
      "source": [
        "print(RandomForestClassifier().explainParams())"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "aDfwRWg2cwF9"
      },
      "outputs": [],
      "source": [
        "n_trees = 25\n",
        "rf = RandomForestClassifier(numTrees=n_trees,\n",
        "                            featureSubsetStrategy='sqrt',\n",
        "                            subsamplingRate=1 / n_trees,\n",
        "                            maxDepth=30, \n",
        "                            impurity='entropy',\n",
        "                            seed=42)\n",
        "\n",
        "learning_pipeline = rf"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "suCyhBPxO5X9"
      },
      "source": [
        "## GBT and OneVsRest"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kzd6j7ilQC4Q"
      },
      "source": [
        "Impossible in this version of Spark!"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "vIiwmjD9OxXX"
      },
      "outputs": [],
      "source": [
        "from pyspark.ml.classification import GBTClassifier, OneVsRest"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "GAJy7bBVPI3j"
      },
      "outputs": [],
      "source": [
        "gbt = GBTClassifier(maxDepth=5, \n",
        "                    lossType='logistic', \n",
        "                    maxIter=20, \n",
        "                    stepSize=0.1, \n",
        "                    seed=42, \n",
        "                    subsamplingRate=0.7, \n",
        "                    impurity='variance', \n",
        "                    featureSubsetStrategy='all', \n",
        "                    validationTol=0.01, \n",
        "                    validationIndicatorCol=None)\n",
        "\n",
        "ovr = OneVsRest(classifier=gbt)\n",
        "learning_pipeline = ovr"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MmcEiCpni0l0"
      },
      "source": [
        "## LogisticRegression"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "kH9tBgH0tQ23",
        "outputId": "364aa6d0-6f2f-4e74-9eb1-1e560a9b83c9"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "aggregationDepth: suggested depth for treeAggregate (>= 2). (default: 2)\n",
            "elasticNetParam: the ElasticNet mixing parameter, in range [0, 1]. For alpha = 0, the penalty is an L2 penalty. For alpha = 1, it is an L1 penalty. (default: 0.0)\n",
            "family: The name of family which is a description of the label distribution to be used in the model. Supported options: auto, binomial, multinomial (default: auto)\n",
            "featuresCol: features column name. (default: features)\n",
            "fitIntercept: whether to fit an intercept term. (default: True)\n",
            "labelCol: label column name. (default: label)\n",
            "lowerBoundsOnCoefficients: The lower bounds on coefficients if fitting under bound constrained optimization. The bound matrix must be compatible with the shape (1, number of features) for binomial regression, or (number of classes, number of features) for multinomial regression. (undefined)\n",
            "lowerBoundsOnIntercepts: The lower bounds on intercepts if fitting under bound constrained optimization. The bounds vector size must beequal with 1 for binomial regression, or the number oflasses for multinomial regression. (undefined)\n",
            "maxBlockSizeInMB: maximum memory in MB for stacking input data into blocks. Data is stacked within partitions. If more than remaining data size in a partition then it is adjusted to the data size. Default 0.0 represents choosing optimal value, depends on specific algorithm. Must be >= 0. (default: 0.0)\n",
            "maxIter: max number of iterations (>= 0). (default: 100)\n",
            "predictionCol: prediction column name. (default: prediction)\n",
            "probabilityCol: Column name for predicted class conditional probabilities. Note: Not all models output well-calibrated probability estimates! These probabilities should be treated as confidences, not precise probabilities. (default: probability)\n",
            "rawPredictionCol: raw prediction (a.k.a. confidence) column name. (default: rawPrediction)\n",
            "regParam: regularization parameter (>= 0). (default: 0.0)\n",
            "standardization: whether to standardize the training features before fitting the model. (default: True)\n",
            "threshold: Threshold in binary classification prediction, in range [0, 1]. If threshold and thresholds are both set, they must match.e.g. if threshold is p, then thresholds must be equal to [1-p, p]. (default: 0.5)\n",
            "thresholds: Thresholds in multi-class classification to adjust the probability of predicting each class. Array must have length equal to the number of classes, with values > 0, excepting that at most one value may be 0. The class with largest value p/t is predicted, where p is the original probability of that class and t is the class's threshold. (undefined)\n",
            "tol: the convergence tolerance for iterative algorithms (>= 0). (default: 1e-06)\n",
            "upperBoundsOnCoefficients: The upper bounds on coefficients if fitting under bound constrained optimization. The bound matrix must be compatible with the shape (1, number of features) for binomial regression, or (number of classes, number of features) for multinomial regression. (undefined)\n",
            "upperBoundsOnIntercepts: The upper bounds on intercepts if fitting under bound constrained optimization. The bound vector size must be equal with 1 for binomial regression, or the number of classes for multinomial regression. (undefined)\n",
            "weightCol: weight column name. If this is not set or empty, we treat all instance weights as 1.0. (undefined)\n"
          ]
        }
      ],
      "source": [
        "print(LogisticRegression().explainParams())"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Up71dSYDAYa1",
        "outputId": "0a231584-bfb7-494a-d5f5-feff4e820c82"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+----------+-------------------+\n",
            "|gt        |(count(1) / 493773)|\n",
            "+----------+-------------------+\n",
            "|stairsup  |0.13363630656192219|\n",
            "|sit       |0.15965230986708467|\n",
            "|stand     |0.14619673412681536|\n",
            "|walk      |0.16959007479145274|\n",
            "|bike      |0.13722500015189165|\n",
            "|stairsdown|0.11996403205521565|\n",
            "|null      |0.13373554244561772|\n",
            "+----------+-------------------+\n",
            "\n"
          ]
        }
      ],
      "source": [
        "data.groupby(\"gt\").agg(F.count(\"*\") / data.count()).alias(\"freq\").show(truncate=False)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "wo_Ht6GYA5M1"
      },
      "outputs": [],
      "source": [
        "thresholds = [0.0, 0.13397597068577488, 0.1577837068723985, 0.1459375223180359, 0.16992288718015394, 0.13840203778546484, 0.12004951628232338, 0.13392835887584853]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "a8RluZMgpRtp"
      },
      "outputs": [],
      "source": [
        "lr = LogisticRegression() \\\n",
        "            .setMaxIter(1) \\\n",
        "            .setRegParam(0.0) \\\n",
        "            .setElasticNetParam(1.0) \\\n",
        "            .setFamily('multinomial')\n",
        "# RegParam sets the proportion between the regularizer and the loss function.\n",
        "# ElasticNet sets the proportion between the Lasso model regularizer and the Ridge model regularizer.\n",
        "learning_pipeline = lr"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "LogisticRegression:\n",
        "*   MaxIter=100, reg=0.0 --> 0.1686\n",
        "*   MaxIter=100, reg=1.0 --> 0.0496\n",
        "*   MaxIter=100, reg=1000.0 --> 0.0496\n",
        "*   MaxIter=1000, reg=100.0, standardization=True --> 0.0496\n",
        "*   MaxIter=1000, reg=100.0, standardization=False --> 0.0857\n",
        "*   standardization=False --> INCREASES ACCURACY BUT ALSO RUNTIME\n",
        "*   MaxIter=1000, reg=1.0 --> 0.0496\n",
        "\n",
        "*   MaxIter=1000, reg=10.0 --> 0.0857\n",
        "*   MaxIter=1000, reg=100.0 --> 0.0857"
      ],
      "metadata": {
        "id": "FLoPhueqlyU3"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zFxeZr-FTvVS"
      },
      "source": [
        "## Full Pipeline"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "RURFtjjRZhz_"
      },
      "outputs": [],
      "source": [
        "basic_pipeline = Pipeline(stages=[\n",
        "    preprocessing_pipeline,\n",
        "    learning_pipeline,\n",
        "    SQLTransformer(\n",
        "        statement=f\"SELECT {f_statement}, (DeviceIndex - 1) AS DeviceIndex, (userIndex - 1) AS userIndex, (label - 1) AS label FROM __THIS__\"),\n",
        "    PredToString\n",
        "])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "NKHwxnAlW8C6"
      },
      "outputs": [],
      "source": [
        "(trainDF, testDF) = data.randomSplit([0.7, 0.3], seed=42)\n",
        "trainDF = trainDF.cache()\n",
        "testDF = testDF.cache()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "2rqQMozcSxNO"
      },
      "outputs": [],
      "source": [
        "evaluator = MulticlassClassificationEvaluator()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "ZmPKGwTXFEfS"
      },
      "outputs": [],
      "source": [
        "n_trees = 20\n",
        "rf = RandomForestClassifier(numTrees=n_trees,\n",
        "                            featureSubsetStrategy='sqrt',\n",
        "                            subsamplingRate=1 / n_trees,\n",
        "                            maxDepth=30, \n",
        "                            impurity='entropy',\n",
        "                            seed=42)\n",
        "\n",
        "learning_pipeline = rf"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "TJ5Apt-_FG6Q"
      },
      "outputs": [],
      "source": [
        "basic_pipeline = Pipeline(stages=[\n",
        "    preprocessing_pipeline,\n",
        "    learning_pipeline,\n",
        "    PredToString\n",
        "])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "rHlmIJNmE-ak",
        "outputId": "9293fbc4-36f3-4df9-fd9b-3bb94f513b05"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.8029160292266365"
            ]
          },
          "metadata": {},
          "execution_count": 56
        }
      ],
      "source": [
        "pred_set = basic_pipeline.fit(trainDF).transform(trainDF)\n",
        "evaluator.evaluate(pred_set)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "LGR8omaoNKz3",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "839cae6b-c7d7-4911-9705-77002d7f0560"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.7737611093768888"
            ]
          },
          "metadata": {},
          "execution_count": 59
        }
      ],
      "source": [
        "pred_set = basic_pipeline.fit(trainDF).transform(testDF)\n",
        "evaluator.evaluate(pred_set)"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import pdb\n",
        "pdb.pm()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "WVh5gyK357wu",
        "outputId": "697cc7b3-6e20-4e96-82b0-f359aa52763a"
      },
      "execution_count": null,
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "> /usr/lib/python3.7/socket.py(589)readinto()\n",
            "-> return self._sock.recv_into(b)\n",
            "(Pdb) timeout\n",
            "<class 'socket.timeout'>\n",
            "(Pdb) self._timeout_occurred\n",
            "False\n",
            "(Pdb) self._sock.recv_into(b)\n",
            "*** AttributeError: 'NoneType' object has no attribute 'recv_into'\n",
            "(Pdb) q\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "\n",
            "PYDEV DEBUGGER WARNING:\n",
            "sys.settrace() should not be used when the debugger is being used.\n",
            "This may cause the debugger to stop working correctly.\n",
            "If this is needed, please check: \n",
            "http://pydev.blogspot.com/2007/06/why-cant-pydev-debugger-work-with.html\n",
            "to see how to restore the debug tracing back correctly.\n",
            "Call Location:\n",
            "  File \"/usr/lib/python3.7/bdb.py\", line 357, in set_quit\n",
            "    sys.settrace(None)\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mRjsPGqj15xV"
      },
      "source": [
        "RandomForest:\n",
        "\n",
        "with Interactions and impurity='entropy':\n",
        "*   numTrees=25, maxDepth=20 --> validation: 0.8124, train: 0.8046\n",
        "*   numTrees=30, maxDepth=20 --> validation: 0.7511, train: 0.7725\n",
        "*   numTrees=40, maxDepth=20 --> validation: 0.7445, train: 0.7622\n",
        "*   numTrees=25, maxDepth=30 --> validation: 0.8188, train: 0.7876 (max possible maxDepth value)\n",
        "*   numTrees=30, maxDepth=30 --> validation: 0.8121, train: 0.7792\n",
        "*   numTrees=40, maxDepth=30 --> validation: 0.8016, train: 0.7675\n",
        "*   numTrees=40, maxDepth=30, subSamplingRate=0.25 --> Fails!\n",
        "\n",
        "with Interactions and impurity='gini':\n",
        "*   numTrees=25, maxDepth=20 --> validation: 0.7633, train: 0.7873\n",
        "*   numTrees=30, maxDepth=20 --> validation: 0.7555, train: 0.7767\n",
        "*   numTrees=40, maxDepth=20 --> validation: 0.7460, train: 0.7637\n",
        "*   numTrees=25, maxDepth=30 --> validation: 0.7722, train: 0.7987 (max possible maxDepth value)\n",
        "*   gini gives worse results in general!"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 397
        },
        "id": "zW4NFIrL42F6",
        "outputId": "a3d71cd5-7046-404b-8984-803022f376d7"
      },
      "outputs": [
        {
          "ename": "Py4JJavaError",
          "evalue": "ignored",
          "output_type": "error",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mPy4JJavaError\u001b[0m                             Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-32-aafd8b2f9d28>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m     21\u001b[0m     \u001b[0mrf\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msetMaxDepth\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mm\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     22\u001b[0m     \u001b[0mtrain_pred\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mrf\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfit\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtrainDF\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtransform\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtrainDF\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 23\u001b[0;31m     \u001b[0mtest_pred\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mrf\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfit\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtrainDF\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtransform\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtestDF\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     24\u001b[0m     \u001b[0mtrain_scores\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mevaluator\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mevaluate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtrain_pred\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     25\u001b[0m     \u001b[0mtest_scores\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mevaluator\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mevaluate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtest_pred\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/pyspark/ml/base.py\u001b[0m in \u001b[0;36mfit\u001b[0;34m(self, dataset, params)\u001b[0m\n\u001b[1;32m    203\u001b[0m                 \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcopy\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mparams\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_fit\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdataset\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    204\u001b[0m             \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 205\u001b[0;31m                 \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_fit\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdataset\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    206\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    207\u001b[0m             raise TypeError(\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/pyspark/ml/wrapper.py\u001b[0m in \u001b[0;36m_fit\u001b[0;34m(self, dataset)\u001b[0m\n\u001b[1;32m    377\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    378\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0m_fit\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdataset\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mDataFrame\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m->\u001b[0m \u001b[0mJM\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 379\u001b[0;31m         \u001b[0mjava_model\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_fit_java\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdataset\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    380\u001b[0m         \u001b[0mmodel\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_create_model\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mjava_model\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    381\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_copyValues\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmodel\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/pyspark/ml/wrapper.py\u001b[0m in \u001b[0;36m_fit_java\u001b[0;34m(self, dataset)\u001b[0m\n\u001b[1;32m    374\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    375\u001b[0m         \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_transfer_params_to_java\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 376\u001b[0;31m         \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_java_obj\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfit\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdataset\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_jdf\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    377\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    378\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0m_fit\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdataset\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mDataFrame\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m->\u001b[0m \u001b[0mJM\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/py4j/java_gateway.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, *args)\u001b[0m\n\u001b[1;32m   1320\u001b[0m         \u001b[0manswer\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgateway_client\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msend_command\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcommand\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1321\u001b[0m         return_value = get_return_value(\n\u001b[0;32m-> 1322\u001b[0;31m             answer, self.gateway_client, self.target_id, self.name)\n\u001b[0m\u001b[1;32m   1323\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1324\u001b[0m         \u001b[0;32mfor\u001b[0m \u001b[0mtemp_arg\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mtemp_args\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/pyspark/sql/utils.py\u001b[0m in \u001b[0;36mdeco\u001b[0;34m(*a, **kw)\u001b[0m\n\u001b[1;32m    188\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mdeco\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0ma\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mAny\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkw\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mAny\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m->\u001b[0m \u001b[0mAny\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    189\u001b[0m         \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 190\u001b[0;31m             \u001b[0;32mreturn\u001b[0m \u001b[0mf\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0ma\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkw\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    191\u001b[0m         \u001b[0;32mexcept\u001b[0m \u001b[0mPy4JJavaError\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    192\u001b[0m             \u001b[0mconverted\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mconvert_exception\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0me\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mjava_exception\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/py4j/protocol.py\u001b[0m in \u001b[0;36mget_return_value\u001b[0;34m(answer, gateway_client, target_id, name)\u001b[0m\n\u001b[1;32m    326\u001b[0m                 raise Py4JJavaError(\n\u001b[1;32m    327\u001b[0m                     \u001b[0;34m\"An error occurred while calling {0}{1}{2}.\\n\"\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 328\u001b[0;31m                     format(target_id, \".\", name), value)\n\u001b[0m\u001b[1;32m    329\u001b[0m             \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    330\u001b[0m                 raise Py4JError(\n",
            "\u001b[0;31mPy4JJavaError\u001b[0m: An error occurred while calling o4679.fit.\n: java.lang.OutOfMemoryError: Java heap space\n"
          ]
        }
      ],
      "source": [
        "import matplotlib.pyplot as plt\n",
        "\n",
        "preprocessed_data = preprocessing_pipeline.fit(data).transform(data).cache()\n",
        "(trainDF, testDF) = preprocessed_data.randomSplit([0.7, 0.3], seed=42)\n",
        "numTreesList = [25, 25, 30, 40, 40]\n",
        "maxDepthList = [20, 30, 30, 30, 30]\n",
        "\n",
        "train_scores =[]\n",
        "test_scores =[]\n",
        "values = list(zip(numTreesList, maxDepthList))\n",
        "rf = RandomForestClassifier(numTrees=25,\n",
        "                            featureSubsetStrategy='sqrt',\n",
        "                            subsamplingRate=1 / 25,\n",
        "                            maxDepth=20, \n",
        "                            impurity='entropy',\n",
        "                            seed=42)\n",
        "\n",
        "for (n, m) in values:\n",
        "    rf.setNumTrees(n)\n",
        "    rf.setSubsamplingRate(1 / n)\n",
        "    rf.setMaxDepth(m)\n",
        "    train_pred = rf.fit(trainDF).transform(trainDF)\n",
        "    test_pred = rf.fit(trainDF).transform(testDF)\n",
        "    train_scores.append(evaluator.evaluate(train_pred))\n",
        "    test_scores.append(evaluator.evaluate(test_pred))\n",
        "\n",
        "# plot the relationship between r and testing accuracy\n",
        "plt.plot(values, train_scores, label=\"train accuracy\")\n",
        "plt.plot(values, test_scores, label=\"test accuracy\")\n",
        "plt.xlabel('Parameters (numTrees, maxDepth)')\n",
        "plt.ylabel('Accuracy')\n",
        "plt.title('Accuracy over Parameters for Random Forest Classifier')\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "background_save": true,
          "base_uri": "https://localhost:8080/"
        },
        "id": "1N4AD_V8im_-",
        "outputId": "1fda16da-50b5-48ca-adef-f3b5dbe9c9aa"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "> /usr/local/lib/python3.7/dist-packages/py4j/protocol.py(328)get_return_value()\n",
            "-> format(target_id, \".\", name), value)\n",
            "> /usr/local/lib/python3.7/dist-packages/pyspark/sql/utils.py(190)deco()\n",
            "-> return f(*a, **kw)\n",
            "> /usr/local/lib/python3.7/dist-packages/py4j/java_gateway.py(1322)__call__()\n",
            "-> answer, self.gateway_client, self.target_id, self.name)\n",
            "> /usr/local/lib/python3.7/dist-packages/pyspark/ml/wrapper.py(376)_fit_java()\n",
            "-> return self._java_obj.fit(dataset._jdf)\n",
            "> /usr/local/lib/python3.7/dist-packages/pyspark/ml/wrapper.py(379)_fit()\n",
            "-> java_model = self._fit_java(dataset)\n",
            "> /usr/local/lib/python3.7/dist-packages/pyspark/ml/base.py(205)fit()\n",
            "-> return self._fit(dataset)\n",
            "> <ipython-input-32-aafd8b2f9d28>(23)<module>()\n",
            "-> test_pred = rf.fit(trainDF).transform(testDF)\n",
            "Param(parent='RandomForestClassifier_556cd39b0a68', name='numTrees', doc='Number of trees to train (>= 1).')\n",
            "<py4j.java_gateway.JavaMember object at 0x7fa1b662bf10>\n",
            "*** NameError: name 'fr' is not defined\n",
            "40\n",
            "30\n"
          ]
        }
      ],
      "source": [
        "import pdb\n",
        "pdb.pm()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "bpOR7pjZd_G9"
      },
      "source": [
        "# Cross Validation"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Njv_pyQWd-e3"
      },
      "outputs": [],
      "source": [
        "paramGrid = ParamGridBuilder() \\\n",
        ".addGrid(rf.numTrees, [10, 20]) \\\n",
        ".addGrid(rf.maxDepth, [15, 25]) \\\n",
        ".addGrid(rf.subsamplingRate, [0.7, 0.8]) \\\n",
        ".build()\n",
        "\n",
        "numFolds = 5\n",
        "crossval = CrossValidator(\n",
        "    estimator=basic_pipeline,\n",
        "    estimatorParamMaps=paramGrid,\n",
        "    evaluator=evaluator,\n",
        "    numFolds=numFolds)\n",
        "\n",
        "cv_model = crossval.fit(trainDF)\n",
        "cv_model.avgMetrics"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "HEtJTA_rHWf3"
      },
      "outputs": [],
      "source": [
        "rf = RandomForestClassifier(numTrees=20, \n",
        "                            maxBins=36,\n",
        "                            minInstancesPerNode=1,\n",
        "                            featureSubsetStrategy='sqrt',\n",
        "                            subsamplingRate=0.7,\n",
        "                            minInfoGain=0.0,\n",
        "                            maxDepth=25, \n",
        "                            impurity='gini',\n",
        "                            labelCol=\"gtIndex\", \n",
        "                            seed=42)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "vcDE_UA1nlgm"
      },
      "outputs": [],
      "source": [
        "# evaluator.evaluate(cv_model.bestModel.transform(testDF))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "t-hxNr0XIs14"
      },
      "outputs": [],
      "source": [
        "preprocessed_data =preprocess_pipeline.fit(data).transform(trainDF).drop(*[\"Arrival_Time\", \n",
        "                                                                        \"Creation_Time\", \n",
        "                                                                        \"Device\", \n",
        "                                                                        \"Device\", \n",
        "                                                                        \"Index\", \n",
        "                                                                        \"Model\", \n",
        "                                                                        \"User\", \n",
        "                                                                        \"gt\", \n",
        "                                                                        \"x\", \n",
        "                                                                        \"y\", \n",
        "                                                                        \"z\", \n",
        "                                                                        \"DeviceIndex\", \n",
        "                                                                        \"userIndex\", \"init_features\"])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "k-usa03DH2wO",
        "outputId": "2af72219-29f8-4516-a69a-c41bd924634c"
      },
      "outputs": [
        {
          "ename": "Py4JJavaError",
          "evalue": "ignored",
          "output_type": "error",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mPy4JJavaError\u001b[0m                             Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-20-c6e3b2502197>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mrf\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfit\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mpreprocessed_data\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/pyspark/ml/base.py\u001b[0m in \u001b[0;36mfit\u001b[0;34m(self, dataset, params)\u001b[0m\n\u001b[1;32m    203\u001b[0m                 \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcopy\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mparams\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_fit\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdataset\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    204\u001b[0m             \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 205\u001b[0;31m                 \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_fit\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdataset\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    206\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    207\u001b[0m             raise TypeError(\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/pyspark/ml/wrapper.py\u001b[0m in \u001b[0;36m_fit\u001b[0;34m(self, dataset)\u001b[0m\n\u001b[1;32m    377\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    378\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0m_fit\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdataset\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mDataFrame\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m->\u001b[0m \u001b[0mJM\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 379\u001b[0;31m         \u001b[0mjava_model\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_fit_java\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdataset\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    380\u001b[0m         \u001b[0mmodel\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_create_model\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mjava_model\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    381\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_copyValues\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmodel\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/pyspark/ml/wrapper.py\u001b[0m in \u001b[0;36m_fit_java\u001b[0;34m(self, dataset)\u001b[0m\n\u001b[1;32m    374\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    375\u001b[0m         \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_transfer_params_to_java\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 376\u001b[0;31m         \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_java_obj\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfit\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdataset\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_jdf\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    377\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    378\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0m_fit\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdataset\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mDataFrame\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m->\u001b[0m \u001b[0mJM\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/py4j/java_gateway.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, *args)\u001b[0m\n\u001b[1;32m   1320\u001b[0m         \u001b[0manswer\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgateway_client\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msend_command\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcommand\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1321\u001b[0m         return_value = get_return_value(\n\u001b[0;32m-> 1322\u001b[0;31m             answer, self.gateway_client, self.target_id, self.name)\n\u001b[0m\u001b[1;32m   1323\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1324\u001b[0m         \u001b[0;32mfor\u001b[0m \u001b[0mtemp_arg\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mtemp_args\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/pyspark/sql/utils.py\u001b[0m in \u001b[0;36mdeco\u001b[0;34m(*a, **kw)\u001b[0m\n\u001b[1;32m    188\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mdeco\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0ma\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mAny\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkw\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mAny\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m->\u001b[0m \u001b[0mAny\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    189\u001b[0m         \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 190\u001b[0;31m             \u001b[0;32mreturn\u001b[0m \u001b[0mf\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0ma\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkw\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    191\u001b[0m         \u001b[0;32mexcept\u001b[0m \u001b[0mPy4JJavaError\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    192\u001b[0m             \u001b[0mconverted\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mconvert_exception\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0me\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mjava_exception\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/py4j/protocol.py\u001b[0m in \u001b[0;36mget_return_value\u001b[0;34m(answer, gateway_client, target_id, name)\u001b[0m\n\u001b[1;32m    326\u001b[0m                 raise Py4JJavaError(\n\u001b[1;32m    327\u001b[0m                     \u001b[0;34m\"An error occurred while calling {0}{1}{2}.\\n\"\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 328\u001b[0;31m                     format(target_id, \".\", name), value)\n\u001b[0m\u001b[1;32m    329\u001b[0m             \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    330\u001b[0m                 raise Py4JError(\n",
            "\u001b[0;31mPy4JJavaError\u001b[0m: An error occurred while calling o1391.fit.\n: org.apache.spark.SparkException: Job aborted due to stage failure: Task 0 in stage 131.0 failed 1 times, most recent failure: Lost task 0.0 in stage 131.0 (TID 244) (f3f7a2451cea executor driver): java.lang.OutOfMemoryError: Java heap space\n\tat scala.collection.immutable.HashMap$HashMap1.get0(HashMap.scala:359)\n\tat scala.collection.immutable.HashMap$HashTrieMap.get0(HashMap.scala:579)\n\tat scala.collection.immutable.HashMap.get(HashMap.scala:70)\n\tat scala.collection.MapLike.apply(MapLike.scala:143)\n\tat scala.collection.MapLike.apply$(MapLike.scala:143)\n\tat scala.collection.AbstractMap.apply(Map.scala:65)\n\tat org.apache.spark.ml.tree.impl.RandomForest$.$anonfun$findBestSplits$8(RandomForest.scala:560)\n\tat org.apache.spark.ml.tree.impl.RandomForest$.$anonfun$findBestSplits$8$adapted(RandomForest.scala:558)\n\tat org.apache.spark.ml.tree.impl.RandomForest$$$Lambda$3901/0x0000000841486840.apply(Unknown Source)\n\tat scala.collection.immutable.HashMap$HashMap1.foreach(HashMap.scala:400)\n\tat scala.collection.immutable.HashMap$HashTrieMap.foreach(HashMap.scala:728)\n\tat org.apache.spark.ml.tree.impl.RandomForest$.binSeqOp$1(RandomForest.scala:558)\n\tat org.apache.spark.ml.tree.impl.RandomForest$.$anonfun$findBestSplits$24(RandomForest.scala:655)\n\tat org.apache.spark.ml.tree.impl.RandomForest$$$Lambda$3900/0x0000000841487040.apply(Unknown Source)\n\tat scala.collection.Iterator.foreach(Iterator.scala:943)\n\tat scala.collection.Iterator.foreach$(Iterator.scala:943)\n\tat org.apache.spark.InterruptibleIterator.foreach(InterruptibleIterator.scala:28)\n\tat org.apache.spark.ml.tree.impl.RandomForest$.$anonfun$findBestSplits$21(RandomForest.scala:655)\n\tat org.apache.spark.ml.tree.impl.RandomForest$$$Lambda$3874/0x000000084163f840.apply(Unknown Source)\n\tat org.apache.spark.rdd.RDD.$anonfun$mapPartitions$2(RDD.scala:855)\n\tat org.apache.spark.rdd.RDD.$anonfun$mapPartitions$2$adapted(RDD.scala:855)\n\tat org.apache.spark.rdd.RDD$$Lambda$2453/0x0000000841073840.apply(Unknown Source)\n\tat org.apache.spark.rdd.MapPartitionsRDD.compute(MapPartitionsRDD.scala:52)\n\tat org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:365)\n\tat org.apache.spark.rdd.RDD.iterator(RDD.scala:329)\n\tat org.apache.spark.shuffle.ShuffleWriteProcessor.write(ShuffleWriteProcessor.scala:59)\n\tat org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:99)\n\tat org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:52)\n\tat org.apache.spark.scheduler.Task.run(Task.scala:136)\n\tat org.apache.spark.executor.Executor$TaskRunner.$anonfun$run$3(Executor.scala:548)\n\tat org.apache.spark.executor.Executor$TaskRunner$$Lambda$2408/0x0000000841018c40.apply(Unknown Source)\n\tat org.apache.spark.util.Utils$.tryWithSafeFinally(Utils.scala:1504)\n\nDriver stacktrace:\n\tat org.apache.spark.scheduler.DAGScheduler.failJobAndIndependentStages(DAGScheduler.scala:2672)\n\tat org.apache.spark.scheduler.DAGScheduler.$anonfun$abortStage$2(DAGScheduler.scala:2608)\n\tat org.apache.spark.scheduler.DAGScheduler.$anonfun$abortStage$2$adapted(DAGScheduler.scala:2607)\n\tat scala.collection.mutable.ResizableArray.foreach(ResizableArray.scala:62)\n\tat scala.collection.mutable.ResizableArray.foreach$(ResizableArray.scala:55)\n\tat scala.collection.mutable.ArrayBuffer.foreach(ArrayBuffer.scala:49)\n\tat org.apache.spark.scheduler.DAGScheduler.abortStage(DAGScheduler.scala:2607)\n\tat org.apache.spark.scheduler.DAGScheduler.$anonfun$handleTaskSetFailed$1(DAGScheduler.scala:1182)\n\tat org.apache.spark.scheduler.DAGScheduler.$anonfun$handleTaskSetFailed$1$adapted(DAGScheduler.scala:1182)\n\tat scala.Option.foreach(Option.scala:407)\n\tat org.apache.spark.scheduler.DAGScheduler.handleTaskSetFailed(DAGScheduler.scala:1182)\n\tat org.apache.spark.scheduler.DAGSchedulerEventProcessLoop.doOnReceive(DAGScheduler.scala:2860)\n\tat org.apache.spark.scheduler.DAGSchedulerEventProcessLoop.onReceive(DAGScheduler.scala:2802)\n\tat org.apache.spark.scheduler.DAGSchedulerEventProcessLoop.onReceive(DAGScheduler.scala:2791)\n\tat org.apache.spark.util.EventLoop$$anon$1.run(EventLoop.scala:49)\n\tat org.apache.spark.scheduler.DAGScheduler.runJob(DAGScheduler.scala:952)\n\tat org.apache.spark.SparkContext.runJob(SparkContext.scala:2228)\n\tat org.apache.spark.SparkContext.runJob(SparkContext.scala:2249)\n\tat org.apache.spark.SparkContext.runJob(SparkContext.scala:2268)\n\tat org.apache.spark.SparkContext.runJob(SparkContext.scala:2293)\n\tat org.apache.spark.rdd.RDD.$anonfun$collect$1(RDD.scala:1021)\n\tat org.apache.spark.rdd.RDDOperationScope$.withScope(RDDOperationScope.scala:151)\n\tat org.apache.spark.rdd.RDDOperationScope$.withScope(RDDOperationScope.scala:112)\n\tat org.apache.spark.rdd.RDD.withScope(RDD.scala:406)\n\tat org.apache.spark.rdd.RDD.collect(RDD.scala:1020)\n\tat org.apache.spark.rdd.PairRDDFunctions.$anonfun$collectAsMap$1(PairRDDFunctions.scala:738)\n\tat org.apache.spark.rdd.RDDOperationScope$.withScope(RDDOperationScope.scala:151)\n\tat org.apache.spark.rdd.RDDOperationScope$.withScope(RDDOperationScope.scala:112)\n\tat org.apache.spark.rdd.RDD.withScope(RDD.scala:406)\n\tat org.apache.spark.rdd.PairRDDFunctions.collectAsMap(PairRDDFunctions.scala:737)\n\tat org.apache.spark.ml.tree.impl.RandomForest$.findBestSplits(RandomForest.scala:663)\n\tat org.apache.spark.ml.tree.impl.RandomForest$.runBagged(RandomForest.scala:208)\n\tat org.apache.spark.ml.tree.impl.RandomForest$.run(RandomForest.scala:302)\n\tat org.apache.spark.ml.classification.RandomForestClassifier.$anonfun$train$1(RandomForestClassifier.scala:161)\n\tat org.apache.spark.ml.util.Instrumentation$.$anonfun$instrumented$1(Instrumentation.scala:191)\n\tat scala.util.Try$.apply(Try.scala:213)\n\tat org.apache.spark.ml.util.Instrumentation$.instrumented(Instrumentation.scala:191)\n\tat org.apache.spark.ml.classification.RandomForestClassifier.train(RandomForestClassifier.scala:138)\n\tat org.apache.spark.ml.classification.RandomForestClassifier.train(RandomForestClassifier.scala:46)\n\tat org.apache.spark.ml.Predictor.fit(Predictor.scala:151)\n\tat org.apache.spark.ml.Predictor.fit(Predictor.scala:115)\n\tat java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke0(Native Method)\n\tat java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)\n\tat java.base/jdk.internal.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)\n\tat java.base/java.lang.reflect.Method.invoke(Method.java:566)\n\tat py4j.reflection.MethodInvoker.invoke(MethodInvoker.java:244)\n\tat py4j.reflection.ReflectionEngine.invoke(ReflectionEngine.java:357)\n\tat py4j.Gateway.invoke(Gateway.java:282)\n\tat py4j.commands.AbstractCommand.invokeMethod(AbstractCommand.java:132)\n\tat py4j.commands.CallCommand.execute(CallCommand.java:79)\n\tat py4j.ClientServerConnection.waitForCommands(ClientServerConnection.java:182)\n\tat py4j.ClientServerConnection.run(ClientServerConnection.java:106)\n\tat java.base/java.lang.Thread.run(Thread.java:829)\nCaused by: java.lang.OutOfMemoryError: Java heap space\n\tat scala.collection.immutable.HashMap$HashMap1.get0(HashMap.scala:359)\n\tat scala.collection.immutable.HashMap$HashTrieMap.get0(HashMap.scala:579)\n\tat scala.collection.immutable.HashMap.get(HashMap.scala:70)\n\tat scala.collection.MapLike.apply(MapLike.scala:143)\n\tat scala.collection.MapLike.apply$(MapLike.scala:143)\n\tat scala.collection.AbstractMap.apply(Map.scala:65)\n\tat org.apache.spark.ml.tree.impl.RandomForest$.$anonfun$findBestSplits$8(RandomForest.scala:560)\n\tat org.apache.spark.ml.tree.impl.RandomForest$.$anonfun$findBestSplits$8$adapted(RandomForest.scala:558)\n\tat org.apache.spark.ml.tree.impl.RandomForest$$$Lambda$3901/0x0000000841486840.apply(Unknown Source)\n\tat scala.collection.immutable.HashMap$HashMap1.foreach(HashMap.scala:400)\n\tat scala.collection.immutable.HashMap$HashTrieMap.foreach(HashMap.scala:728)\n\tat org.apache.spark.ml.tree.impl.RandomForest$.binSeqOp$1(RandomForest.scala:558)\n\tat org.apache.spark.ml.tree.impl.RandomForest$.$anonfun$findBestSplits$24(RandomForest.scala:655)\n\tat org.apache.spark.ml.tree.impl.RandomForest$$$Lambda$3900/0x0000000841487040.apply(Unknown Source)\n\tat scala.collection.Iterator.foreach(Iterator.scala:943)\n\tat scala.collection.Iterator.foreach$(Iterator.scala:943)\n\tat org.apache.spark.InterruptibleIterator.foreach(InterruptibleIterator.scala:28)\n\tat org.apache.spark.ml.tree.impl.RandomForest$.$anonfun$findBestSplits$21(RandomForest.scala:655)\n\tat org.apache.spark.ml.tree.impl.RandomForest$$$Lambda$3874/0x000000084163f840.apply(Unknown Source)\n\tat org.apache.spark.rdd.RDD.$anonfun$mapPartitions$2(RDD.scala:855)\n\tat org.apache.spark.rdd.RDD.$anonfun$mapPartitions$2$adapted(RDD.scala:855)\n\tat org.apache.spark.rdd.RDD$$Lambda$2453/0x0000000841073840.apply(Unknown Source)\n\tat org.apache.spark.rdd.MapPartitionsRDD.compute(MapPartitionsRDD.scala:52)\n\tat org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:365)\n\tat org.apache.spark.rdd.RDD.iterator(RDD.scala:329)\n\tat org.apache.spark.shuffle.ShuffleWriteProcessor.write(ShuffleWriteProcessor.scala:59)\n\tat org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:99)\n\tat org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:52)\n\tat org.apache.spark.scheduler.Task.run(Task.scala:136)\n\tat org.apache.spark.executor.Executor$TaskRunner.$anonfun$run$3(Executor.scala:548)\n\tat org.apache.spark.executor.Executor$TaskRunner$$Lambda$2408/0x0000000841018c40.apply(Unknown Source)\n\tat org.apache.spark.util.Utils$.tryWithSafeFinally(Utils.scala:1504)\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "ERROR:root:Exception while sending command.\n",
            "Traceback (most recent call last):\n",
            "  File \"/usr/local/lib/python3.7/dist-packages/py4j/clientserver.py\", line 516, in send_command\n",
            "    raise Py4JNetworkError(\"Answer from Java side is empty\")\n",
            "py4j.protocol.Py4JNetworkError: Answer from Java side is empty\n",
            "\n",
            "During handling of the above exception, another exception occurred:\n",
            "\n",
            "Traceback (most recent call last):\n",
            "  File \"/usr/local/lib/python3.7/dist-packages/py4j/java_gateway.py\", line 1038, in send_command\n",
            "    response = connection.send_command(command)\n",
            "  File \"/usr/local/lib/python3.7/dist-packages/py4j/clientserver.py\", line 540, in send_command\n",
            "    \"Error while sending or receiving\", e, proto.ERROR_ON_RECEIVE)\n",
            "py4j.protocol.Py4JNetworkError: Error while sending or receiving\n"
          ]
        }
      ],
      "source": [
        "rf.fit(preprocessed_data)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 165
        },
        "id": "AKVaPAzaIjMw",
        "outputId": "f51b4919-5a97-49e9-f982-f8fd8dd3dfcd"
      },
      "outputs": [
        {
          "ename": "AttributeError",
          "evalue": "ignored",
          "output_type": "error",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mAttributeError\u001b[0m                            Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-21-d84ad56f1f56>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mevaluator\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mevaluate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mrf\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtransform\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtestDF\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
            "\u001b[0;31mAttributeError\u001b[0m: 'RandomForestClassifier' object has no attribute 'transform'"
          ]
        }
      ],
      "source": [
        "evaluator.evaluate(rf.transform(testDF))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FpzeYxQNqBJl"
      },
      "source": [
        "much better!"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "RdPq3yyPp-41"
      },
      "outputs": [],
      "source": [
        "best_rf = cv_model.bestModel.stages[1].stages[0]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "xtc8wjgLqux6"
      },
      "outputs": [],
      "source": [
        "print(best_rf._java_obj.getNumTrees())\n",
        "print(best_rf._java_obj.getMaxDepth())\n",
        "print(best_rf._java_obj.getSubsamplingRate())"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "sgwyg0PV6GOa"
      },
      "source": [
        "# Implementing Grarient Boosted Tree in Pyspark"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "fXmgdiKO_FkA"
      },
      "source": [
        "## Trying Stuff"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "rb7qNVoW_ZC2"
      },
      "outputs": [],
      "source": [
        "import pandas as pd\n",
        "from sklearn.ensemble import GradientBoostingClassifier\n",
        "import numpy as np\n",
        "from sklearn import metrics"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "5yOdy7Q3_vRL"
      },
      "outputs": [],
      "source": [
        "import xgboost as xgb"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "0urYRPRn6RhH"
      },
      "outputs": [],
      "source": [
        "data_pd = data.toPandas()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "MHy_i-7WELXI",
        "outputId": "6edd609d-86ff-41af-f85a-f4d8293825db"
      },
      "outputs": [
        {
          "data": {
            "text/html": [
              "\n",
              "  <div id=\"df-766f2e76-8545-4bc2-8c06-424aca1edb67\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Arrival_Time</th>\n",
              "      <th>Creation_Time</th>\n",
              "      <th>Device</th>\n",
              "      <th>Index</th>\n",
              "      <th>Model</th>\n",
              "      <th>User</th>\n",
              "      <th>gt</th>\n",
              "      <th>x</th>\n",
              "      <th>y</th>\n",
              "      <th>z</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>1424686735175</td>\n",
              "      <td>1424686733176178965</td>\n",
              "      <td>nexus4_1</td>\n",
              "      <td>35</td>\n",
              "      <td>nexus4</td>\n",
              "      <td>g</td>\n",
              "      <td>stand</td>\n",
              "      <td>0.001404</td>\n",
              "      <td>0.000504</td>\n",
              "      <td>-0.012405</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1424686735378</td>\n",
              "      <td>1424686733382813486</td>\n",
              "      <td>nexus4_1</td>\n",
              "      <td>76</td>\n",
              "      <td>nexus4</td>\n",
              "      <td>g</td>\n",
              "      <td>stand</td>\n",
              "      <td>-0.003937</td>\n",
              "      <td>0.026138</td>\n",
              "      <td>-0.011337</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>1424686735577</td>\n",
              "      <td>1424686733579072031</td>\n",
              "      <td>nexus4_1</td>\n",
              "      <td>115</td>\n",
              "      <td>nexus4</td>\n",
              "      <td>g</td>\n",
              "      <td>stand</td>\n",
              "      <td>0.003540</td>\n",
              "      <td>-0.034744</td>\n",
              "      <td>-0.019882</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>1424686735779</td>\n",
              "      <td>1424688581834321412</td>\n",
              "      <td>nexus4_2</td>\n",
              "      <td>163</td>\n",
              "      <td>nexus4</td>\n",
              "      <td>g</td>\n",
              "      <td>stand</td>\n",
              "      <td>0.002823</td>\n",
              "      <td>0.005585</td>\n",
              "      <td>0.017319</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>1424686735982</td>\n",
              "      <td>1424688582035859498</td>\n",
              "      <td>nexus4_2</td>\n",
              "      <td>203</td>\n",
              "      <td>nexus4</td>\n",
              "      <td>g</td>\n",
              "      <td>stand</td>\n",
              "      <td>0.001755</td>\n",
              "      <td>-0.018982</td>\n",
              "      <td>-0.022202</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>546078</th>\n",
              "      <td>1424789459800</td>\n",
              "      <td>1424789463282162951</td>\n",
              "      <td>nexus4_1</td>\n",
              "      <td>371676</td>\n",
              "      <td>nexus4</td>\n",
              "      <td>e</td>\n",
              "      <td>bike</td>\n",
              "      <td>-0.699982</td>\n",
              "      <td>0.223953</td>\n",
              "      <td>-0.088852</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>546079</th>\n",
              "      <td>1424789459999</td>\n",
              "      <td>1424791305316714089</td>\n",
              "      <td>nexus4_2</td>\n",
              "      <td>396217</td>\n",
              "      <td>nexus4</td>\n",
              "      <td>e</td>\n",
              "      <td>bike</td>\n",
              "      <td>-1.009277</td>\n",
              "      <td>0.855545</td>\n",
              "      <td>-0.293488</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>546080</th>\n",
              "      <td>1424789460198</td>\n",
              "      <td>1424789463679929064</td>\n",
              "      <td>nexus4_1</td>\n",
              "      <td>371755</td>\n",
              "      <td>nexus4</td>\n",
              "      <td>e</td>\n",
              "      <td>bike</td>\n",
              "      <td>0.308319</td>\n",
              "      <td>-0.329330</td>\n",
              "      <td>-0.193527</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>546081</th>\n",
              "      <td>1424789460407</td>\n",
              "      <td>1424789463886533068</td>\n",
              "      <td>nexus4_1</td>\n",
              "      <td>371796</td>\n",
              "      <td>nexus4</td>\n",
              "      <td>e</td>\n",
              "      <td>bike</td>\n",
              "      <td>1.356140</td>\n",
              "      <td>0.657608</td>\n",
              "      <td>0.009415</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>546082</th>\n",
              "      <td>1424789460601</td>\n",
              "      <td>1424791305920931618</td>\n",
              "      <td>nexus4_2</td>\n",
              "      <td>396337</td>\n",
              "      <td>nexus4</td>\n",
              "      <td>e</td>\n",
              "      <td>bike</td>\n",
              "      <td>1.418549</td>\n",
              "      <td>-0.184799</td>\n",
              "      <td>0.287567</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>546083 rows × 10 columns</p>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-766f2e76-8545-4bc2-8c06-424aca1edb67')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-766f2e76-8545-4bc2-8c06-424aca1edb67 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-766f2e76-8545-4bc2-8c06-424aca1edb67');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ],
            "text/plain": [
              "         Arrival_Time        Creation_Time    Device   Index   Model User  \\\n",
              "0       1424686735175  1424686733176178965  nexus4_1      35  nexus4    g   \n",
              "1       1424686735378  1424686733382813486  nexus4_1      76  nexus4    g   \n",
              "2       1424686735577  1424686733579072031  nexus4_1     115  nexus4    g   \n",
              "3       1424686735779  1424688581834321412  nexus4_2     163  nexus4    g   \n",
              "4       1424686735982  1424688582035859498  nexus4_2     203  nexus4    g   \n",
              "...               ...                  ...       ...     ...     ...  ...   \n",
              "546078  1424789459800  1424789463282162951  nexus4_1  371676  nexus4    e   \n",
              "546079  1424789459999  1424791305316714089  nexus4_2  396217  nexus4    e   \n",
              "546080  1424789460198  1424789463679929064  nexus4_1  371755  nexus4    e   \n",
              "546081  1424789460407  1424789463886533068  nexus4_1  371796  nexus4    e   \n",
              "546082  1424789460601  1424791305920931618  nexus4_2  396337  nexus4    e   \n",
              "\n",
              "           gt         x         y         z  \n",
              "0       stand  0.001404  0.000504 -0.012405  \n",
              "1       stand -0.003937  0.026138 -0.011337  \n",
              "2       stand  0.003540 -0.034744 -0.019882  \n",
              "3       stand  0.002823  0.005585  0.017319  \n",
              "4       stand  0.001755 -0.018982 -0.022202  \n",
              "...       ...       ...       ...       ...  \n",
              "546078   bike -0.699982  0.223953 -0.088852  \n",
              "546079   bike -1.009277  0.855545 -0.293488  \n",
              "546080   bike  0.308319 -0.329330 -0.193527  \n",
              "546081   bike  1.356140  0.657608  0.009415  \n",
              "546082   bike  1.418549 -0.184799  0.287567  \n",
              "\n",
              "[546083 rows x 10 columns]"
            ]
          },
          "execution_count": 18,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "data_pd"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "z51Uhq2ZDgqH"
      },
      "outputs": [],
      "source": [
        "gt_idx = {\"bike\": 1,\n",
        "          \"null\": 2,\n",
        "          \"sit\": 3,\n",
        "          \"stairsdown\": 4,\n",
        "          \"stairsup\": 5,\n",
        "          \"stand\": 6,\n",
        "          \"walk\": 7}\n",
        "Device_idx = {\"nexus4_1\": 1, \n",
        "              \"nexus4_2\": 2}\n",
        "Model_idx = {\"nexus4\": 0}\n",
        "User_idx = {\"a\": 1,\n",
        "            \"b\": 2,\n",
        "            \"c\": 3,\n",
        "            \"d\": 4,\n",
        "            \"e\": 5,\n",
        "            \"f\": 6,\n",
        "            \"g\": 7,\n",
        "            \"h\": 8,\n",
        "            \"i\": 9,}"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "r0oSCvIXD7JQ",
        "outputId": "5e37bb69-3d74-4890-a055-91e76bd325f8"
      },
      "outputs": [
        {
          "data": {
            "text/html": [
              "\n",
              "  <div id=\"df-9fe63159-1f69-4c42-a909-58d4e603840e\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Arrival_Time</th>\n",
              "      <th>Creation_Time</th>\n",
              "      <th>Device</th>\n",
              "      <th>Index</th>\n",
              "      <th>Model</th>\n",
              "      <th>User</th>\n",
              "      <th>gt</th>\n",
              "      <th>x</th>\n",
              "      <th>y</th>\n",
              "      <th>z</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>1424686735175</td>\n",
              "      <td>1424686733176178965</td>\n",
              "      <td>1</td>\n",
              "      <td>35</td>\n",
              "      <td>0</td>\n",
              "      <td>7</td>\n",
              "      <td>6</td>\n",
              "      <td>0.001404</td>\n",
              "      <td>0.000504</td>\n",
              "      <td>-0.012405</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1424686735378</td>\n",
              "      <td>1424686733382813486</td>\n",
              "      <td>1</td>\n",
              "      <td>76</td>\n",
              "      <td>0</td>\n",
              "      <td>7</td>\n",
              "      <td>6</td>\n",
              "      <td>-0.003937</td>\n",
              "      <td>0.026138</td>\n",
              "      <td>-0.011337</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>1424686735577</td>\n",
              "      <td>1424686733579072031</td>\n",
              "      <td>1</td>\n",
              "      <td>115</td>\n",
              "      <td>0</td>\n",
              "      <td>7</td>\n",
              "      <td>6</td>\n",
              "      <td>0.003540</td>\n",
              "      <td>-0.034744</td>\n",
              "      <td>-0.019882</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>1424686735779</td>\n",
              "      <td>1424688581834321412</td>\n",
              "      <td>2</td>\n",
              "      <td>163</td>\n",
              "      <td>0</td>\n",
              "      <td>7</td>\n",
              "      <td>6</td>\n",
              "      <td>0.002823</td>\n",
              "      <td>0.005585</td>\n",
              "      <td>0.017319</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>1424686735982</td>\n",
              "      <td>1424688582035859498</td>\n",
              "      <td>2</td>\n",
              "      <td>203</td>\n",
              "      <td>0</td>\n",
              "      <td>7</td>\n",
              "      <td>6</td>\n",
              "      <td>0.001755</td>\n",
              "      <td>-0.018982</td>\n",
              "      <td>-0.022202</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>546078</th>\n",
              "      <td>1424789459800</td>\n",
              "      <td>1424789463282162951</td>\n",
              "      <td>1</td>\n",
              "      <td>371676</td>\n",
              "      <td>0</td>\n",
              "      <td>5</td>\n",
              "      <td>1</td>\n",
              "      <td>-0.699982</td>\n",
              "      <td>0.223953</td>\n",
              "      <td>-0.088852</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>546079</th>\n",
              "      <td>1424789459999</td>\n",
              "      <td>1424791305316714089</td>\n",
              "      <td>2</td>\n",
              "      <td>396217</td>\n",
              "      <td>0</td>\n",
              "      <td>5</td>\n",
              "      <td>1</td>\n",
              "      <td>-1.009277</td>\n",
              "      <td>0.855545</td>\n",
              "      <td>-0.293488</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>546080</th>\n",
              "      <td>1424789460198</td>\n",
              "      <td>1424789463679929064</td>\n",
              "      <td>1</td>\n",
              "      <td>371755</td>\n",
              "      <td>0</td>\n",
              "      <td>5</td>\n",
              "      <td>1</td>\n",
              "      <td>0.308319</td>\n",
              "      <td>-0.329330</td>\n",
              "      <td>-0.193527</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>546081</th>\n",
              "      <td>1424789460407</td>\n",
              "      <td>1424789463886533068</td>\n",
              "      <td>1</td>\n",
              "      <td>371796</td>\n",
              "      <td>0</td>\n",
              "      <td>5</td>\n",
              "      <td>1</td>\n",
              "      <td>1.356140</td>\n",
              "      <td>0.657608</td>\n",
              "      <td>0.009415</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>546082</th>\n",
              "      <td>1424789460601</td>\n",
              "      <td>1424791305920931618</td>\n",
              "      <td>2</td>\n",
              "      <td>396337</td>\n",
              "      <td>0</td>\n",
              "      <td>5</td>\n",
              "      <td>1</td>\n",
              "      <td>1.418549</td>\n",
              "      <td>-0.184799</td>\n",
              "      <td>0.287567</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>546083 rows × 10 columns</p>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-9fe63159-1f69-4c42-a909-58d4e603840e')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-9fe63159-1f69-4c42-a909-58d4e603840e button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-9fe63159-1f69-4c42-a909-58d4e603840e');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ],
            "text/plain": [
              "         Arrival_Time        Creation_Time  Device   Index  Model  User  gt  \\\n",
              "0       1424686735175  1424686733176178965       1      35      0     7   6   \n",
              "1       1424686735378  1424686733382813486       1      76      0     7   6   \n",
              "2       1424686735577  1424686733579072031       1     115      0     7   6   \n",
              "3       1424686735779  1424688581834321412       2     163      0     7   6   \n",
              "4       1424686735982  1424688582035859498       2     203      0     7   6   \n",
              "...               ...                  ...     ...     ...    ...   ...  ..   \n",
              "546078  1424789459800  1424789463282162951       1  371676      0     5   1   \n",
              "546079  1424789459999  1424791305316714089       2  396217      0     5   1   \n",
              "546080  1424789460198  1424789463679929064       1  371755      0     5   1   \n",
              "546081  1424789460407  1424789463886533068       1  371796      0     5   1   \n",
              "546082  1424789460601  1424791305920931618       2  396337      0     5   1   \n",
              "\n",
              "               x         y         z  \n",
              "0       0.001404  0.000504 -0.012405  \n",
              "1      -0.003937  0.026138 -0.011337  \n",
              "2       0.003540 -0.034744 -0.019882  \n",
              "3       0.002823  0.005585  0.017319  \n",
              "4       0.001755 -0.018982 -0.022202  \n",
              "...          ...       ...       ...  \n",
              "546078 -0.699982  0.223953 -0.088852  \n",
              "546079 -1.009277  0.855545 -0.293488  \n",
              "546080  0.308319 -0.329330 -0.193527  \n",
              "546081  1.356140  0.657608  0.009415  \n",
              "546082  1.418549 -0.184799  0.287567  \n",
              "\n",
              "[546083 rows x 10 columns]"
            ]
          },
          "execution_count": 20,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "data_pd[\"gt\"] = data_pd.apply(func=lambda row: gt_idx[row[\"gt\"]], axis=1)\n",
        "data_pd[\"Device\"] = data_pd.apply(func=lambda row: Device_idx[row[\"Device\"]], axis=1)\n",
        "data_pd[\"Model\"] = data_pd.apply(func=lambda row: Model_idx[row[\"Model\"]], axis=1)\n",
        "data_pd[\"User\"] = data_pd.apply(func=lambda row: User_idx[row[\"User\"]], axis=1)\n",
        "data_pd"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "O1bMGx8d_7Kr"
      },
      "outputs": [],
      "source": [
        "from sklearn.model_selection import train_test_split\n",
        "X_train, X_validation = train_test_split(data_pd, train_size=0.7, random_state=0)\n",
        "X_train = pd.DataFrame(X_train)\n",
        "X_validation = pd.DataFrame(X_validation)\n",
        "X_train, y_train = X_train.drop(\"gt\", axis=1), X_train[\"gt\"]\n",
        "X_validation, y_validation = X_validation.drop(\"gt\", axis=1), X_validation[\"gt\"]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "iETFvkRrCQ2U"
      },
      "outputs": [],
      "source": [
        "evallist = [(X_validation, y_validation), (X_train, y_train)]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "N7WEZh1hPr4X",
        "outputId": "b9ea0af6-72f5-4187-c5b9-5275c7af2899"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Parameter optimization\n",
            "[0]\tvalidation_0-merror:0.254857\tvalidation_1-merror:0.253538\n",
            "[1]\tvalidation_0-merror:0.258996\tvalidation_1-merror:0.257601\n",
            "[2]\tvalidation_0-merror:0.250945\tvalidation_1-merror:0.249965\n",
            "[3]\tvalidation_0-merror:0.240641\tvalidation_1-merror:0.239336\n",
            "[4]\tvalidation_0-merror:0.238047\tvalidation_1-merror:0.236631\n",
            "[5]\tvalidation_0-merror:0.236667\tvalidation_1-merror:0.235223\n",
            "[6]\tvalidation_0-merror:0.236099\tvalidation_1-merror:0.2347\n",
            "[7]\tvalidation_0-merror:0.235092\tvalidation_1-merror:0.23373\n",
            "[8]\tvalidation_0-merror:0.23505\tvalidation_1-merror:0.233698\n",
            "[9]\tvalidation_0-merror:0.234769\tvalidation_1-merror:0.233319\n",
            "[10]\tvalidation_0-merror:0.235111\tvalidation_1-merror:0.23368\n",
            "[11]\tvalidation_0-merror:0.234775\tvalidation_1-merror:0.233444\n",
            "[12]\tvalidation_0-merror:0.235349\tvalidation_1-merror:0.234114\n",
            "[13]\tvalidation_0-merror:0.234708\tvalidation_1-merror:0.233233\n",
            "[14]\tvalidation_0-merror:0.234\tvalidation_1-merror:0.232806\n",
            "[15]\tvalidation_0-merror:0.228988\tvalidation_1-merror:0.227812\n",
            "[16]\tvalidation_0-merror:0.226211\tvalidation_1-merror:0.225081\n",
            "[17]\tvalidation_0-merror:0.224734\tvalidation_1-merror:0.22366\n",
            "[18]\tvalidation_0-merror:0.224569\tvalidation_1-merror:0.223684\n",
            "[19]\tvalidation_0-merror:0.223202\tvalidation_1-merror:0.222337\n",
            "[20]\tvalidation_0-merror:0.222597\tvalidation_1-merror:0.221751\n",
            "[21]\tvalidation_0-merror:0.222426\tvalidation_1-merror:0.221651\n",
            "[22]\tvalidation_0-merror:0.221352\tvalidation_1-merror:0.220466\n",
            "[23]\tvalidation_0-merror:0.222036\tvalidation_1-merror:0.221314\n",
            "[24]\tvalidation_0-merror:0.236588\tvalidation_1-merror:0.235788\n",
            "[25]\tvalidation_0-merror:0.236948\tvalidation_1-merror:0.236223\n",
            "[26]\tvalidation_0-merror:0.233658\tvalidation_1-merror:0.232746\n",
            "[27]\tvalidation_0-merror:0.233652\tvalidation_1-merror:0.232699\n",
            "[28]\tvalidation_0-merror:0.231241\tvalidation_1-merror:0.230172\n",
            "[29]\tvalidation_0-merror:0.229336\tvalidation_1-merror:0.228202\n",
            "[30]\tvalidation_0-merror:0.227664\tvalidation_1-merror:0.226269\n",
            "[31]\tvalidation_0-merror:0.223568\tvalidation_1-merror:0.222206\n",
            "[32]\tvalidation_0-merror:0.207337\tvalidation_1-merror:0.206348\n",
            "[33]\tvalidation_0-merror:0.206599\tvalidation_1-merror:0.205746\n",
            "[34]\tvalidation_0-merror:0.206489\tvalidation_1-merror:0.205699\n",
            "[35]\tvalidation_0-merror:0.205292\tvalidation_1-merror:0.204511\n",
            "[36]\tvalidation_0-merror:0.204279\tvalidation_1-merror:0.203096\n",
            "[37]\tvalidation_0-merror:0.203174\tvalidation_1-merror:0.202185\n",
            "[38]\tvalidation_0-merror:0.201648\tvalidation_1-merror:0.200493\n",
            "[39]\tvalidation_0-merror:0.201196\tvalidation_1-merror:0.200158\n",
            "[40]\tvalidation_0-merror:0.198663\tvalidation_1-merror:0.197168\n",
            "[41]\tvalidation_0-merror:0.198505\tvalidation_1-merror:0.196974\n",
            "[42]\tvalidation_0-merror:0.198419\tvalidation_1-merror:0.196966\n",
            "[43]\tvalidation_0-merror:0.19729\tvalidation_1-merror:0.195941\n",
            "[44]\tvalidation_0-merror:0.195568\tvalidation_1-merror:0.194314\n",
            "[45]\tvalidation_0-merror:0.192883\tvalidation_1-merror:0.19214\n",
            "[46]\tvalidation_0-merror:0.1919\tvalidation_1-merror:0.191096\n",
            "[47]\tvalidation_0-merror:0.188488\tvalidation_1-merror:0.187795\n",
            "[48]\tvalidation_0-merror:0.187883\tvalidation_1-merror:0.18723\n",
            "[49]\tvalidation_0-merror:0.186241\tvalidation_1-merror:0.185793\n",
            "[50]\tvalidation_0-merror:0.185893\tvalidation_1-merror:0.185506\n",
            "[51]\tvalidation_0-merror:0.184563\tvalidation_1-merror:0.184043\n",
            "[52]\tvalidation_0-merror:0.18065\tvalidation_1-merror:0.180964\n",
            "[53]\tvalidation_0-merror:0.179728\tvalidation_1-merror:0.179834\n",
            "[54]\tvalidation_0-merror:0.179722\tvalidation_1-merror:0.179784\n",
            "[55]\tvalidation_0-merror:0.179429\tvalidation_1-merror:0.179528\n",
            "[56]\tvalidation_0-merror:0.179222\tvalidation_1-merror:0.179248\n",
            "[57]\tvalidation_0-merror:0.17888\tvalidation_1-merror:0.178905\n",
            "[58]\tvalidation_0-merror:0.178978\tvalidation_1-merror:0.179256\n",
            "[59]\tvalidation_0-merror:0.17722\tvalidation_1-merror:0.176847\n",
            "[60]\tvalidation_0-merror:0.176487\tvalidation_1-merror:0.176245\n",
            "[61]\tvalidation_0-merror:0.176457\tvalidation_1-merror:0.175947\n",
            "[62]\tvalidation_0-merror:0.176396\tvalidation_1-merror:0.175879\n",
            "[63]\tvalidation_0-merror:0.173673\tvalidation_1-merror:0.173558\n",
            "[64]\tvalidation_0-merror:0.172806\tvalidation_1-merror:0.172705\n",
            "[65]\tvalidation_0-merror:0.168747\tvalidation_1-merror:0.16863\n",
            "[66]\tvalidation_0-merror:0.168485\tvalidation_1-merror:0.168303\n",
            "[67]\tvalidation_0-merror:0.168155\tvalidation_1-merror:0.167962\n",
            "[68]\tvalidation_0-merror:0.165646\tvalidation_1-merror:0.165103\n",
            "[69]\tvalidation_0-merror:0.165475\tvalidation_1-merror:0.164808\n",
            "[70]\tvalidation_0-merror:0.165121\tvalidation_1-merror:0.16463\n",
            "[71]\tvalidation_0-merror:0.164694\tvalidation_1-merror:0.164109\n",
            "[72]\tvalidation_0-merror:0.164773\tvalidation_1-merror:0.164198\n",
            "[73]\tvalidation_0-merror:0.164664\tvalidation_1-merror:0.163926\n",
            "[74]\tvalidation_0-merror:0.164187\tvalidation_1-merror:0.163565\n",
            "[75]\tvalidation_0-merror:0.163321\tvalidation_1-merror:0.162843\n",
            "[76]\tvalidation_0-merror:0.162936\tvalidation_1-merror:0.162524\n",
            "[77]\tvalidation_0-merror:0.162295\tvalidation_1-merror:0.161825\n",
            "[78]\tvalidation_0-merror:0.162167\tvalidation_1-merror:0.16165\n",
            "[79]\tvalidation_0-merror:0.161013\tvalidation_1-merror:0.160677\n",
            "[80]\tvalidation_0-merror:0.1587\tvalidation_1-merror:0.158294\n",
            "[81]\tvalidation_0-merror:0.158572\tvalidation_1-merror:0.158076\n",
            "[82]\tvalidation_0-merror:0.157888\tvalidation_1-merror:0.157307\n",
            "[83]\tvalidation_0-merror:0.156881\tvalidation_1-merror:0.156455\n",
            "[84]\tvalidation_0-merror:0.15682\tvalidation_1-merror:0.156162\n",
            "[85]\tvalidation_0-merror:0.15627\tvalidation_1-merror:0.155497\n",
            "[86]\tvalidation_0-merror:0.156222\tvalidation_1-merror:0.155508\n",
            "[87]\tvalidation_0-merror:0.155831\tvalidation_1-merror:0.155269\n",
            "[88]\tvalidation_0-merror:0.155507\tvalidation_1-merror:0.155052\n",
            "[89]\tvalidation_0-merror:0.154879\tvalidation_1-merror:0.154401\n",
            "[90]\tvalidation_0-merror:0.1545\tvalidation_1-merror:0.154113\n",
            "[91]\tvalidation_0-merror:0.15428\tvalidation_1-merror:0.153988\n",
            "[92]\tvalidation_0-merror:0.15298\tvalidation_1-merror:0.153017\n",
            "[93]\tvalidation_0-merror:0.151399\tvalidation_1-merror:0.151332\n",
            "[94]\tvalidation_0-merror:0.149275\tvalidation_1-merror:0.149172\n",
            "[95]\tvalidation_0-merror:0.14781\tvalidation_1-merror:0.147801\n",
            "[96]\tvalidation_0-merror:0.147169\tvalidation_1-merror:0.147011\n",
            "[97]\tvalidation_0-merror:0.147096\tvalidation_1-merror:0.147037\n",
            "[98]\tvalidation_0-merror:0.147615\tvalidation_1-merror:0.147604\n",
            "[99]\tvalidation_0-merror:0.146736\tvalidation_1-merror:0.146699\n",
            "[100]\tvalidation_0-merror:0.146589\tvalidation_1-merror:0.146647\n",
            "[101]\tvalidation_0-merror:0.146016\tvalidation_1-merror:0.146134\n",
            "[102]\tvalidation_0-merror:0.145832\tvalidation_1-merror:0.145894\n",
            "[103]\tvalidation_0-merror:0.145942\tvalidation_1-merror:0.145925\n",
            "[104]\tvalidation_0-merror:0.145875\tvalidation_1-merror:0.14587\n",
            "[105]\tvalidation_0-merror:0.145747\tvalidation_1-merror:0.14582\n",
            "[106]\tvalidation_0-merror:0.145692\tvalidation_1-merror:0.145742\n",
            "[107]\tvalidation_0-merror:0.145405\tvalidation_1-merror:0.145339\n",
            "[108]\tvalidation_0-merror:0.145112\tvalidation_1-merror:0.144881\n",
            "[109]\tvalidation_0-merror:0.144905\tvalidation_1-merror:0.144735\n",
            "[110]\tvalidation_0-merror:0.144862\tvalidation_1-merror:0.144884\n",
            "[111]\tvalidation_0-merror:0.144593\tvalidation_1-merror:0.144366\n",
            "[112]\tvalidation_0-merror:0.144502\tvalidation_1-merror:0.144559\n",
            "[113]\tvalidation_0-merror:0.144599\tvalidation_1-merror:0.144829\n",
            "[114]\tvalidation_0-merror:0.144508\tvalidation_1-merror:0.144376\n",
            "[115]\tvalidation_0-merror:0.144227\tvalidation_1-merror:0.143963\n",
            "[116]\tvalidation_0-merror:0.14394\tvalidation_1-merror:0.143762\n",
            "[117]\tvalidation_0-merror:0.143867\tvalidation_1-merror:0.143824\n",
            "[118]\tvalidation_0-merror:0.143684\tvalidation_1-merror:0.143516\n",
            "[119]\tvalidation_0-merror:0.143635\tvalidation_1-merror:0.14344\n",
            "[120]\tvalidation_0-merror:0.143611\tvalidation_1-merror:0.143398\n",
            "[121]\tvalidation_0-merror:0.143452\tvalidation_1-merror:0.143217\n",
            "[122]\tvalidation_0-merror:0.143397\tvalidation_1-merror:0.143147\n",
            "[123]\tvalidation_0-merror:0.142512\tvalidation_1-merror:0.142676\n",
            "[124]\tvalidation_0-merror:0.14184\tvalidation_1-merror:0.141917\n",
            "[125]\tvalidation_0-merror:0.141676\tvalidation_1-merror:0.141624\n",
            "[126]\tvalidation_0-merror:0.140876\tvalidation_1-merror:0.140787\n",
            "[127]\tvalidation_0-merror:0.140943\tvalidation_1-merror:0.140952\n",
            "[128]\tvalidation_0-merror:0.140534\tvalidation_1-merror:0.1404\n",
            "[129]\tvalidation_0-merror:0.140436\tvalidation_1-merror:0.140337\n",
            "[130]\tvalidation_0-merror:0.140424\tvalidation_1-merror:0.140157\n",
            "[131]\tvalidation_0-merror:0.140125\tvalidation_1-merror:0.139942\n",
            "[132]\tvalidation_0-merror:0.139576\tvalidation_1-merror:0.139233\n",
            "[133]\tvalidation_0-merror:0.139203\tvalidation_1-merror:0.13876\n",
            "[134]\tvalidation_0-merror:0.13899\tvalidation_1-merror:0.138597\n",
            "[135]\tvalidation_0-merror:0.138929\tvalidation_1-merror:0.13866\n",
            "[136]\tvalidation_0-merror:0.13888\tvalidation_1-merror:0.138386\n",
            "[137]\tvalidation_0-merror:0.138917\tvalidation_1-merror:0.13842\n",
            "[138]\tvalidation_0-merror:0.138697\tvalidation_1-merror:0.138226\n",
            "[139]\tvalidation_0-merror:0.138428\tvalidation_1-merror:0.137938\n",
            "[140]\tvalidation_0-merror:0.138465\tvalidation_1-merror:0.138069\n",
            "[141]\tvalidation_0-merror:0.138257\tvalidation_1-merror:0.13786\n",
            "[142]\tvalidation_0-merror:0.137824\tvalidation_1-merror:0.13719\n",
            "[143]\tvalidation_0-merror:0.137519\tvalidation_1-merror:0.136973\n",
            "[144]\tvalidation_0-merror:0.137653\tvalidation_1-merror:0.13701\n",
            "[145]\tvalidation_0-merror:0.137268\tvalidation_1-merror:0.136672\n",
            "[146]\tvalidation_0-merror:0.137299\tvalidation_1-merror:0.136766\n",
            "[147]\tvalidation_0-merror:0.136908\tvalidation_1-merror:0.136431\n",
            "[148]\tvalidation_0-merror:0.136798\tvalidation_1-merror:0.136261\n",
            "[149]\tvalidation_0-merror:0.136609\tvalidation_1-merror:0.136138\n",
            "[150]\tvalidation_0-merror:0.136426\tvalidation_1-merror:0.135961\n",
            "[151]\tvalidation_0-merror:0.136566\tvalidation_1-merror:0.136081\n",
            "[152]\tvalidation_0-merror:0.136377\tvalidation_1-merror:0.135874\n",
            "[153]\tvalidation_0-merror:0.136371\tvalidation_1-merror:0.135825\n",
            "[154]\tvalidation_0-merror:0.136005\tvalidation_1-merror:0.13573\n",
            "[155]\tvalidation_0-merror:0.136127\tvalidation_1-merror:0.135825\n",
            "[156]\tvalidation_0-merror:0.136426\tvalidation_1-merror:0.136112\n",
            "[157]\tvalidation_0-merror:0.13642\tvalidation_1-merror:0.136104\n",
            "[158]\tvalidation_0-merror:0.136249\tvalidation_1-merror:0.135911\n",
            "[159]\tvalidation_0-merror:0.136212\tvalidation_1-merror:0.135885\n",
            "[160]\tvalidation_0-merror:0.136023\tvalidation_1-merror:0.135759\n",
            "[161]\tvalidation_0-merror:0.136249\tvalidation_1-merror:0.135919\n",
            "[162]\tvalidation_0-merror:0.13523\tvalidation_1-merror:0.135042\n",
            "[163]\tvalidation_0-merror:0.135053\tvalidation_1-merror:0.134702\n",
            "[164]\tvalidation_0-merror:0.134882\tvalidation_1-merror:0.134516\n",
            "[165]\tvalidation_0-merror:0.134906\tvalidation_1-merror:0.134579\n",
            "[166]\tvalidation_0-merror:0.134534\tvalidation_1-merror:0.134278\n",
            "[167]\tvalidation_0-merror:0.134741\tvalidation_1-merror:0.134527\n",
            "[168]\tvalidation_0-merror:0.134668\tvalidation_1-merror:0.134498\n",
            "[169]\tvalidation_0-merror:0.134442\tvalidation_1-merror:0.134326\n",
            "[170]\tvalidation_0-merror:0.134515\tvalidation_1-merror:0.134281\n",
            "[171]\tvalidation_0-merror:0.134497\tvalidation_1-merror:0.134386\n",
            "[172]\tvalidation_0-merror:0.134174\tvalidation_1-merror:0.134051\n",
            "[173]\tvalidation_0-merror:0.134674\tvalidation_1-merror:0.134566\n",
            "[174]\tvalidation_0-merror:0.134668\tvalidation_1-merror:0.134566\n",
            "[175]\tvalidation_0-merror:0.13432\tvalidation_1-merror:0.134221\n",
            "[176]\tvalidation_0-merror:0.134302\tvalidation_1-merror:0.134229\n",
            "[177]\tvalidation_0-merror:0.134375\tvalidation_1-merror:0.134318\n",
            "[178]\tvalidation_0-merror:0.134363\tvalidation_1-merror:0.134305\n",
            "[179]\tvalidation_0-merror:0.134369\tvalidation_1-merror:0.13431\n",
            "[180]\tvalidation_0-merror:0.134094\tvalidation_1-merror:0.134046\n",
            "[181]\tvalidation_0-merror:0.134143\tvalidation_1-merror:0.13415\n",
            "[182]\tvalidation_0-merror:0.134003\tvalidation_1-merror:0.133978\n",
            "[183]\tvalidation_0-merror:0.134088\tvalidation_1-merror:0.133991\n",
            "[184]\tvalidation_0-merror:0.1341\tvalidation_1-merror:0.133988\n",
            "[185]\tvalidation_0-merror:0.133929\tvalidation_1-merror:0.133805\n",
            "[186]\tvalidation_0-merror:0.133838\tvalidation_1-merror:0.133768\n",
            "[187]\tvalidation_0-merror:0.133551\tvalidation_1-merror:0.133528\n",
            "[188]\tvalidation_0-merror:0.133545\tvalidation_1-merror:0.133478\n",
            "[189]\tvalidation_0-merror:0.133398\tvalidation_1-merror:0.133352\n",
            "[190]\tvalidation_0-merror:0.133435\tvalidation_1-merror:0.133394\n",
            "[191]\tvalidation_0-merror:0.133423\tvalidation_1-merror:0.133384\n",
            "[192]\tvalidation_0-merror:0.133112\tvalidation_1-merror:0.133054\n",
            "[193]\tvalidation_0-merror:0.132953\tvalidation_1-merror:0.13302\n",
            "[194]\tvalidation_0-merror:0.132855\tvalidation_1-merror:0.132939\n",
            "[195]\tvalidation_0-merror:0.133191\tvalidation_1-merror:0.133185\n",
            "[196]\tvalidation_0-merror:0.133063\tvalidation_1-merror:0.133083\n",
            "[197]\tvalidation_0-merror:0.133154\tvalidation_1-merror:0.133114\n",
            "[198]\tvalidation_0-merror:0.133166\tvalidation_1-merror:0.133135\n",
            "[199]\tvalidation_0-merror:0.133264\tvalidation_1-merror:0.133245\n",
            "[200]\tvalidation_0-merror:0.133136\tvalidation_1-merror:0.132955\n",
            "[201]\tvalidation_0-merror:0.133124\tvalidation_1-merror:0.132947\n",
            "[202]\tvalidation_0-merror:0.133032\tvalidation_1-merror:0.132923\n",
            "[203]\tvalidation_0-merror:0.133057\tvalidation_1-merror:0.132863\n",
            "[204]\tvalidation_0-merror:0.132849\tvalidation_1-merror:0.132821\n",
            "[205]\tvalidation_0-merror:0.132355\tvalidation_1-merror:0.132366\n",
            "[206]\tvalidation_0-merror:0.132361\tvalidation_1-merror:0.132348\n",
            "[207]\tvalidation_0-merror:0.132379\tvalidation_1-merror:0.132374\n",
            "[208]\tvalidation_0-merror:0.132403\tvalidation_1-merror:0.132395\n",
            "[209]\tvalidation_0-merror:0.132501\tvalidation_1-merror:0.132502\n",
            "[210]\tvalidation_0-merror:0.132495\tvalidation_1-merror:0.132536\n",
            "[211]\tvalidation_0-merror:0.132648\tvalidation_1-merror:0.13257\n",
            "[212]\tvalidation_0-merror:0.1323\tvalidation_1-merror:0.132374\n",
            "[213]\tvalidation_0-merror:0.132403\tvalidation_1-merror:0.132227\n",
            "[214]\tvalidation_0-merror:0.132367\tvalidation_1-merror:0.132209\n",
            "[215]\tvalidation_0-merror:0.132519\tvalidation_1-merror:0.132329\n",
            "[216]\tvalidation_0-merror:0.132507\tvalidation_1-merror:0.132309\n",
            "[217]\tvalidation_0-merror:0.132477\tvalidation_1-merror:0.132285\n",
            "[218]\tvalidation_0-merror:0.132471\tvalidation_1-merror:0.132288\n",
            "[219]\tvalidation_0-merror:0.132458\tvalidation_1-merror:0.132269\n",
            "[220]\tvalidation_0-merror:0.132465\tvalidation_1-merror:0.132256\n",
            "[221]\tvalidation_0-merror:0.132452\tvalidation_1-merror:0.132243\n",
            "[222]\tvalidation_0-merror:0.132434\tvalidation_1-merror:0.132217\n",
            "[223]\tvalidation_0-merror:0.132379\tvalidation_1-merror:0.132193\n",
            "[224]\tvalidation_0-merror:0.132172\tvalidation_1-merror:0.132036\n",
            "[225]\tvalidation_0-merror:0.132153\tvalidation_1-merror:0.132042\n",
            "[226]\tvalidation_0-merror:0.132129\tvalidation_1-merror:0.132034\n",
            "[227]\tvalidation_0-merror:0.13211\tvalidation_1-merror:0.132021\n",
            "[228]\tvalidation_0-merror:0.132001\tvalidation_1-merror:0.13183\n",
            "[229]\tvalidation_0-merror:0.131848\tvalidation_1-merror:0.131686\n",
            "[230]\tvalidation_0-merror:0.131494\tvalidation_1-merror:0.131291\n",
            "[231]\tvalidation_0-merror:0.131415\tvalidation_1-merror:0.131246\n",
            "[232]\tvalidation_0-merror:0.131604\tvalidation_1-merror:0.131427\n",
            "[233]\tvalidation_0-merror:0.13161\tvalidation_1-merror:0.131424\n",
            "[234]\tvalidation_0-merror:0.131604\tvalidation_1-merror:0.131419\n",
            "[235]\tvalidation_0-merror:0.13161\tvalidation_1-merror:0.131427\n",
            "[236]\tvalidation_0-merror:0.131604\tvalidation_1-merror:0.131414\n",
            "[237]\tvalidation_0-merror:0.131402\tvalidation_1-merror:0.13122\n",
            "[238]\tvalidation_0-merror:0.131463\tvalidation_1-merror:0.131233\n",
            "[239]\tvalidation_0-merror:0.131451\tvalidation_1-merror:0.131215\n",
            "[240]\tvalidation_0-merror:0.131451\tvalidation_1-merror:0.13121\n",
            "[241]\tvalidation_0-merror:0.131439\tvalidation_1-merror:0.131202\n",
            "[242]\tvalidation_0-merror:0.131634\tvalidation_1-merror:0.131273\n",
            "[243]\tvalidation_0-merror:0.131634\tvalidation_1-merror:0.13127\n",
            "[244]\tvalidation_0-merror:0.13164\tvalidation_1-merror:0.131296\n",
            "[245]\tvalidation_0-merror:0.131628\tvalidation_1-merror:0.131283\n",
            "[246]\tvalidation_0-merror:0.131622\tvalidation_1-merror:0.131273\n",
            "[247]\tvalidation_0-merror:0.131512\tvalidation_1-merror:0.131291\n",
            "[248]\tvalidation_0-merror:0.131317\tvalidation_1-merror:0.131014\n",
            "[249]\tvalidation_0-merror:0.131299\tvalidation_1-merror:0.131008\n",
            "[250]\tvalidation_0-merror:0.131268\tvalidation_1-merror:0.130974\n",
            "[251]\tvalidation_0-merror:0.131293\tvalidation_1-merror:0.130995\n",
            "[252]\tvalidation_0-merror:0.131286\tvalidation_1-merror:0.13098\n",
            "[253]\tvalidation_0-merror:0.131244\tvalidation_1-merror:0.130964\n",
            "[254]\tvalidation_0-merror:0.131238\tvalidation_1-merror:0.130948\n",
            "[255]\tvalidation_0-merror:0.131238\tvalidation_1-merror:0.130951\n",
            "[256]\tvalidation_0-merror:0.130731\tvalidation_1-merror:0.130514\n",
            "[257]\tvalidation_0-merror:0.13078\tvalidation_1-merror:0.13048\n",
            "[258]\tvalidation_0-merror:0.130743\tvalidation_1-merror:0.130362\n",
            "[259]\tvalidation_0-merror:0.13056\tvalidation_1-merror:0.130252\n",
            "[260]\tvalidation_0-merror:0.130584\tvalidation_1-merror:0.130231\n",
            "[261]\tvalidation_0-merror:0.130548\tvalidation_1-merror:0.130166\n",
            "[262]\tvalidation_0-merror:0.130548\tvalidation_1-merror:0.130171\n",
            "[263]\tvalidation_0-merror:0.130554\tvalidation_1-merror:0.130182\n",
            "[264]\tvalidation_0-merror:0.130548\tvalidation_1-merror:0.130174\n",
            "[265]\tvalidation_0-merror:0.13056\tvalidation_1-merror:0.130174\n",
            "[266]\tvalidation_0-merror:0.13056\tvalidation_1-merror:0.130166\n",
            "[267]\tvalidation_0-merror:0.130566\tvalidation_1-merror:0.130166\n",
            "[268]\tvalidation_0-merror:0.130548\tvalidation_1-merror:0.130158\n",
            "[269]\tvalidation_0-merror:0.130542\tvalidation_1-merror:0.130156\n",
            "[270]\tvalidation_0-merror:0.130426\tvalidation_1-merror:0.130108\n",
            "[271]\tvalidation_0-merror:0.130426\tvalidation_1-merror:0.130106\n",
            "[272]\tvalidation_0-merror:0.130426\tvalidation_1-merror:0.130106\n",
            "[273]\tvalidation_0-merror:0.130432\tvalidation_1-merror:0.130114\n",
            "[274]\tvalidation_0-merror:0.130432\tvalidation_1-merror:0.130122\n",
            "[275]\tvalidation_0-merror:0.130438\tvalidation_1-merror:0.130111\n",
            "[276]\tvalidation_0-merror:0.130456\tvalidation_1-merror:0.130124\n",
            "[277]\tvalidation_0-merror:0.130487\tvalidation_1-merror:0.130137\n",
            "[278]\tvalidation_0-merror:0.130481\tvalidation_1-merror:0.130129\n",
            "[279]\tvalidation_0-merror:0.130487\tvalidation_1-merror:0.130137\n",
            "[280]\tvalidation_0-merror:0.130481\tvalidation_1-merror:0.130132\n",
            "[281]\tvalidation_0-merror:0.130475\tvalidation_1-merror:0.130127\n",
            "[282]\tvalidation_0-merror:0.130475\tvalidation_1-merror:0.130127\n",
            "[283]\tvalidation_0-merror:0.130462\tvalidation_1-merror:0.130116\n",
            "[284]\tvalidation_0-merror:0.130468\tvalidation_1-merror:0.130111\n",
            "[285]\tvalidation_0-merror:0.130462\tvalidation_1-merror:0.130114\n",
            "[286]\tvalidation_0-merror:0.130462\tvalidation_1-merror:0.130111\n",
            "[287]\tvalidation_0-merror:0.13045\tvalidation_1-merror:0.130095\n",
            "[288]\tvalidation_0-merror:0.130468\tvalidation_1-merror:0.130101\n",
            "[289]\tvalidation_0-merror:0.13034\tvalidation_1-merror:0.130035\n",
            "[290]\tvalidation_0-merror:0.130322\tvalidation_1-merror:0.129996\n",
            "[291]\tvalidation_0-merror:0.130414\tvalidation_1-merror:0.130048\n",
            "[292]\tvalidation_0-merror:0.130377\tvalidation_1-merror:0.130033\n",
            "[293]\tvalidation_0-merror:0.130407\tvalidation_1-merror:0.130027\n",
            "[294]\tvalidation_0-merror:0.130401\tvalidation_1-merror:0.130012\n",
            "[295]\tvalidation_0-merror:0.130401\tvalidation_1-merror:0.130012\n",
            "[296]\tvalidation_0-merror:0.130395\tvalidation_1-merror:0.130012\n",
            "[297]\tvalidation_0-merror:0.130383\tvalidation_1-merror:0.130012\n",
            "[298]\tvalidation_0-merror:0.130383\tvalidation_1-merror:0.130012\n",
            "[299]\tvalidation_0-merror:0.130377\tvalidation_1-merror:0.130017\n",
            "[300]\tvalidation_0-merror:0.130371\tvalidation_1-merror:0.130012\n",
            "[301]\tvalidation_0-merror:0.130365\tvalidation_1-merror:0.130014\n",
            "[302]\tvalidation_0-merror:0.130383\tvalidation_1-merror:0.130014\n",
            "[303]\tvalidation_0-merror:0.130383\tvalidation_1-merror:0.130012\n",
            "[304]\tvalidation_0-merror:0.130383\tvalidation_1-merror:0.130014\n",
            "[305]\tvalidation_0-merror:0.130395\tvalidation_1-merror:0.13002\n",
            "[306]\tvalidation_0-merror:0.130395\tvalidation_1-merror:0.130022\n",
            "[307]\tvalidation_0-merror:0.130395\tvalidation_1-merror:0.130012\n",
            "[308]\tvalidation_0-merror:0.130389\tvalidation_1-merror:0.130009\n",
            "[309]\tvalidation_0-merror:0.130389\tvalidation_1-merror:0.130017\n",
            "[310]\tvalidation_0-merror:0.130401\tvalidation_1-merror:0.130022\n",
            "[311]\tvalidation_0-merror:0.130401\tvalidation_1-merror:0.130025\n",
            "[312]\tvalidation_0-merror:0.130395\tvalidation_1-merror:0.130014\n",
            "[313]\tvalidation_0-merror:0.130395\tvalidation_1-merror:0.13002\n",
            "[314]\tvalidation_0-merror:0.130383\tvalidation_1-merror:0.130014\n",
            "[315]\tvalidation_0-merror:0.130383\tvalidation_1-merror:0.130014\n",
            "[316]\tvalidation_0-merror:0.130371\tvalidation_1-merror:0.130012\n",
            "[317]\tvalidation_0-merror:0.130371\tvalidation_1-merror:0.130012\n",
            "[318]\tvalidation_0-merror:0.130377\tvalidation_1-merror:0.130006\n",
            "[319]\tvalidation_0-merror:0.130371\tvalidation_1-merror:0.130001\n",
            "[320]\tvalidation_0-merror:0.130365\tvalidation_1-merror:0.130001\n",
            "[321]\tvalidation_0-merror:0.130365\tvalidation_1-merror:0.130132\n",
            "[322]\tvalidation_0-merror:0.13006\tvalidation_1-merror:0.129771\n",
            "[323]\tvalidation_0-merror:0.130066\tvalidation_1-merror:0.12975\n",
            "[324]\tvalidation_0-merror:0.130182\tvalidation_1-merror:0.129813\n",
            "[325]\tvalidation_0-merror:0.130175\tvalidation_1-merror:0.12981\n",
            "[326]\tvalidation_0-merror:0.130175\tvalidation_1-merror:0.129808\n",
            "[327]\tvalidation_0-merror:0.130145\tvalidation_1-merror:0.129792\n",
            "[328]\tvalidation_0-merror:0.130145\tvalidation_1-merror:0.129787\n",
            "[329]\tvalidation_0-merror:0.130145\tvalidation_1-merror:0.129789\n",
            "[330]\tvalidation_0-merror:0.130151\tvalidation_1-merror:0.129779\n",
            "[331]\tvalidation_0-merror:0.130145\tvalidation_1-merror:0.129779\n",
            "[332]\tvalidation_0-merror:0.130157\tvalidation_1-merror:0.129787\n",
            "[333]\tvalidation_0-merror:0.130151\tvalidation_1-merror:0.129787\n",
            "[334]\tvalidation_0-merror:0.130169\tvalidation_1-merror:0.129805\n",
            "[335]\tvalidation_0-merror:0.130182\tvalidation_1-merror:0.12981\n",
            "[336]\tvalidation_0-merror:0.130182\tvalidation_1-merror:0.129813\n",
            "[337]\tvalidation_0-merror:0.130169\tvalidation_1-merror:0.129805\n",
            "[338]\tvalidation_0-merror:0.130169\tvalidation_1-merror:0.1298\n",
            "[339]\tvalidation_0-merror:0.130169\tvalidation_1-merror:0.129802\n",
            "[340]\tvalidation_0-merror:0.130163\tvalidation_1-merror:0.129802\n",
            "[341]\tvalidation_0-merror:0.130194\tvalidation_1-merror:0.129779\n",
            "[342]\tvalidation_0-merror:0.130182\tvalidation_1-merror:0.129776\n",
            "[343]\tvalidation_0-merror:0.130175\tvalidation_1-merror:0.129774\n",
            "[344]\tvalidation_0-merror:0.130175\tvalidation_1-merror:0.129774\n",
            "[345]\tvalidation_0-merror:0.130175\tvalidation_1-merror:0.129771\n",
            "[346]\tvalidation_0-merror:0.130175\tvalidation_1-merror:0.129771\n",
            "[347]\tvalidation_0-merror:0.130175\tvalidation_1-merror:0.129766\n",
            "[348]\tvalidation_0-merror:0.129974\tvalidation_1-merror:0.129695\n",
            "[349]\tvalidation_0-merror:0.129974\tvalidation_1-merror:0.129695\n",
            "[350]\tvalidation_0-merror:0.130096\tvalidation_1-merror:0.129742\n",
            "[351]\tvalidation_0-merror:0.130084\tvalidation_1-merror:0.129747\n",
            "[352]\tvalidation_0-merror:0.130084\tvalidation_1-merror:0.12974\n",
            "[353]\tvalidation_0-merror:0.130084\tvalidation_1-merror:0.12974\n",
            "[354]\tvalidation_0-merror:0.130084\tvalidation_1-merror:0.129742\n",
            "[355]\tvalidation_0-merror:0.130072\tvalidation_1-merror:0.129747\n",
            "[356]\tvalidation_0-merror:0.130072\tvalidation_1-merror:0.129747\n",
            "[357]\tvalidation_0-merror:0.130072\tvalidation_1-merror:0.129747\n",
            "[358]\tvalidation_0-merror:0.130072\tvalidation_1-merror:0.129745\n",
            "[359]\tvalidation_0-merror:0.130078\tvalidation_1-merror:0.129734\n",
            "[360]\tvalidation_0-merror:0.130084\tvalidation_1-merror:0.129734\n",
            "[361]\tvalidation_0-merror:0.130072\tvalidation_1-merror:0.129732\n",
            "[362]\tvalidation_0-merror:0.130066\tvalidation_1-merror:0.129727\n",
            "[363]\tvalidation_0-merror:0.13006\tvalidation_1-merror:0.129729\n",
            "[364]\tvalidation_0-merror:0.13006\tvalidation_1-merror:0.129727\n",
            "[365]\tvalidation_0-merror:0.130005\tvalidation_1-merror:0.12958\n",
            "[366]\tvalidation_0-merror:0.129998\tvalidation_1-merror:0.12958\n",
            "[367]\tvalidation_0-merror:0.130005\tvalidation_1-merror:0.129583\n",
            "[368]\tvalidation_0-merror:0.129992\tvalidation_1-merror:0.12958\n",
            "[369]\tvalidation_0-merror:0.129992\tvalidation_1-merror:0.129583\n",
            "[370]\tvalidation_0-merror:0.129992\tvalidation_1-merror:0.129585\n",
            "[371]\tvalidation_0-merror:0.129992\tvalidation_1-merror:0.129588\n",
            "[372]\tvalidation_0-merror:0.129992\tvalidation_1-merror:0.129593\n",
            "[373]\tvalidation_0-merror:0.129962\tvalidation_1-merror:0.129549\n",
            "[374]\tvalidation_0-merror:0.129962\tvalidation_1-merror:0.129546\n",
            "[375]\tvalidation_0-merror:0.129962\tvalidation_1-merror:0.129536\n",
            "[376]\tvalidation_0-merror:0.129962\tvalidation_1-merror:0.129533\n",
            "[377]\tvalidation_0-merror:0.129956\tvalidation_1-merror:0.129517\n",
            "[378]\tvalidation_0-merror:0.129962\tvalidation_1-merror:0.129517\n",
            "[379]\tvalidation_0-merror:0.129962\tvalidation_1-merror:0.129504\n",
            "[380]\tvalidation_0-merror:0.129962\tvalidation_1-merror:0.129509\n",
            "[381]\tvalidation_0-merror:0.129956\tvalidation_1-merror:0.129507\n",
            "[382]\tvalidation_0-merror:0.129962\tvalidation_1-merror:0.129507\n",
            "[383]\tvalidation_0-merror:0.12995\tvalidation_1-merror:0.129494\n",
            "[384]\tvalidation_0-merror:0.129944\tvalidation_1-merror:0.129486\n",
            "[385]\tvalidation_0-merror:0.129944\tvalidation_1-merror:0.129488\n",
            "[386]\tvalidation_0-merror:0.129937\tvalidation_1-merror:0.129481\n",
            "[387]\tvalidation_0-merror:0.129937\tvalidation_1-merror:0.129483\n",
            "[388]\tvalidation_0-merror:0.129944\tvalidation_1-merror:0.129491\n",
            "[389]\tvalidation_0-merror:0.129944\tvalidation_1-merror:0.129488\n",
            "[390]\tvalidation_0-merror:0.129944\tvalidation_1-merror:0.129494\n",
            "[391]\tvalidation_0-merror:0.12995\tvalidation_1-merror:0.129496\n",
            "[392]\tvalidation_0-merror:0.12995\tvalidation_1-merror:0.129502\n",
            "[393]\tvalidation_0-merror:0.12995\tvalidation_1-merror:0.129504\n",
            "[394]\tvalidation_0-merror:0.129937\tvalidation_1-merror:0.129499\n",
            "[395]\tvalidation_0-merror:0.129937\tvalidation_1-merror:0.129499\n",
            "[396]\tvalidation_0-merror:0.129937\tvalidation_1-merror:0.129499\n",
            "[397]\tvalidation_0-merror:0.129931\tvalidation_1-merror:0.129494\n",
            "[398]\tvalidation_0-merror:0.129931\tvalidation_1-merror:0.129494\n",
            "[399]\tvalidation_0-merror:0.129925\tvalidation_1-merror:0.129488\n",
            "[400]\tvalidation_0-merror:0.129925\tvalidation_1-merror:0.129488\n",
            "[401]\tvalidation_0-merror:0.129925\tvalidation_1-merror:0.129486\n",
            "[402]\tvalidation_0-merror:0.129925\tvalidation_1-merror:0.129491\n",
            "[403]\tvalidation_0-merror:0.129925\tvalidation_1-merror:0.129491\n",
            "[404]\tvalidation_0-merror:0.129925\tvalidation_1-merror:0.129488\n",
            "[405]\tvalidation_0-merror:0.129925\tvalidation_1-merror:0.129488\n",
            "[406]\tvalidation_0-merror:0.129925\tvalidation_1-merror:0.129483\n",
            "[407]\tvalidation_0-merror:0.129925\tvalidation_1-merror:0.129483\n",
            "[408]\tvalidation_0-merror:0.129913\tvalidation_1-merror:0.129468\n",
            "[409]\tvalidation_0-merror:0.129901\tvalidation_1-merror:0.12946\n",
            "[410]\tvalidation_0-merror:0.129907\tvalidation_1-merror:0.129465\n",
            "[411]\tvalidation_0-merror:0.129907\tvalidation_1-merror:0.129457\n",
            "[412]\tvalidation_0-merror:0.129913\tvalidation_1-merror:0.129457\n",
            "[413]\tvalidation_0-merror:0.129913\tvalidation_1-merror:0.129457\n",
            "[414]\tvalidation_0-merror:0.129913\tvalidation_1-merror:0.129457\n",
            "[415]\tvalidation_0-merror:0.129913\tvalidation_1-merror:0.129457\n",
            "[416]\tvalidation_0-merror:0.129913\tvalidation_1-merror:0.129454\n",
            "[417]\tvalidation_0-merror:0.129913\tvalidation_1-merror:0.129454\n",
            "[418]\tvalidation_0-merror:0.129907\tvalidation_1-merror:0.129454\n",
            "[419]\tvalidation_0-merror:0.129846\tvalidation_1-merror:0.12942\n",
            "[420]\tvalidation_0-merror:0.129846\tvalidation_1-merror:0.12942\n",
            "[421]\tvalidation_0-merror:0.12984\tvalidation_1-merror:0.12941\n",
            "[422]\tvalidation_0-merror:0.12984\tvalidation_1-merror:0.12941\n",
            "[423]\tvalidation_0-merror:0.12984\tvalidation_1-merror:0.12941\n",
            "[424]\tvalidation_0-merror:0.12984\tvalidation_1-merror:0.129413\n",
            "[425]\tvalidation_0-merror:0.12984\tvalidation_1-merror:0.12941\n",
            "[426]\tvalidation_0-merror:0.12984\tvalidation_1-merror:0.12941\n",
            "[427]\tvalidation_0-merror:0.129828\tvalidation_1-merror:0.12941\n",
            "[428]\tvalidation_0-merror:0.129828\tvalidation_1-merror:0.12941\n",
            "[429]\tvalidation_0-merror:0.129828\tvalidation_1-merror:0.12941\n",
            "[430]\tvalidation_0-merror:0.129821\tvalidation_1-merror:0.129413\n",
            "[431]\tvalidation_0-merror:0.129821\tvalidation_1-merror:0.129405\n",
            "[432]\tvalidation_0-merror:0.129821\tvalidation_1-merror:0.129405\n",
            "[433]\tvalidation_0-merror:0.129815\tvalidation_1-merror:0.129405\n",
            "[434]\tvalidation_0-merror:0.129821\tvalidation_1-merror:0.129394\n",
            "[435]\tvalidation_0-merror:0.129821\tvalidation_1-merror:0.129394\n",
            "[436]\tvalidation_0-merror:0.129821\tvalidation_1-merror:0.129394\n",
            "[437]\tvalidation_0-merror:0.129815\tvalidation_1-merror:0.129394\n",
            "[438]\tvalidation_0-merror:0.129815\tvalidation_1-merror:0.129394\n",
            "[439]\tvalidation_0-merror:0.129815\tvalidation_1-merror:0.129392\n",
            "[440]\tvalidation_0-merror:0.129815\tvalidation_1-merror:0.129389\n",
            "[441]\tvalidation_0-merror:0.129815\tvalidation_1-merror:0.129389\n",
            "[442]\tvalidation_0-merror:0.129815\tvalidation_1-merror:0.129389\n",
            "[443]\tvalidation_0-merror:0.129815\tvalidation_1-merror:0.129384\n",
            "[444]\tvalidation_0-merror:0.129815\tvalidation_1-merror:0.129381\n",
            "[445]\tvalidation_0-merror:0.129809\tvalidation_1-merror:0.129371\n",
            "[446]\tvalidation_0-merror:0.129809\tvalidation_1-merror:0.129368\n",
            "[447]\tvalidation_0-merror:0.129809\tvalidation_1-merror:0.129368\n",
            "[448]\tvalidation_0-merror:0.129809\tvalidation_1-merror:0.129368\n",
            "[449]\tvalidation_0-merror:0.129809\tvalidation_1-merror:0.129368\n",
            "[450]\tvalidation_0-merror:0.129809\tvalidation_1-merror:0.129379\n",
            "[451]\tvalidation_0-merror:0.129809\tvalidation_1-merror:0.129379\n",
            "[452]\tvalidation_0-merror:0.129809\tvalidation_1-merror:0.129379\n",
            "[453]\tvalidation_0-merror:0.129809\tvalidation_1-merror:0.129379\n",
            "[454]\tvalidation_0-merror:0.129809\tvalidation_1-merror:0.129368\n",
            "[455]\tvalidation_0-merror:0.129809\tvalidation_1-merror:0.129368\n",
            "[456]\tvalidation_0-merror:0.129809\tvalidation_1-merror:0.129368\n",
            "[457]\tvalidation_0-merror:0.129809\tvalidation_1-merror:0.129368\n",
            "[458]\tvalidation_0-merror:0.129809\tvalidation_1-merror:0.129368\n",
            "[459]\tvalidation_0-merror:0.129809\tvalidation_1-merror:0.129368\n",
            "[460]\tvalidation_0-merror:0.129809\tvalidation_1-merror:0.129368\n",
            "[461]\tvalidation_0-merror:0.129809\tvalidation_1-merror:0.129368\n",
            "[462]\tvalidation_0-merror:0.129803\tvalidation_1-merror:0.129366\n",
            "[463]\tvalidation_0-merror:0.129803\tvalidation_1-merror:0.129366\n",
            "[464]\tvalidation_0-merror:0.129803\tvalidation_1-merror:0.129366\n",
            "[465]\tvalidation_0-merror:0.129803\tvalidation_1-merror:0.129366\n",
            "[466]\tvalidation_0-merror:0.129803\tvalidation_1-merror:0.129366\n",
            "[467]\tvalidation_0-merror:0.129803\tvalidation_1-merror:0.129366\n",
            "[468]\tvalidation_0-merror:0.129797\tvalidation_1-merror:0.129358\n",
            "[469]\tvalidation_0-merror:0.129797\tvalidation_1-merror:0.129358\n",
            "[470]\tvalidation_0-merror:0.129797\tvalidation_1-merror:0.129355\n",
            "[471]\tvalidation_0-merror:0.129797\tvalidation_1-merror:0.129355\n",
            "[472]\tvalidation_0-merror:0.129797\tvalidation_1-merror:0.129355\n",
            "[473]\tvalidation_0-merror:0.129803\tvalidation_1-merror:0.129363\n",
            "[474]\tvalidation_0-merror:0.129797\tvalidation_1-merror:0.129358\n",
            "[475]\tvalidation_0-merror:0.129797\tvalidation_1-merror:0.129358\n",
            "[476]\tvalidation_0-merror:0.129797\tvalidation_1-merror:0.129358\n",
            "[477]\tvalidation_0-merror:0.129797\tvalidation_1-merror:0.129358\n",
            "[478]\tvalidation_0-merror:0.129797\tvalidation_1-merror:0.129358\n",
            "[479]\tvalidation_0-merror:0.129669\tvalidation_1-merror:0.12924\n",
            "[480]\tvalidation_0-merror:0.129644\tvalidation_1-merror:0.129219\n",
            "[481]\tvalidation_0-merror:0.129644\tvalidation_1-merror:0.129219\n",
            "[482]\tvalidation_0-merror:0.129651\tvalidation_1-merror:0.129235\n",
            "[483]\tvalidation_0-merror:0.129657\tvalidation_1-merror:0.129235\n",
            "[484]\tvalidation_0-merror:0.129644\tvalidation_1-merror:0.129211\n",
            "[485]\tvalidation_0-merror:0.129644\tvalidation_1-merror:0.129211\n",
            "[486]\tvalidation_0-merror:0.129638\tvalidation_1-merror:0.129209\n",
            "[487]\tvalidation_0-merror:0.129638\tvalidation_1-merror:0.129211\n",
            "[488]\tvalidation_0-merror:0.129638\tvalidation_1-merror:0.129209\n",
            "[489]\tvalidation_0-merror:0.129638\tvalidation_1-merror:0.129209\n",
            "[490]\tvalidation_0-merror:0.129638\tvalidation_1-merror:0.129209\n",
            "[491]\tvalidation_0-merror:0.129638\tvalidation_1-merror:0.129209\n",
            "[492]\tvalidation_0-merror:0.129638\tvalidation_1-merror:0.129209\n",
            "[493]\tvalidation_0-merror:0.129638\tvalidation_1-merror:0.129209\n",
            "[494]\tvalidation_0-merror:0.129632\tvalidation_1-merror:0.129209\n",
            "[495]\tvalidation_0-merror:0.129632\tvalidation_1-merror:0.129209\n",
            "[496]\tvalidation_0-merror:0.129602\tvalidation_1-merror:0.129161\n",
            "[497]\tvalidation_0-merror:0.129602\tvalidation_1-merror:0.129161\n",
            "[498]\tvalidation_0-merror:0.129602\tvalidation_1-merror:0.129159\n",
            "[499]\tvalidation_0-merror:0.129602\tvalidation_1-merror:0.129159\n",
            "[500]\tvalidation_0-merror:0.129602\tvalidation_1-merror:0.129159\n",
            "[501]\tvalidation_0-merror:0.129596\tvalidation_1-merror:0.129154\n",
            "[502]\tvalidation_0-merror:0.129596\tvalidation_1-merror:0.129154\n",
            "[503]\tvalidation_0-merror:0.129596\tvalidation_1-merror:0.129156\n",
            "[504]\tvalidation_0-merror:0.129596\tvalidation_1-merror:0.129156\n",
            "[505]\tvalidation_0-merror:0.129596\tvalidation_1-merror:0.129156\n",
            "[506]\tvalidation_0-merror:0.129596\tvalidation_1-merror:0.129156\n",
            "[507]\tvalidation_0-merror:0.129596\tvalidation_1-merror:0.129156\n",
            "[508]\tvalidation_0-merror:0.129596\tvalidation_1-merror:0.129156\n",
            "[509]\tvalidation_0-merror:0.129596\tvalidation_1-merror:0.129159\n",
            "[510]\tvalidation_0-merror:0.129596\tvalidation_1-merror:0.129159\n",
            "[511]\tvalidation_0-merror:0.129589\tvalidation_1-merror:0.129154\n",
            "[512]\tvalidation_0-merror:0.129589\tvalidation_1-merror:0.129154\n",
            "[513]\tvalidation_0-merror:0.129589\tvalidation_1-merror:0.129154\n",
            "[514]\tvalidation_0-merror:0.129589\tvalidation_1-merror:0.129154\n",
            "[515]\tvalidation_0-merror:0.129589\tvalidation_1-merror:0.129154\n",
            "[516]\tvalidation_0-merror:0.129589\tvalidation_1-merror:0.129154\n",
            "[517]\tvalidation_0-merror:0.129589\tvalidation_1-merror:0.129154\n",
            "[518]\tvalidation_0-merror:0.129589\tvalidation_1-merror:0.129154\n",
            "[519]\tvalidation_0-merror:0.129589\tvalidation_1-merror:0.129151\n",
            "[520]\tvalidation_0-merror:0.129583\tvalidation_1-merror:0.129151\n",
            "[521]\tvalidation_0-merror:0.129583\tvalidation_1-merror:0.129154\n",
            "[522]\tvalidation_0-merror:0.129583\tvalidation_1-merror:0.129151\n",
            "[523]\tvalidation_0-merror:0.129583\tvalidation_1-merror:0.129151\n",
            "[524]\tvalidation_0-merror:0.129577\tvalidation_1-merror:0.129146\n",
            "[525]\tvalidation_0-merror:0.129577\tvalidation_1-merror:0.129146\n",
            "[526]\tvalidation_0-merror:0.129577\tvalidation_1-merror:0.129148\n",
            "[527]\tvalidation_0-merror:0.129583\tvalidation_1-merror:0.129151\n",
            "[528]\tvalidation_0-merror:0.129571\tvalidation_1-merror:0.129148\n",
            "[529]\tvalidation_0-merror:0.129571\tvalidation_1-merror:0.129146\n",
            "[530]\tvalidation_0-merror:0.129547\tvalidation_1-merror:0.129135\n",
            "[531]\tvalidation_0-merror:0.129553\tvalidation_1-merror:0.129138\n",
            "[532]\tvalidation_0-merror:0.129553\tvalidation_1-merror:0.129138\n",
            "[533]\tvalidation_0-merror:0.129553\tvalidation_1-merror:0.129138\n",
            "[534]\tvalidation_0-merror:0.129553\tvalidation_1-merror:0.129138\n",
            "[535]\tvalidation_0-merror:0.129565\tvalidation_1-merror:0.129138\n",
            "[536]\tvalidation_0-merror:0.129565\tvalidation_1-merror:0.129138\n",
            "[537]\tvalidation_0-merror:0.129565\tvalidation_1-merror:0.129138\n",
            "[538]\tvalidation_0-merror:0.129565\tvalidation_1-merror:0.129146\n",
            "[539]\tvalidation_0-merror:0.129565\tvalidation_1-merror:0.129146\n",
            "[540]\tvalidation_0-merror:0.129535\tvalidation_1-merror:0.129127\n",
            "[541]\tvalidation_0-merror:0.129492\tvalidation_1-merror:0.129112\n",
            "[542]\tvalidation_0-merror:0.129498\tvalidation_1-merror:0.129112\n",
            "[543]\tvalidation_0-merror:0.129498\tvalidation_1-merror:0.129112\n",
            "[544]\tvalidation_0-merror:0.129522\tvalidation_1-merror:0.129114\n",
            "[545]\tvalidation_0-merror:0.129516\tvalidation_1-merror:0.129114\n",
            "[546]\tvalidation_0-merror:0.129516\tvalidation_1-merror:0.129114\n",
            "[547]\tvalidation_0-merror:0.129522\tvalidation_1-merror:0.129125\n",
            "[548]\tvalidation_0-merror:0.129516\tvalidation_1-merror:0.129125\n",
            "[549]\tvalidation_0-merror:0.129516\tvalidation_1-merror:0.129125\n",
            "[550]\tvalidation_0-merror:0.129516\tvalidation_1-merror:0.129125\n",
            "[551]\tvalidation_0-merror:0.129486\tvalidation_1-merror:0.129112\n",
            "[552]\tvalidation_0-merror:0.12951\tvalidation_1-merror:0.129114\n",
            "[553]\tvalidation_0-merror:0.12951\tvalidation_1-merror:0.129112\n",
            "[554]\tvalidation_0-merror:0.129498\tvalidation_1-merror:0.129112\n",
            "[555]\tvalidation_0-merror:0.129498\tvalidation_1-merror:0.129109\n",
            "[556]\tvalidation_0-merror:0.129498\tvalidation_1-merror:0.129109\n",
            "[557]\tvalidation_0-merror:0.129467\tvalidation_1-merror:0.129107\n",
            "[558]\tvalidation_0-merror:0.129467\tvalidation_1-merror:0.129104\n",
            "[559]\tvalidation_0-merror:0.129467\tvalidation_1-merror:0.129104\n",
            "[560]\tvalidation_0-merror:0.129467\tvalidation_1-merror:0.129104\n",
            "[561]\tvalidation_0-merror:0.129467\tvalidation_1-merror:0.129104\n",
            "[562]\tvalidation_0-merror:0.129467\tvalidation_1-merror:0.129107\n",
            "[563]\tvalidation_0-merror:0.129467\tvalidation_1-merror:0.129107\n",
            "[564]\tvalidation_0-merror:0.129467\tvalidation_1-merror:0.129107\n",
            "[565]\tvalidation_0-merror:0.129461\tvalidation_1-merror:0.129107\n",
            "[566]\tvalidation_0-merror:0.129467\tvalidation_1-merror:0.129107\n",
            "[567]\tvalidation_0-merror:0.129467\tvalidation_1-merror:0.129107\n",
            "[568]\tvalidation_0-merror:0.129467\tvalidation_1-merror:0.129104\n",
            "[569]\tvalidation_0-merror:0.129467\tvalidation_1-merror:0.129101\n",
            "[570]\tvalidation_0-merror:0.129467\tvalidation_1-merror:0.129101\n",
            "[571]\tvalidation_0-merror:0.129467\tvalidation_1-merror:0.129101\n",
            "[572]\tvalidation_0-merror:0.129467\tvalidation_1-merror:0.129101\n",
            "[573]\tvalidation_0-merror:0.129467\tvalidation_1-merror:0.129104\n",
            "[574]\tvalidation_0-merror:0.129467\tvalidation_1-merror:0.129107\n",
            "[575]\tvalidation_0-merror:0.129467\tvalidation_1-merror:0.129107\n",
            "[576]\tvalidation_0-merror:0.129449\tvalidation_1-merror:0.129104\n",
            "[577]\tvalidation_0-merror:0.129449\tvalidation_1-merror:0.129104\n",
            "[578]\tvalidation_0-merror:0.129449\tvalidation_1-merror:0.129101\n",
            "[579]\tvalidation_0-merror:0.129449\tvalidation_1-merror:0.129101\n",
            "[580]\tvalidation_0-merror:0.129443\tvalidation_1-merror:0.129093\n",
            "[581]\tvalidation_0-merror:0.129443\tvalidation_1-merror:0.129093\n",
            "[582]\tvalidation_0-merror:0.129443\tvalidation_1-merror:0.129093\n",
            "[583]\tvalidation_0-merror:0.129455\tvalidation_1-merror:0.129096\n",
            "[584]\tvalidation_0-merror:0.129425\tvalidation_1-merror:0.129091\n",
            "[585]\tvalidation_0-merror:0.129437\tvalidation_1-merror:0.129093\n",
            "[586]\tvalidation_0-merror:0.129437\tvalidation_1-merror:0.129091\n",
            "[587]\tvalidation_0-merror:0.129437\tvalidation_1-merror:0.129091\n",
            "[588]\tvalidation_0-merror:0.129425\tvalidation_1-merror:0.129088\n",
            "[589]\tvalidation_0-merror:0.129406\tvalidation_1-merror:0.129067\n",
            "[590]\tvalidation_0-merror:0.129419\tvalidation_1-merror:0.129067\n",
            "[591]\tvalidation_0-merror:0.129419\tvalidation_1-merror:0.129067\n",
            "[592]\tvalidation_0-merror:0.129419\tvalidation_1-merror:0.129067\n",
            "[593]\tvalidation_0-merror:0.129406\tvalidation_1-merror:0.129067\n",
            "[594]\tvalidation_0-merror:0.129406\tvalidation_1-merror:0.129067\n",
            "[595]\tvalidation_0-merror:0.129406\tvalidation_1-merror:0.129067\n",
            "[596]\tvalidation_0-merror:0.129406\tvalidation_1-merror:0.129067\n",
            "[597]\tvalidation_0-merror:0.129406\tvalidation_1-merror:0.129057\n",
            "[598]\tvalidation_0-merror:0.129406\tvalidation_1-merror:0.129057\n",
            "[599]\tvalidation_0-merror:0.129406\tvalidation_1-merror:0.129057\n",
            "[600]\tvalidation_0-merror:0.129419\tvalidation_1-merror:0.129057\n",
            "[601]\tvalidation_0-merror:0.129406\tvalidation_1-merror:0.129046\n",
            "[602]\tvalidation_0-merror:0.129419\tvalidation_1-merror:0.129046\n",
            "[603]\tvalidation_0-merror:0.1294\tvalidation_1-merror:0.129033\n",
            "[604]\tvalidation_0-merror:0.1294\tvalidation_1-merror:0.129033\n",
            "[605]\tvalidation_0-merror:0.129406\tvalidation_1-merror:0.129046\n",
            "[606]\tvalidation_0-merror:0.129406\tvalidation_1-merror:0.129046\n",
            "[607]\tvalidation_0-merror:0.129406\tvalidation_1-merror:0.129046\n",
            "[608]\tvalidation_0-merror:0.129406\tvalidation_1-merror:0.129046\n",
            "[609]\tvalidation_0-merror:0.129406\tvalidation_1-merror:0.129046\n",
            "[610]\tvalidation_0-merror:0.129406\tvalidation_1-merror:0.129044\n",
            "[611]\tvalidation_0-merror:0.129388\tvalidation_1-merror:0.129036\n",
            "[612]\tvalidation_0-merror:0.129388\tvalidation_1-merror:0.129036\n",
            "[613]\tvalidation_0-merror:0.129382\tvalidation_1-merror:0.129023\n",
            "[614]\tvalidation_0-merror:0.129382\tvalidation_1-merror:0.129023\n",
            "[615]\tvalidation_0-merror:0.129382\tvalidation_1-merror:0.129023\n",
            "[616]\tvalidation_0-merror:0.129382\tvalidation_1-merror:0.129023\n",
            "[617]\tvalidation_0-merror:0.129376\tvalidation_1-merror:0.129023\n",
            "[618]\tvalidation_0-merror:0.129339\tvalidation_1-merror:0.128952\n",
            "[619]\tvalidation_0-merror:0.129339\tvalidation_1-merror:0.128952\n",
            "[620]\tvalidation_0-merror:0.129339\tvalidation_1-merror:0.128952\n",
            "[621]\tvalidation_0-merror:0.129376\tvalidation_1-merror:0.12902\n",
            "[622]\tvalidation_0-merror:0.129339\tvalidation_1-merror:0.128952\n",
            "[623]\tvalidation_0-merror:0.129333\tvalidation_1-merror:0.12895\n",
            "[624]\tvalidation_0-merror:0.129333\tvalidation_1-merror:0.12895\n",
            "[625]\tvalidation_0-merror:0.129333\tvalidation_1-merror:0.12895\n",
            "[626]\tvalidation_0-merror:0.129333\tvalidation_1-merror:0.12895\n",
            "[627]\tvalidation_0-merror:0.129333\tvalidation_1-merror:0.12895\n",
            "[628]\tvalidation_0-merror:0.129333\tvalidation_1-merror:0.12895\n",
            "[629]\tvalidation_0-merror:0.129333\tvalidation_1-merror:0.12895\n",
            "[630]\tvalidation_0-merror:0.129333\tvalidation_1-merror:0.12895\n",
            "[631]\tvalidation_0-merror:0.129333\tvalidation_1-merror:0.12895\n",
            "[632]\tvalidation_0-merror:0.129333\tvalidation_1-merror:0.12895\n",
            "[633]\tvalidation_0-merror:0.129321\tvalidation_1-merror:0.128939\n",
            "[634]\tvalidation_0-merror:0.129321\tvalidation_1-merror:0.128939\n",
            "[635]\tvalidation_0-merror:0.129321\tvalidation_1-merror:0.128939\n",
            "[636]\tvalidation_0-merror:0.129321\tvalidation_1-merror:0.128939\n",
            "[637]\tvalidation_0-merror:0.129321\tvalidation_1-merror:0.128939\n",
            "[638]\tvalidation_0-merror:0.129321\tvalidation_1-merror:0.128939\n",
            "[639]\tvalidation_0-merror:0.129321\tvalidation_1-merror:0.128939\n",
            "[640]\tvalidation_0-merror:0.129321\tvalidation_1-merror:0.128939\n",
            "[641]\tvalidation_0-merror:0.129321\tvalidation_1-merror:0.128936\n",
            "[642]\tvalidation_0-merror:0.129321\tvalidation_1-merror:0.128936\n",
            "[643]\tvalidation_0-merror:0.129321\tvalidation_1-merror:0.128936\n",
            "[644]\tvalidation_0-merror:0.129321\tvalidation_1-merror:0.128936\n",
            "[645]\tvalidation_0-merror:0.129321\tvalidation_1-merror:0.128936\n",
            "[646]\tvalidation_0-merror:0.129309\tvalidation_1-merror:0.128936\n",
            "[647]\tvalidation_0-merror:0.129309\tvalidation_1-merror:0.128936\n",
            "[648]\tvalidation_0-merror:0.129309\tvalidation_1-merror:0.128936\n",
            "[649]\tvalidation_0-merror:0.129309\tvalidation_1-merror:0.128936\n",
            "[650]\tvalidation_0-merror:0.129309\tvalidation_1-merror:0.128936\n",
            "[651]\tvalidation_0-merror:0.129309\tvalidation_1-merror:0.128936\n",
            "[652]\tvalidation_0-merror:0.129309\tvalidation_1-merror:0.128936\n",
            "[653]\tvalidation_0-merror:0.129309\tvalidation_1-merror:0.128936\n",
            "[654]\tvalidation_0-merror:0.129309\tvalidation_1-merror:0.128936\n",
            "[655]\tvalidation_0-merror:0.129309\tvalidation_1-merror:0.128936\n",
            "[656]\tvalidation_0-merror:0.129309\tvalidation_1-merror:0.128936\n",
            "[657]\tvalidation_0-merror:0.129309\tvalidation_1-merror:0.128934\n",
            "[658]\tvalidation_0-merror:0.129309\tvalidation_1-merror:0.128934\n",
            "[659]\tvalidation_0-merror:0.129309\tvalidation_1-merror:0.128934\n",
            "[660]\tvalidation_0-merror:0.129309\tvalidation_1-merror:0.128934\n",
            "[661]\tvalidation_0-merror:0.129309\tvalidation_1-merror:0.128934\n",
            "[662]\tvalidation_0-merror:0.129309\tvalidation_1-merror:0.128934\n",
            "[663]\tvalidation_0-merror:0.129309\tvalidation_1-merror:0.128934\n",
            "[664]\tvalidation_0-merror:0.129309\tvalidation_1-merror:0.128934\n",
            "[665]\tvalidation_0-merror:0.129309\tvalidation_1-merror:0.128934\n",
            "[666]\tvalidation_0-merror:0.129309\tvalidation_1-merror:0.128934\n",
            "[667]\tvalidation_0-merror:0.129309\tvalidation_1-merror:0.128934\n",
            "[668]\tvalidation_0-merror:0.129303\tvalidation_1-merror:0.128929\n",
            "[669]\tvalidation_0-merror:0.129303\tvalidation_1-merror:0.128929\n",
            "[670]\tvalidation_0-merror:0.129303\tvalidation_1-merror:0.128929\n",
            "[671]\tvalidation_0-merror:0.129303\tvalidation_1-merror:0.128929\n",
            "[672]\tvalidation_0-merror:0.129303\tvalidation_1-merror:0.128929\n",
            "[673]\tvalidation_0-merror:0.129303\tvalidation_1-merror:0.128929\n",
            "[674]\tvalidation_0-merror:0.129303\tvalidation_1-merror:0.128929\n",
            "[675]\tvalidation_0-merror:0.129303\tvalidation_1-merror:0.128929\n",
            "[676]\tvalidation_0-merror:0.129303\tvalidation_1-merror:0.128929\n",
            "[677]\tvalidation_0-merror:0.129303\tvalidation_1-merror:0.128929\n",
            "[678]\tvalidation_0-merror:0.129303\tvalidation_1-merror:0.128929\n",
            "[679]\tvalidation_0-merror:0.129303\tvalidation_1-merror:0.128929\n",
            "[680]\tvalidation_0-merror:0.129303\tvalidation_1-merror:0.128929\n",
            "[681]\tvalidation_0-merror:0.129303\tvalidation_1-merror:0.128929\n",
            "[682]\tvalidation_0-merror:0.129303\tvalidation_1-merror:0.128929\n",
            "[683]\tvalidation_0-merror:0.129303\tvalidation_1-merror:0.128929\n",
            "[684]\tvalidation_0-merror:0.129303\tvalidation_1-merror:0.128929\n",
            "[685]\tvalidation_0-merror:0.129303\tvalidation_1-merror:0.128929\n",
            "[686]\tvalidation_0-merror:0.129303\tvalidation_1-merror:0.128929\n",
            "[687]\tvalidation_0-merror:0.129303\tvalidation_1-merror:0.128929\n",
            "[688]\tvalidation_0-merror:0.129303\tvalidation_1-merror:0.128929\n",
            "[689]\tvalidation_0-merror:0.129303\tvalidation_1-merror:0.128929\n",
            "[690]\tvalidation_0-merror:0.129303\tvalidation_1-merror:0.128929\n",
            "[691]\tvalidation_0-merror:0.129303\tvalidation_1-merror:0.128929\n",
            "[692]\tvalidation_0-merror:0.129303\tvalidation_1-merror:0.128929\n",
            "[693]\tvalidation_0-merror:0.129303\tvalidation_1-merror:0.128929\n",
            "[694]\tvalidation_0-merror:0.129303\tvalidation_1-merror:0.128929\n",
            "[695]\tvalidation_0-merror:0.12929\tvalidation_1-merror:0.128929\n",
            "[696]\tvalidation_0-merror:0.12929\tvalidation_1-merror:0.128929\n",
            "[697]\tvalidation_0-merror:0.12929\tvalidation_1-merror:0.128929\n",
            "[698]\tvalidation_0-merror:0.12929\tvalidation_1-merror:0.128929\n",
            "[699]\tvalidation_0-merror:0.12929\tvalidation_1-merror:0.128929\n",
            "[700]\tvalidation_0-merror:0.12929\tvalidation_1-merror:0.128929\n",
            "[701]\tvalidation_0-merror:0.12929\tvalidation_1-merror:0.128929\n",
            "[702]\tvalidation_0-merror:0.12929\tvalidation_1-merror:0.128929\n",
            "[703]\tvalidation_0-merror:0.12929\tvalidation_1-merror:0.128929\n",
            "[704]\tvalidation_0-merror:0.12929\tvalidation_1-merror:0.128929\n",
            "[705]\tvalidation_0-merror:0.129278\tvalidation_1-merror:0.128923\n",
            "[706]\tvalidation_0-merror:0.129278\tvalidation_1-merror:0.128923\n",
            "[707]\tvalidation_0-merror:0.129278\tvalidation_1-merror:0.128923\n",
            "[708]\tvalidation_0-merror:0.129278\tvalidation_1-merror:0.128923\n",
            "[709]\tvalidation_0-merror:0.129278\tvalidation_1-merror:0.128923\n",
            "[710]\tvalidation_0-merror:0.129278\tvalidation_1-merror:0.128923\n",
            "[711]\tvalidation_0-merror:0.129278\tvalidation_1-merror:0.128923\n",
            "[712]\tvalidation_0-merror:0.129278\tvalidation_1-merror:0.128923\n",
            "[713]\tvalidation_0-merror:0.129278\tvalidation_1-merror:0.128923\n",
            "[714]\tvalidation_0-merror:0.129278\tvalidation_1-merror:0.128923\n",
            "[715]\tvalidation_0-merror:0.129278\tvalidation_1-merror:0.128926\n",
            "[716]\tvalidation_0-merror:0.129278\tvalidation_1-merror:0.128923\n",
            "[717]\tvalidation_0-merror:0.129278\tvalidation_1-merror:0.128923\n",
            "[718]\tvalidation_0-merror:0.129278\tvalidation_1-merror:0.128923\n",
            "[719]\tvalidation_0-merror:0.129278\tvalidation_1-merror:0.128923\n",
            "[720]\tvalidation_0-merror:0.129272\tvalidation_1-merror:0.128921\n",
            "[721]\tvalidation_0-merror:0.129272\tvalidation_1-merror:0.128921\n",
            "[722]\tvalidation_0-merror:0.129229\tvalidation_1-merror:0.128866\n",
            "[723]\tvalidation_0-merror:0.129229\tvalidation_1-merror:0.128866\n",
            "[724]\tvalidation_0-merror:0.129229\tvalidation_1-merror:0.128866\n",
            "[725]\tvalidation_0-merror:0.129229\tvalidation_1-merror:0.128866\n",
            "[726]\tvalidation_0-merror:0.129229\tvalidation_1-merror:0.128866\n",
            "[727]\tvalidation_0-merror:0.129229\tvalidation_1-merror:0.128866\n",
            "[728]\tvalidation_0-merror:0.129229\tvalidation_1-merror:0.128866\n",
            "[729]\tvalidation_0-merror:0.129229\tvalidation_1-merror:0.128866\n",
            "[730]\tvalidation_0-merror:0.129229\tvalidation_1-merror:0.128866\n",
            "[731]\tvalidation_0-merror:0.129229\tvalidation_1-merror:0.128866\n",
            "[732]\tvalidation_0-merror:0.129229\tvalidation_1-merror:0.128866\n",
            "[733]\tvalidation_0-merror:0.129229\tvalidation_1-merror:0.128866\n",
            "[734]\tvalidation_0-merror:0.129229\tvalidation_1-merror:0.128866\n",
            "[735]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128866\n",
            "[736]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128866\n",
            "[737]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128866\n",
            "[738]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128866\n",
            "[739]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128866\n",
            "[740]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128866\n",
            "[741]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128866\n",
            "[742]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128866\n",
            "[743]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128866\n",
            "[744]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128866\n",
            "[745]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128866\n",
            "[746]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128866\n",
            "[747]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128866\n",
            "[748]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128866\n",
            "[749]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128866\n",
            "[750]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128866\n",
            "[751]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128866\n",
            "[752]\tvalidation_0-merror:0.129229\tvalidation_1-merror:0.128866\n",
            "[753]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128866\n",
            "[754]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128858\n",
            "[755]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128858\n",
            "[756]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128858\n",
            "[757]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128858\n",
            "[758]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128858\n",
            "[759]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128858\n",
            "[760]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128858\n",
            "[761]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128858\n",
            "[762]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128858\n",
            "[763]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128858\n",
            "[764]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128858\n",
            "[765]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128858\n",
            "[766]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128858\n",
            "[767]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128858\n",
            "[768]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128858\n",
            "[769]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128858\n",
            "[770]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128858\n",
            "[771]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128858\n",
            "[772]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128858\n",
            "[773]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128858\n",
            "[774]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128858\n",
            "[775]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128858\n",
            "[776]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128858\n",
            "[777]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128858\n",
            "[778]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128858\n",
            "[779]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128858\n",
            "[780]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128858\n",
            "[781]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128858\n",
            "[782]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128858\n",
            "[783]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128858\n",
            "[784]\tvalidation_0-merror:0.129223\tvalidation_1-merror:0.128858\n",
            "[785]\tvalidation_0-merror:0.129217\tvalidation_1-merror:0.128855\n",
            "[786]\tvalidation_0-merror:0.129217\tvalidation_1-merror:0.128855\n",
            "[787]\tvalidation_0-merror:0.129217\tvalidation_1-merror:0.128855\n",
            "[788]\tvalidation_0-merror:0.129217\tvalidation_1-merror:0.128855\n",
            "[789]\tvalidation_0-merror:0.129217\tvalidation_1-merror:0.128855\n",
            "[790]\tvalidation_0-merror:0.129217\tvalidation_1-merror:0.128855\n",
            "[791]\tvalidation_0-merror:0.129199\tvalidation_1-merror:0.128829\n",
            "[792]\tvalidation_0-merror:0.129199\tvalidation_1-merror:0.128829\n",
            "[793]\tvalidation_0-merror:0.129199\tvalidation_1-merror:0.128829\n",
            "[794]\tvalidation_0-merror:0.129199\tvalidation_1-merror:0.128829\n",
            "[795]\tvalidation_0-merror:0.129199\tvalidation_1-merror:0.128829\n",
            "[796]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[797]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[798]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[799]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[800]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[801]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[802]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[803]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[804]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[805]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[806]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[807]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[808]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[809]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[810]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[811]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[812]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[813]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[814]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[815]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[816]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[817]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[818]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[819]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[820]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[821]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[822]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[823]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[824]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[825]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[826]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[827]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[828]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[829]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[830]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[831]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[832]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[833]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[834]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[835]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[836]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[837]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[838]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[839]\tvalidation_0-merror:0.129205\tvalidation_1-merror:0.128832\n",
            "[840]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[841]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[842]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[843]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[844]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[845]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[846]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[847]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[848]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[849]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[850]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[851]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[852]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[853]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[854]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[855]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[856]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[857]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[858]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[859]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[860]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[861]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[862]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[863]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[864]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[865]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[866]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[867]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[868]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[869]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[870]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[871]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[872]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[873]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[874]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[875]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[876]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[877]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[878]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[879]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[880]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[881]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[882]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[883]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[884]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[885]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[886]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[887]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[888]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[889]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[890]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[891]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[892]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[893]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[894]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[895]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[896]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[897]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[898]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[899]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[900]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[901]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[902]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[903]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[904]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[905]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[906]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[907]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[908]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[909]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[910]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[911]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[912]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[913]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[914]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[915]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[916]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[917]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[918]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[919]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[920]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[921]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[922]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[923]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[924]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[925]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[926]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[927]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[928]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[929]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[930]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[931]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[932]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[933]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[934]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[935]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[936]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[937]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[938]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[939]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[940]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[941]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[942]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[943]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[944]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[945]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[946]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[947]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[948]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[949]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[950]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[951]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[952]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[953]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[954]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[955]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[956]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[957]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[958]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[959]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[960]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[961]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[962]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[963]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[964]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[965]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[966]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[967]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[968]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[969]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[970]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[971]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[972]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[973]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[974]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[975]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[976]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[977]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[978]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[979]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[980]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[981]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[982]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[983]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[984]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[985]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[986]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[987]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[988]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[989]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[990]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[991]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[992]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[993]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[994]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[995]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[996]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[997]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[998]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n",
            "[999]\tvalidation_0-merror:0.129187\tvalidation_1-merror:0.128821\n"
          ]
        },
        {
          "data": {
            "text/plain": [
              "XGBClassifier(early_stopping_rounds =10, eta=0.1, gamma=0.1, lambda_=1.0,\n",
              "              maxDepth=10, n_estimators=1000, numRound=1000, numWorkers=1,\n",
              "              objective='multi:softprob', subsample=1.0)"
            ]
          },
          "execution_count": 23,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "print(\"Parameter optimization\")\n",
        "params = { \n",
        "    'eta': 0.1,\n",
        "    'gamma': 0.1,\n",
        "    'early_stopping_rounds': 10,\n",
        "    'maxDepth': 10, \n",
        "    'lambda_': 1.0,\n",
        "    \"n_estimators\": 1000,\n",
        "    'subsample': 1.0,\n",
        "    'numRound': 1000,\n",
        "    'numWorkers': 1,\n",
        "    'verbosity': 1\n",
        "}\n",
        "xgb_model = xgb.XGBClassifier(**params)\n",
        "\n",
        "xgb_model.fit(X=X_train, \n",
        "              y=y_train, \n",
        "              eval_set=evallist)  # merror = 1 - accuracy"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-Y7uzK2Wa1NJ",
        "outputId": "d28fbc99-a144-4d5a-9132-87654d429719"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "array([0.18515559, 0.01203109, 0.00476563, 0.33467907, 0.        ,\n",
              "       0.14832245, 0.13303672, 0.06019406, 0.12181537], dtype=float32)"
            ]
          },
          "execution_count": 24,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "xgb_model.feature_importances_"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "_6R00cHnba5s",
        "outputId": "57bdff8a-5369-4d74-c4dd-a654a53c4c57"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "<bound method XGBClassifier.evals_result of XGBClassifier(early_stopping_rounds =10, eta=0.1, gamma=0.1, lambda_=1.0,\n",
              "              maxDepth=10, n_estimators=1000, numRound=1000, numWorkers=1,\n",
              "              objective='multi:softprob', subsample=1.0)>"
            ]
          },
          "execution_count": 26,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "xgb_model.evals_result"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "MBnC6ubobdfa",
        "outputId": "77e51746-9c76-4b0f-97c7-ef9739c662ce"
      },
      "outputs": [
        {
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAcEAAAEWCAYAAAAegCx/AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3deXhV5bn+8e/NIKIMiqAFqwyCyhyFCs7BFhywVar1aPm14HDQ9uDQ1oEeZ1oVrbRosbUKTkhxpDjVgSPEsYogASwW0UIFtYgoSlK0gTy/P/ZKugkBAiTZCev+XNe+sta7pud92MnDetfaeykiMDMzS6MGuQ7AzMwsV1wEzcwstVwEzcwstVwEzcwstVwEzcwstVwEzcwstVwEzQxJ/ytpQq7jMKtt8ucEzbaPpKXAXsD6rOb9I+LD7dznORHxf9sXXf0j6Rqgc0T8v1zHYjs+nwmaVY9vR0SzrNc2F8DqIKlRLo+/repr3FZ/uQia1RBJLSVNlPSRpA8k/VJSw2TZfpJmSFol6RNJkyXtliybBOwLPCGpSNKlkvIlLa+w/6WSvpVMXyPpEUn3S/oCGL6541cS6zWS7k+mO0gKSWdKWibpM0nnSfqGpPmSVksan7XtcEmvSBov6XNJf5P0zazl7SQ9LulTSe9K+u8Kx82O+zzgf4H/Svo+L1nvTElvS1oj6e+Szs3aR76k5ZJ+JunjpL9nZi1vKmmspH8k8b0sqWmyrL+kV5M+zZOUv03/2FZvuQia1Zx7gHVAZ+AgYBBwTrJMwA1AO6ArsA9wDUBE/AB4n/+cXd5UxeOdBDwC7AZM3sLxq6If0AX4L2AccDnwLaA7cJqkoyus+x7QGrgamCqpVbLsAWB50tdTgeslHbOJuCcC1wMPJn3vnazzMXAi0AI4E/iNpIOz9vE1oCWwN3A2cJuk3ZNlNwN9gMOAVsClQKmkvYGngF8m7RcDj0pqsxU5snrORdCsekxLziZWS5omaS/gBOCiiCiOiI+B3wCnA0TEuxExPSK+ioiVwK+Boze9+yr5S0RMi4hSMsVik8evol9ExJcR8RxQDEyJiI8j4gPgJTKFtczHwLiIKImIB4FFwGBJ+wCHA5cl+yoEJgA/rCzuiFhbWSAR8VREvBcZLwDPAUdmrVICjE6O/2egCDhAUgPgLODCiPggItZHxKsR8RXw/4A/R8Sfk2NPB2YnebOU8Pi7WfU4OfsmFkmHAI2BjySVNTcAliXL9wJuIfOHvHmy7LPtjGFZ1nT7zR2/ilZkTa+tZL5Z1vwHseFddv8gc+bXDvg0ItZUWNZ3E3FXStLxZM4w9yfTj12ABVmrrIqIdVnz/0riaw3sTOYstaL2wPckfTurrTEwc0vx2I7DRdCsZiwDvgJaV/jjXOZ6IICeEfGppJOB8VnLK962XUzmDz8AybW9isN22dts6fjVbW9JyiqE+wKPAx8CrSQ1zyqE+wIfZG1bsa8bzEtqAjxK5uzxsYgokTSNzJDylnwCfAnsB8yrsGwZMCki/nujrSw1PBxqVgMi4iMyQ3ZjJbWQ1CC5GaZsyLM5mSG7z5NrU5dU2MUKoFPW/DvAzpIGS2oMXAE02Y7jV7c9gQskNZb0PTLXOf8cEcuAV4EbJO0sqReZa3b3b2ZfK4AOyVAmwE5k+roSWJecFQ6qSlDJ0PBdwK+TG3QaSjo0Kaz3A9+WdGzSvnNyk83Xt777Vl+5CJrVnB+S+QO+kMxQ5yNA22TZtcDBwOdkbs6YWmHbG4ArkmuMF0fE58CPyVxP+4DMmeFyNm9zx69ur5O5ieYT4Drg1IhYlSw7A+hA5qzwT8DVW/j848PJz1WS3kzOIC8AHiLTj++TOcusqovJDJ2+AXwK3Ag0SAr0SWTuRl1J5szwEvx3MVX8YXkz2y6ShpP5YP8RuY7FbGv5fzxmZpZaLoJmZpZaHg41M7PU8pmgmZmllj8nWMfstttu0blz51yHkVPFxcXsuuuuuQ4j55wH56CM87DlHMyZM+eTiNjqr7xzEaxj9tprL2bPnp3rMHKqoKCA/Pz8XIeRc86Dc1DGedhyDiT9Y1v26+FQMzNLLRdBMzNLLRdBMzNLLRdBMzNLLRdBMzNLLRdBMzNLLRdBMzNLLRdBMzNLLRdBMzNLLRdBMzNLLRdBMzNLLRdBMzNLLRdBMzNLLRdBMzNLLRdBMzNLLRdBMzNLLRdBMzNLLRdBMzNLLRdBMzOrNsuWLWPAgAF069aN7t27c8sttwBw5ZVX0qtXL/Ly8hg0aBAffvghAJMnT6ZXr1707NmTww47jHnz5pXvq0OHDvTs2ZO8vDzOPffc8vaHH36Y7t2706BBA2bPnr1d8boImplZtWnUqBFjx45l4cKFvPbaa9x2220sXLiQSy65hPnz51NYWMiJJ57I6NGjAejYsSMvvPACCxYs4Morr2TEiBEb7G/mzJkUFhbyhz/8obytR48eTJ06laOOOmr7493uPVQDSScDfwK6RsTfqrjNqxFx2DYcqwPwZET0qGTZscCNyWxn4ANgLTAfeBX4V0Tct7XH3BprS9bTYdRTNXmIOu9nPdcxPOU5AOcBnIMy9SUPS8cMpm3btrRt2xaA5s2b07VrVz744AO6detWvl5xcTGSADjssP/8Ge/fvz/Lly/f4nG6du1abTHXiSIInAG8nPy8OnuBpEYRsa7i/LYUwC2JiGeBZ5PjFAAXR8T2nWubmaXU0qVLmTt3Lv369QPg8ssv57777qNly5bMnDlzo/UnTpzI8ccfXz4viUGDBiGJ/Px88vPzqz3GnA+HSmoGHAGcDZyetOVLeknS48DCivPJOkXJzwckDc7a3z2STpXUIdnmzeS1XUVT0jWSLk6mCyT9RtJsSW9L+oakqZIWS/pl1jb/T9IsSYWS/iCp4fbEYGZWXxQVFXHKKacwbtw4WrRoAcB1113HsmXLGDp0KOPHj99g/ZkzZzJx4kRuvPHG8raXX36ZN998k6effppp06bx4osvVnucdeFM8CTgmYh4R9IqSX2S9oOBHhGxRFJ+9nyF7R8ETgOekrQT8E3gR4CAgRHxpaQuwBSgbzXG/e+I6CvpQuAxoA/wKfCepN8AewL/BRweESWSfgcMBTYaTpU0AhgB0Lp1G67qua7iKqmyV9PM8E/aOQ/OQZn6koeCggIA1q1bx89//nP69etHq1atytvLdOrUiVGjRjFgwAAA3nvvPa666irGjBnDggULNlh38eLFAPTr148pU6ZQWlpavmz16tXMmTOHoqKibY65LhTBM4BbkukHkvkngVkVCl7F+TJPA7dIagIcB7wYEWsltQTGS8oD1gP7V3Pcjyc/FwB/jYiPACT9HdiHzNltH+CNZOy7KfBxZTuKiDuAOwD27dQ5xi6oC/8sufOznutIew7AeQDnoEx9ycPSoflEBMOGDePwww9n3Lhx5csWL15Mly5dAPjtb39Lnz59yM/P5/333+ecc87h4Ycf3uD6YHFxMaWlpTRv3pzi4mJGjhzJzTffvMGQ6G677UafPn3o23fbz29ymlVJrYBjgJ6SAmgIBPAUUFxh9YrzACRnegXAsWTOvB5IFv0EWAH0JjPs+2U1h/9V8rM0a7psvhGZM9F7I+LnW7PTpo0bsmjM4C2vuAMrKChg6dD8XIeRc86Dc1CmPuXhlVdeYdKkSeUfbQC4/vrrmThxIosWLaJBgwa0b9+e22+/HYDRo0ezatUqfvzjHwOZu0tnz57NihUrGDJkCJA5szz00EM57rjjAPjTn/7E+eefz8qVKxk8eHD5cbZFrv9rcSowKSLKPwAi6QXgyK3cz4PAOWSGO4cnbS2B5RFRKmkYmQJbm54HHpP0m4j4OCn4zSPiH7Uch5lZrTniiCOIiI3aTzjhhErXnzBhAhMmTNiovVOnTht8ZjB7SHXIkCHlBbJM2d2mWyvXN8acQeajEdkeTdq3xnPA0cD/RcS/k7bfAcMkzQMOZBNnkjUlIhYCVwDPSZoPTAfa1mYMZma2eTk9E4yIAZW03QrcWqGtACio0NYsa7oEaFVh+WKgV1bTZUn7UmCjzwhWEkd+hflrKltWMbYKyx4kc5ZqZmZ1UK7PBM3MzHIm19cEc6bCt8OUWRIRQypb38zMdjypLYLZ3w5jZmbp5OFQMzNLLRdBMzNLLRdBMzNLLRdBMzNLLRdBMzNLLRdBMzNLLRdBMzNLLRdBMzNLLRdBMzNLLRdBMzNLLRdBMzNLLRdBq1c6dOhQ/sTqvn37AnDllVfSq1cv8vLyGDRoEB9++CEAn3/+Od/+9rfp3bs33bt35+677y7fT8OGDcnLyyMvL4/vfOc7OemLmeVejX6BtqSvAeOAbwCrgRXARRHxznbsczfg+xHxu2S+HXBrRJxaDSEj6XWgCZnnEzYFPkgW/Rj4aXUdZ1PWlqynw6inavIQdd7Peq5jeIUcLB0zuHx65syZtG7dunz+kksu4Re/+AUAt956K6NHj+b222/ntttuo1u3bjzxxBOsXLmSAw44gKFDh7LTTjvRtGlTCgsLa6dDZlZn1VgRVOZZ938C7o2I05O23sBewDvJfKOIWLeVu96NTEH6HUBEfAhUW2GKiH5JbMOBvhExMmvxq9V1HKs+LVq0KJ8uLi4m89YDSaxZs4aIoKioiFatWtGoUWofnGJmlajJ4dABQElE3F7WEBHzgIaSXpL0OLBQUkNJv5L0hqT5ks4FkNRM0vOS3pS0QNJJyW7GAPtJKky26yDprWSbnSXdnaw/V9KApH24pKmSnpG0WNJNW9uZCscZLmmapOmSlkoaKemnyTFfk9QqWW+/5Jhzkj4fuB35NDKFbdCgQfTp04c77rijvP3yyy9nn332YfLkyYwePRqAkSNH8vbbb9OuXTt69uzJLbfcQoMGmbf8l19+Sd++fenfvz/Tpk3LSV/MLPdq8r/FPYA5m1h2MNAjIpZIGgF8HhHfkNQEeEXSc8AyYEhEfCGpNfBaUjhHJdvmQaY4Ze33f4CIiJ5JwXlO0v7JsjzgIOArYJGk30bEsu3s30HAzsC7wGURcZCk3wA/JDMMfAdwXkQsltSPzNnrMRV3lORgBEDr1m24qufWnhzvWPZqmhkSzVZQUADATTfdRJs2bfjss8+4+OKLWbt2Lb1792bgwIEMHDiQyZMnc/HFF3PmmWfywgsv0Lp1a/74xz/y4Ycfcs455zBhwgR23XVXpkyZQps2bfjwww8577zzKC4uZu+9985BbzetqKiovN9p5RxkOA81l4NcjQ3NioglyfQgoJeksiHNlkAXYDlwvaSjgFJgbzJDqZtzBPBbgIj4m6R/AGVF8PmI+BxA0kKgPZlCu61mRsQaYI2kz4EnkvYFSX+aAYcBD5cNz5G51riRiLiDTMFk306dY+yCdA/Z/aznOirmYOnQ/I3WmzdvHiUlJeTn/2dZp06dOOGEE7j33nv51a9+xahRozjyyCMBmDhxIm3atOGQQw7ZYD/PPfccTZo02WA/dUFBQUGdi6m2OQcZzkPN5aAm/9r+lU1fqyvOmhZwfvKk9/80Zq7JtQH6RESJpKVkzrq21VdZ0+vZ/r5n7680a7402XcDYHXZGWtVNW3ckEVZN4GkUUFBQaVFr7i4mNLSUpo3b05xcTHPPfccV111FYsXL6ZLly4APPbYYxx4YGbUed999+X555/nyCOPZMWKFSxatIhOnTrx2Wefscsuu9CkSRM++eQTXnnlFS699NLa7KKZ1RE1WQRnkDmTG5Gc6SCpF3BkhfWeBX4kaUZS7PYnc0dmS+DjpG0AmTM3gDVA800c8yVgKDAj2c++wCIyw6+1KhnGXSLpexHxcHKjUK/kuqhtgxUrVjBkyBAA1q1bx/e//32OO+44TjnlFBYtWkSDBg1o3749t9+euQx95ZVXMnz4cHr27ElEcOONN9K6dWteffVVzj33XBo0aEBpaSmjRo2iW7duueyameVIjRXBiAhJQ4Bxki4DvgSWAhXvQpgAdADeTArFSuBkYDLwhKQFwGzgb8l+V0l6JblJ5Wngtqx9/Q74fbLNOmB4RHyVNRxZ24Ym8VwBNAYeAFwEt1GnTp2YN2/j9D366KOVrt+uXTuee+65jdoPO+wwFixYUO3xmVn9U6MXn5KPL5xWyaI7s9YpBf43eVV06Cb2+/0KTT2S9i+BMytZ/x7gnqz5EzcfeaXbLM06TsVlHSrbLrnuedyWjmVmZrnhb4wxM7PUSvVtiFnfDpPtBxHhsTIzsxRIdREs+3YYMzNLJw+HmplZarkImplZarkImplZarkImplZarkImplZarkImplZarkImplZarkImplZarkImplZarkImplZarkImplZarkIWrX58ssvOeSQQ+jduzfdu3fn6quvBmD48OF07NiRvLw88vLyKCwsBGDy5Mn06tWLnj17cthhh5U/K/D9998vXzcvL48WLVowbty4nPXLzHZcqf4C7WySiiKi2Vasnw9cXJVnE26NtSXr6TDqqercZa1YOmYwTZo0YcaMGTRr1oySkhKOOOIIjj/+eAB+9atfceqpp26wTceOHXnhhRfYfffdefrppxkxYgSvv/46++67b3mhXL9+PXvvvXf5E+XNzKqTi6BVG0k0a5b5f0RJSQklJSVI2uT6hx12WPl0//79Wb58+UbrPP/88+y33360b9+++gM2s9TzcGgFkvIlFUh6RNLfJE1W8pdc0nFJ25vAd7O22VXSXZJmSZor6aSk/RZJVyXTx0p6UdIOnfP169eTl5fHnnvuycCBA+nXL/O0qssvv5xevXrxk5/8hK+++mqj7SZOnFh+1pjtgQce4IwzzqjxuM0snRQRuY6hTigbDk2GOR8DugMfAq8AlwCzgcXAMcC7wIPALhFxoqTrgYURcb+k3YBZwEFAAG8AI4HbgRMi4r1Kjj0CGAHQunWbPleNu7NG+1oTeu7dcoP5oqIirrzySi644AJatGhBq1atKCkpYezYsbRr145hw4aVrzt37lzGjRvHrbfeSsuWLSkqKiofUj311FO5++67adWqVW13KefK8pBmzkGG87DlHAwYMGBORPTd2v16OLRysyJiOYCkQqADUAQsiYjFSfv9JIULGAR8R9LFyfzOwL4R8bak/wZeBH5SWQEEiIg7gDsA9u3UOcYuqH//LEuH5m/U9uabb7Jq1SrOPPPM8raddtqJm2++mfz8zPrz589n/PjxTJ8+nf333x+AgoIC8vPzeeyxx+jXrx/f/e53N9p3GpTlIc2cgwznoeZyUP/+2taO7PG69Ww5TwJOiYhFlSzrCawC2lXlwE0bN2TRmMFVCrKuWblyJY0bN2a33XZj7dq1TJ8+ncsuu4yPPvqItm3bEhFMmzaNHj16AJm7QL/73e8yadKk8gKYbcqUKR4KNbMa5SJYdX8DOkjaLzmjy/7r/CxwvqTzIyIkHRQRcyW1B35GZmj0z5KmRcTrOYi9Vnz00UcMGzaM9evXU1paymmnncaJJ57IMcccw8qVK4kI8vLyuP322wEYPXo0q1at4sc//jEAjRo1Yvbs2QAUFxczffp0/vCHP+SsP2a243MRrKKI+DK5dveUpH8BLwHNk8W/AMYB85MbX5ZI+jYwkczHKD6UdDZwj6RvRMSXuehDTevVqxdz587dqH3GjBmVrj9hwgQmTJhQ6bJdd92VVatWVWt8ZmYVuQgmyj4jGBEFQEFW+8is6WeAAyvZdi1wbiW7/VbWOnPIDI2amVkdsUPfrm9mZrY5LoJmZpZaLoJmZpZaLoJmZpZaLoJmZpZaLoJmZpZaLoJmZpZaLoJmZpZaLoJmZpZaLoJmZpZaLoJmZpZaVSqCkvaT1CSZzpd0QfLwWDMzs3qrqmeCjwLrJXUm8/DXfYA/1lhUZmZmtaCqRbA0ItYBQ4DfRsQlQNuaC8vMzKzmVbUIlkg6AxgGPJm0Na6ZkGxz1q9fz0EHHcSJJ54IwNChQznggAPo0aMHZ511FiUlJQB89tlnDBkyhF69enHIIYfw1ltv5TJsM7M6qapF8EzgUOC6iFgiqSMwqebCqhskdZD0VoW2ayRdnKuYbrnlFrp27Vo+P3ToUP72t7+xYMEC1q5dW/6Q2uuvv568vDzmz5/Pfffdx4UXXpirkM3M6qwqFcGIWAhcBryZzC+JiBtrMrAdlaTNPsh4bcl6Oox6aoNXmeXLl/PUU09xzjnnlLedcMIJSEIShxxyCMuXLwdg4cKFHHPMMQAceOCBLF26lBUrVtREl8zM6q2q3h36baAQeCaZz5P0eE0GVtcld8gulDRf0gNJ266S7pI0S9JcSScl7cMlPS5pBvD8th7zoosu4qabbqJBg43/2UpKSpg0aRLHHXccAL1792bq1KkAzJo1i3/84x/lBdLMzDI2e1aS5RrgEKAAICIKJXWqoZjqi1FAx4j4KuvjIpcDMyLirKRtlqT/S5YdDPSKiE8r7kjSCGAEQOvWbbiq57oNlhcUFPCXv/yFkpIS1qxZQ2FhIatWraKgoKB8nZtvvplOnTqxfv16CgoKOPzwwxk/fjydO3emU6dOdO7cmblz57JmzZrqzkO1Kyoq2qBvaeU8OAdlnIeay0FVi2BJRHwuKbuttNqjqXtiM+3zgcmSpgHTkvZBwHeyrhnuDOybTE+vrAACRMQdZD56wr6dOsfYBRv+sywdms+zzz7LnDlzGD58OF9++SVffPEFEyZM4P777+faa6+lUaNGPPTQQxucJQ4ePLhs/3Ts2JHTTjuNFi1abGUKal9BQQH5+fm5DiPnnAfnoIzzUHM5qGoR/Kuk7wMNJXUBLgBerfZo6p5VwO4V2loBS4DBwFHAt4HLJfUEBJwSEYuyN5DUDyiuygGbNm7IojGDN2q/4YYbuOGGG4DMm+Hmm2/m/vvvZ8KECTz77LM8//zzGxTA1atXs8suu7DTTjsxYcIEjjrqqHpRAM3MalNV7w49H+gOfEXmQ/KfAxfVVFB1RUQUAR9JOgZAUivgOOBlYJ+ImEnmhqGWQDPgWeB8JafMkg6q6RjPO+88VqxYwaGHHkpeXh6jR48G4O2336ZHjx4ccMABPP3009xyyy01HYqZWb2zxTNBSQ2BpyJiAJlrXmnzQ+A2Sb9O5q8F3gdmSmpJ5uzv1ohYLekXwDhgvqQGZM4YT6zugPLz88uHBdatW1fpOoceeijvvPNOdR/azGyHssUiGBHrJZVKahkRn9dGUHVJ8vGQAZUsOqKSddcC51bSfg9wT3XHZmZm26eq1wSLgAWSppN1bSsiLqiRqMzMzGpBVYvg1ORlZma2w6hSEYyIe2s6EDMzs9pWpSIoaQmVfGYuItL+gXkzM6vHqjoc2jdremfge2Q+L2dmZlZvVfULtFdlvT6IiHFkPixuZmZWb1V1OPTgrNkGZM4Mq3oWaWZmVidVtZCNzZpeR+ZD4KdVfzhmZma1p6pF8OyI+Ht2Q/JgXTMzs3qrqt8d+kgV28zMzOqNzZ4JSjqQzBdnt5T03axFLcjcJWpmZlZvbWk49AAyXwC9G5lHBpVZA/x3TQVlZmZWGzZbBCPiMeAxSYdGxF9qKSYzM7NaUdUbY+ZK+h8yQ6Plw6ARcVaNRGVmZlYLqnpjzCTga8CxwAvA18kMiVoNWrZsGQMGDKBbt2507969/MG4Dz/8MN27d6dBgwbMnj27fP3JkyeTl5dX/mrQoAGFhYW5Ct/MrM6r6plg54j4nqSTIuJeSX8EXqrJwAwaNWrE2LFjOfjgg1mzZg19+vRh4MCB9OjRg6lTp3LuuRs+unDo0KEMHToUgAULFnDyySeTl5eXi9DNzOqFqhbBkuTnakk9gH8Ce9ZMSOm2tmQ9HUY9BcDSMYNp27YtAM2bN6dr16588MEHDBw4cIv7mTJlCqeffnqNxmpmVt9VdTj0Dkm7A1cCjwMLgZtqLKodhKRvSJovaWdJu0r6a/KfiK22dOlS5s6dS79+/aq0/oMPPsgZZ5yxLYcyM0uNqj5PcEIy+QLgxydVUUS8Ielx4JdAU+D+iHir4nqSRgAjAFq3bsNVPdcBUFBQAMDatWu58MILOeecc3jzzTfLt1u9ejVz5syhqKhog/0tXLiQiOCTTz4p30d9UlRUVC/jrm7Og3NQxnmowRxExBZfwF7ARODpZL4bma9Sq9L2aX4BOwHzgNeBhltaf5+O+0X7y56M9pc9GRER//73v2PQoEExduzYqOjoo4+ON954Y6P2iy66KK677rqN2uuLmTNn5jqEOsF5cA7KOA9bzgEwO7bhb3RVh0PvAZ4F2iXz7wAXVVMd3tHtATQDmrOV37ITEZx99tl07dqVn/70p1XaprS0lIceesjXA83MqqCqN8a0joiHJP0cICLWSVpfg3HtSP5A5lpqR+BGYOTmVm7auCGLxmQe1fjyyy8zadIkevbsWX6X5/XXX89XX33F+eefz8qVKxk8eDB5eXk8++yzALz44ovss88+dOrkUWszsy2pahEslrQHEACS+gOf11hUOwhJPwRKIuKPkhoCr0o6JiJmVGX7I444omxIdSNDhgyptD0/P5/XXnttW0M2M0uVqhbBn5K5K3Q/Sa8AbYBTayyqHURE3Afcl0yvB6p2a6eZmdWKLT1FYt+IeD8i3pR0NJkv1BawKCJKNretmZlZXbelG2OmZU0/GBF/jYi3XADNzGxHsKUiqKxp32lhZmY7lC0VwdjEtJmZWb23pRtjekv6gswZYdNkmmQ+IqJFjUZnZmZWg7b0UN2GtRWImZlZbavqN8aYmZntcFwEzcwstVwEzcwstVwEzcwstVwEzcwstVwEzcwstVwEzcwstVwEzcwstVwE67CzzjqLPffckx49epS3FRYW0r9/f/Ly8ujbty+zZs0CoKCggJYtW5KXl0deXh6jR4/OVdhmZvWGi2AdNnz4cJ555pkN2i699FKuvvpqCgsLGT16NJdeemn5siOPPJLCwkIKCwu56qqrajtcM7N6x0Wwjllbsp4Oo54C4KijjqJVq1YbLJfEF19kvsL1888/p127drUeo5nZjqKqT5a3bSTpPOC8ZLYlsDQiBmzr/saNG8exxx7LxRdfTGlpKa+++mr5sr/85S/07t2bdu3acfPNN9O9e/ftC97MbAenCMDnw/kAAA6rSURBVD8hqTZIagzMAG6KiCcqLBsBjABo3bpNn6vG3UnPvVsC8M9//pOf//zn3H333QDceuut9O7dm6OPPpqZM2fy5JNPMnbsWIqLi2nQoAFNmzbltddeY/z48dx///212sfqUlRURLNmzXIdRs45D85BGedhyzkYMGDAnIjou9U7jgi/auEF/A64dkvr7dNxv2h/2ZNRZsmSJdG9e/fy+RYtWkRpaWlERJSWlkbz5s2jMu3bt4+VK1dWuqyumzlzZq5DqBOcB+egjPOw5RwAs2Mb/jb7mmAtkDQcaA9cu737ateuHS+88AIAM2bMoEuXLkDmjDGSs/pZs2ZRWlrKHnvssb2HMzPbofmaYA2T1Ae4GDgyIkq3tH7Txg1ZNGYwAGeccQYFBQV88sknfP3rX+faa6/lzjvv5MILL2TdunXsvPPO3HHHHQA88sgj/P73v6dRo0Y0bdqUBx54AEk12TUzs3rPRbDmjQRaATOTojQ7Is6pyoZTpkyptH3OnDkbH2TkSEaOHLntUZqZpZCLYA2LiDNzHYOZmVXO1wTNzCy1XATNzCy1XATNzCy1XATNzCy1XATNzCy1XATNzCy1XATNzCy1XATNzCy1XATNzCy1XATNzCy1XATNzCy1XATNzCy1XATNzCy1XATNzCy1XATrqLPOOos999yTHj16lLddcsklHHjggfTq1YshQ4awevVqAKZPn06fPn3o2bMnffr0YcaMGbkK28ysXklVEZS0XlKhpL9KmifpZ5K2KQeS+kq6tbpjLDN8+HCeeeaZDdoGDhzIW2+9xfz589l///254YYbAGjdujVPPPEECxYs4N577+UHP/hBTYVlZrZDSVURBNZGRF5EdAcGAscDV2/LjiJidkRcUK3RAWtL1gNw1FFH0apVqw2WDRo0iEaNMs9B7t+/P8uXLwfgoIMOol27dgB0796dtWvX8tVXX1V3aGZmO5y0FcFyEfExMAIYqYyGkn4l6Q1J8yWdCyDpAUmDy7aTdI+kUyXlS3oyaWsm6W5JC5JtT0naB0n6i6Q3JT0sqVl1xX/XXXdx/PHHb9T+6KOPcvDBB9OkSZPqOpSZ2Q6rUa4DyKWI+LukhsCewEnA5xHxDUlNgFckPQc8CJwGPCVpJ+CbwI+Aflm7ujLZtieApN0ltQauAL4VEcWSLgN+CoyuGIekEWQKMq1bt6GgoACAf/7znxQXF5fPl7n//vtZvXo1e++99wbLlixZwhVXXMFNN9200Tb1SVFRUb2Ov7o4D85BGeehBnMQEal5AUWVtK0G9gIeAd4BCpPXEmAQsDPwPtCETKGcnGyXDzyZTM8BulTY74nAJ1n7WwhM3FKM+3TcL8osWbIkunfvHtnuvvvu6N+/fxQXF2/QvmzZsujSpUu8/PLLUd/NnDkz1yHUCc6Dc1DGedhyDoDZsQ11IdVngpI6AeuBjwEB50fEs5WsVwAcC/wX8EBVdw9Mj4gzqidaeOaZZ7jpppt44YUX2GWXXcrbV69ezeDBgxkzZgyHH354dR3OzGyHl9prgpLaALcD45P/RTwL/EhS42T5/pJ2TVZ/EDgTOBJ4ppLdTQf+J2vfuwOvAYdL6py07Spp/y3F1bRxQwDOOOMMDj30UBYtWsTXv/51Jk6cyMiRI1mzZg0DBw4kLy+P8847D4Dx48fz7rvvMnr0aPLy8sjLy+Pjjz/elrSYmaVK2s4Em0oqBBoD64BJwK+TZROADsCbkgSsBE5Olj2XrPtYRPy7kv3+ErhN0ltkziyvjYipkoYDU5JrjJC5RvhOVQKdMmXKRm1nn312peteccUVXHHFFVXZrZmZZUlVEYyIhptZVgr8b/KquKwEaFWhrQAoSKaLgGGVbDcD+Mb2xGxmZjUntcOhZmZmLoJmZpZaLoJmZpZaLoJmZpZaLoJmZpZaLoJmZpZaLoJmZpZaLoJmZpZaLoJmZpZaLoJmZpZaLoJmZpZaLoJmZpZaLoJmZpZaLoJmZpZaLoJ12C233EKPHj3o3r0748aNA+DTTz9l4MCBdOnShYEDB/LZZ5/lOEozs/rLRbCOeuutt7jzzjuZNWsW8+bN48knn+Tdd99lzJgxfPOb32Tx4sV885vfZMyYMbkO1cys3nIRrGPWlqwH4O2336Zfv37ssssuNGrUiKOPPpqpU6fy2GOPMWxY5vm9w4YNY9q0abkM18ysXnMRrEGSRku6KGv+OkkXVmXbHj168NJLL7Fq1Sr+9a9/8ec//5lly5axYsUK2rZtC8DXvvY1VqxYUUPRm5nt+BrlOoAd3F3AVGCcpAbA6cAhVdmwa9euXHbZZQwaNIhdd92VvLw8GjZsuME6kpBU7UGbmaWFi2ANioilklZJOgjYC5gbEasqridpBDACoHXrNhQUFACw3377MXbsWADuvPNO2rRpQ4sWLXj00UfZY489WLVqFc2bNy9ff0dRVFS0w/VpWzgPzkEZ56HmcuAiWPMmAMOBr5E5M9xIRNwB3AGwb6fOkZ+fD8DHH3/Mnnvuyfvvv8+cOXN47bXXaNy4MYsXL+aUU05hzJgxnH766ZStv6MoKCjY4fq0LZwH56CM81BzOXARrHl/AkYDjYHvb2nlpo3/M+R5yimnsGrVKho3bsxtt93GbrvtxqhRozjttNOYOHEi7du356GHHqq5yM3MdnAugjUsIv4taSawOiLWb822L7300kZte+yxB88//3x1hWdmlmougjUsuSGmP/C9XMdiZmYb8kckapCkbsC7wPMRsTjX8ZiZ2YZ8JliDImIh0CnXcZiZWeV8JmhmZqnlImhmZqnlImhmZqnlImhmZqnlImhmZqnlImhmZqnlImhmZqnlImhmZqnlImhmZqnlImhmZqnlImhmZqnlImhmZqnlImhmZqnlImhmZqnlImhmZqnlImhmZqnlImhmZqmliMh1DJZF0hpgUa7jyLHWwCe5DqIOcB6cgzLOw5Zz0D4i2mztThttezxWQxZFRN9cB5FLkmanPQfgPIBzUMZ5qLkceDjUzMxSy0XQzMxSy0Ww7rkj1wHUAc5BhvPgHJRxHmooB74xxszMUstngmZmllougmZmllougnWEpOMkLZL0rqRRuY6nuklaKmmBpEJJs5O2VpKmS1qc/Nw9aZekW5NczJd0cNZ+hiXrL5Y0LFf9qSpJd0n6WNJbWW3V1m9JfZK8vptsq9rt4ZZtIgfXSPogeT8USjoha9nPk/4sknRsVnulvyOSOkp6PWl/UNJOtde7qpO0j6SZkhZK+qukC5P21LwfNpOD3L0fIsKvHL+AhsB7QCdgJ2Ae0C3XcVVzH5cCrSu03QSMSqZHATcm0ycATwMC+gOvJ+2tgL8nP3dPpnfPdd+20O+jgIOBt2qi38CsZF0l2x6f6z5XMQfXABdXsm635P3fBOiY/F403NzvCPAQcHoyfTvwo1z3eRN5aAscnEw3B95J+pua98NmcpCz94PPBOuGQ4B3I+LvEfFv4AHgpBzHVBtOAu5Npu8FTs5qvy8yXgN2k9QWOBaYHhGfRsRnwHTguNoOemtExIvApxWaq6XfybIWEfFaZH7j78vaV52xiRxsyknAAxHxVUQsAd4l8/tR6e9IcqZzDPBIsn12PuuUiPgoIt5MptcAbwN7k6L3w2ZysCk1/n5wEawb9gaWZc0vZ/NvjPoogOckzZE0ImnbKyI+Sqb/CeyVTG8qHztKnqqr33sn0xXb64uRyTDfXWVDgGx9DvYAVkfEugrtdZqkDsBBwOuk9P1QIQeQo/eDi6DVliMi4mDgeOB/JB2VvTD5n2vqPq+T1n4Dvwf2A/KAj4CxuQ2n9khqBjwKXBQRX2QvS8v7oZIc5Oz94CJYN3wA7JM1//WkbYcRER8kPz8G/kRmOGNFMoRD8vPjZPVN5WNHyVN19fuDZLpie50XESsiYn1ElAJ3knk/wNbnYBWZYcJGFdrrJEmNyfzxnxwRU5PmVL0fKstBLt8PLoJ1wxtAl+Supp2A04HHcxxTtZG0q6TmZdPAIOAtMn0su7NtGPBYMv048MPk7rj+wOfJcNGzwCBJuyfDJYOStvqmWvqdLPtCUv/kWsgPs/ZVp5X90U8MIfN+gEwOTpfURFJHoAuZmz0q/R1JzpxmAqcm22fns05J/o0mAm9HxK+zFqXm/bCpHOT0/ZDru4X8Kr8L6gQyd0q9B1ye63iquW+dyNy9NQ/4a1n/yIzfPw8sBv4PaJW0C7gtycUCoG/Wvs4ic3H8XeDMXPetCn2fQmZ4p4TM9Ymzq7PfQN/kD8Z7wHiSb4GqS69N5GBS0sf5yR+6tlnrX570ZxFZdzdu6nckeX/NSnLzMNAk133eRB6OIDPUOR8oTF4npOn9sJkc5Oz94K9NMzOz1PJwqJmZpZaLoJmZpZaLoJmZpZaLoJmZpZaLoJmZpVajLa9iZvWdpPVkbkEvc3JELM1ROGZ1hj8iYZYCkooiolktHq9R/Of7G83qLA+HmhmS2kp6MXmW21uSjkzaj5P0pqR5kp5P2lpJmpZ82fFrknol7ddImiTpFWCSpDaSHpX0RvI6PIddNKuUh0PN0qGppMJkeklEDKmw/PtkvnrrOkkNgV0ktSHzPY5HRcQSSa2Sda8F5kbEyZKOIfPInrxkWTcyX5a+VtIfgd9ExMuS9iXzdV9da7CPZlvNRdAsHdZGRN5mlr8B3JV8ufG0iCiUlA+8GJnnuBERZc8EPAI4JWmbIWkPSS2SZY9HxNpk+ltAN/3n4eYtJDWLiKLq65bZ9nERNDMi4sXk8VaDgXsk/Rr4bBt2VZw13QDoHxFfVkeMZjXB1wTNDEntgRURcScwATgYeA04Kvn2frKGQ18ChiZt+cAnUeG5eInngPOzjrG5M1GznPCZoJkB5AOXSCoBioAfRsRKSSOAqZIakHnO3UDgGjJDp/OBf/GfxwBVdAFwW7JeI+BF4Lwa7YXZVvJHJMzMLLU8HGpmZqnlImhmZqnlImhmZqnlImhmZqnlImhmZqnlImhmZqnlImhmZqn1/wFaDTgYovNBigAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "import matplotlib.pylab as plt\n",
        "from matplotlib import pyplot\n",
        "from xgboost import plot_importance\n",
        "plot_importance(xgb_model, max_num_features=14) # top 10 most important features\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "D4i6G3gZHxEk",
        "outputId": "d3e04854-3384-4f17-ce51-b3190f55defc"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[[ 1.42468674e+12  1.42468673e+18  1.00000000e+00 ...  7.06873782e-07\n",
            "  -1.74148009e-05 -6.24661285e-06]\n",
            " [ 1.42468674e+12  1.42468673e+18  1.00000000e+00 ... -1.02900436e-04\n",
            "   4.46322366e-05 -2.96337294e-04]\n",
            " [ 1.42468674e+12  1.42468673e+18  1.00000000e+00 ... -1.22996046e-04\n",
            "  -7.03837705e-05  6.90792455e-04]\n",
            " ...\n",
            " [ 1.42478946e+12  1.42478946e+18  1.00000000e+00 ... -1.01538865e-01\n",
            "  -5.96681383e-02  6.37344045e-02]\n",
            " [ 1.42478946e+12  1.42478946e+18  1.00000000e+00 ...  8.91808620e-01\n",
            "   1.27676156e-02  6.19116456e-03]\n",
            " [ 1.42478946e+12  1.42479131e+18  2.00000000e+00 ... -2.62146646e-01\n",
            "   4.07927964e-01 -5.31421774e-02]]\n",
            "['Arrival_Time' 'Creation_Time' 'Device' 'Index' 'Model' 'User' 'gt' 'x'\n",
            " 'y' 'z' 'Arrival_Time Creation_Time' 'Arrival_Time Device'\n",
            " 'Arrival_Time Index' 'Arrival_Time Model' 'Arrival_Time User'\n",
            " 'Arrival_Time gt' 'Arrival_Time x' 'Arrival_Time y' 'Arrival_Time z'\n",
            " 'Creation_Time Device' 'Creation_Time Index' 'Creation_Time Model'\n",
            " 'Creation_Time User' 'Creation_Time gt' 'Creation_Time x'\n",
            " 'Creation_Time y' 'Creation_Time z' 'Device Index' 'Device Model'\n",
            " 'Device User' 'Device gt' 'Device x' 'Device y' 'Device z' 'Index Model'\n",
            " 'Index User' 'Index gt' 'Index x' 'Index y' 'Index z' 'Model User'\n",
            " 'Model gt' 'Model x' 'Model y' 'Model z' 'User gt' 'User x' 'User y'\n",
            " 'User z' 'gt x' 'gt y' 'gt z' 'x y' 'x z' 'y z']\n",
            "55\n"
          ]
        }
      ],
      "source": [
        "from sklearn.preprocessing import PolynomialFeatures\n",
        "poly = PolynomialFeatures(interaction_only=True, include_bias=False)\n",
        "print(poly.fit_transform(data_pd))\n",
        "print(poly.get_feature_names_out())\n",
        "print(len(poly.get_feature_names_out()))\n",
        "poly_data = pd.DataFrame(poly.fit_transform(data_pd), columns=poly.get_feature_names_out())"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "i7a19O53WCtF",
        "outputId": "ad564751-c431-4e07-8931-de8e6521e78a"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Parameter optimization\n",
            "[0]\tvalidation_0-merror:0.247252\tvalidation_1-merror:0.246585\n",
            "[1]\tvalidation_0-merror:0.253771\tvalidation_1-merror:0.252523\n",
            "[2]\tvalidation_0-merror:0.249113\tvalidation_1-merror:0.24879\n",
            "[3]\tvalidation_0-merror:0.232944\tvalidation_1-merror:0.231535\n",
            "[4]\tvalidation_0-merror:0.231955\tvalidation_1-merror:0.230329\n",
            "[5]\tvalidation_0-merror:0.228109\tvalidation_1-merror:0.226745\n",
            "[6]\tvalidation_0-merror:0.22958\tvalidation_1-merror:0.228343\n",
            "[7]\tvalidation_0-merror:0.229825\tvalidation_1-merror:0.228202\n",
            "[8]\tvalidation_0-merror:0.227639\tvalidation_1-merror:0.225905\n",
            "[9]\tvalidation_0-merror:0.228677\tvalidation_1-merror:0.22708\n",
            "[10]\tvalidation_0-merror:0.227908\tvalidation_1-merror:0.22624\n",
            "[11]\tvalidation_0-merror:0.226772\tvalidation_1-merror:0.225159\n",
            "[12]\tvalidation_0-merror:0.225008\tvalidation_1-merror:0.223462\n",
            "[13]\tvalidation_0-merror:0.225735\tvalidation_1-merror:0.224359\n",
            "[14]\tvalidation_0-merror:0.224087\tvalidation_1-merror:0.222601\n",
            "[15]\tvalidation_0-merror:0.222469\tvalidation_1-merror:0.220935\n",
            "[16]\tvalidation_0-merror:0.220272\tvalidation_1-merror:0.218622\n",
            "[17]\tvalidation_0-merror:0.220119\tvalidation_1-merror:0.21852\n",
            "[18]\tvalidation_0-merror:0.21758\tvalidation_1-merror:0.216249\n",
            "[19]\tvalidation_0-merror:0.217128\tvalidation_1-merror:0.215731\n",
            "[20]\tvalidation_0-merror:0.217\tvalidation_1-merror:0.215553\n",
            "[21]\tvalidation_0-merror:0.220034\tvalidation_1-merror:0.218292\n",
            "[22]\tvalidation_0-merror:0.214192\tvalidation_1-merror:0.212843\n",
            "[23]\tvalidation_0-merror:0.217702\tvalidation_1-merror:0.2161\n",
            "[24]\tvalidation_0-merror:0.212709\tvalidation_1-merror:0.211284\n",
            "[25]\tvalidation_0-merror:0.212977\tvalidation_1-merror:0.211054\n",
            "[26]\tvalidation_0-merror:0.212782\tvalidation_1-merror:0.210865\n",
            "[27]\tvalidation_0-merror:0.20705\tvalidation_1-merror:0.205503\n",
            "[28]\tvalidation_0-merror:0.207032\tvalidation_1-merror:0.205055\n",
            "[29]\tvalidation_0-merror:0.202344\tvalidation_1-merror:0.200634\n",
            "[30]\tvalidation_0-merror:0.201044\tvalidation_1-merror:0.199054\n",
            "[31]\tvalidation_0-merror:0.205585\tvalidation_1-merror:0.203658\n",
            "[32]\tvalidation_0-merror:0.205042\tvalidation_1-merror:0.202996\n",
            "[33]\tvalidation_0-merror:0.198669\tvalidation_1-merror:0.197095\n",
            "[34]\tvalidation_0-merror:0.195977\tvalidation_1-merror:0.19463\n",
            "[35]\tvalidation_0-merror:0.195208\tvalidation_1-merror:0.193866\n",
            "[36]\tvalidation_0-merror:0.194909\tvalidation_1-merror:0.193516\n",
            "[37]\tvalidation_0-merror:0.193969\tvalidation_1-merror:0.192663\n",
            "[38]\tvalidation_0-merror:0.192809\tvalidation_1-merror:0.191376\n",
            "[39]\tvalidation_0-merror:0.191857\tvalidation_1-merror:0.190447\n",
            "[40]\tvalidation_0-merror:0.190734\tvalidation_1-merror:0.189414\n",
            "[41]\tvalidation_0-merror:0.188665\tvalidation_1-merror:0.187326\n",
            "[42]\tvalidation_0-merror:0.187828\tvalidation_1-merror:0.186515\n",
            "[43]\tvalidation_0-merror:0.181334\tvalidation_1-merror:0.180679\n",
            "[44]\tvalidation_0-merror:0.179576\tvalidation_1-merror:0.178856\n",
            "[45]\tvalidation_0-merror:0.178672\tvalidation_1-merror:0.177807\n",
            "[46]\tvalidation_0-merror:0.176017\tvalidation_1-merror:0.175133\n",
            "[47]\tvalidation_0-merror:0.173789\tvalidation_1-merror:0.172951\n",
            "[48]\tvalidation_0-merror:0.17288\tvalidation_1-merror:0.172093\n",
            "[49]\tvalidation_0-merror:0.173063\tvalidation_1-merror:0.172336\n",
            "[50]\tvalidation_0-merror:0.172245\tvalidation_1-merror:0.171434\n",
            "[51]\tvalidation_0-merror:0.170658\tvalidation_1-merror:0.169938\n",
            "[52]\tvalidation_0-merror:0.169712\tvalidation_1-merror:0.169263\n",
            "[53]\tvalidation_0-merror:0.168619\tvalidation_1-merror:0.168282\n",
            "[54]\tvalidation_0-merror:0.167575\tvalidation_1-merror:0.167439\n",
            "[55]\tvalidation_0-merror:0.167386\tvalidation_1-merror:0.167274\n",
            "[56]\tvalidation_0-merror:0.166623\tvalidation_1-merror:0.166395\n",
            "[57]\tvalidation_0-merror:0.165878\tvalidation_1-merror:0.16559\n",
            "[58]\tvalidation_0-merror:0.165579\tvalidation_1-merror:0.165213\n",
            "[59]\tvalidation_0-merror:0.164493\tvalidation_1-merror:0.164156\n",
            "[60]\tvalidation_0-merror:0.164572\tvalidation_1-merror:0.16431\n",
            "[61]\tvalidation_0-merror:0.162387\tvalidation_1-merror:0.162118\n",
            "[62]\tvalidation_0-merror:0.161727\tvalidation_1-merror:0.161488\n",
            "[63]\tvalidation_0-merror:0.161715\tvalidation_1-merror:0.161415\n",
            "[64]\tvalidation_0-merror:0.160476\tvalidation_1-merror:0.16008\n",
            "[65]\tvalidation_0-merror:0.160067\tvalidation_1-merror:0.159544\n",
            "[66]\tvalidation_0-merror:0.159658\tvalidation_1-merror:0.159097\n",
            "[67]\tvalidation_0-merror:0.159377\tvalidation_1-merror:0.158806\n",
            "[68]\tvalidation_0-merror:0.159518\tvalidation_1-merror:0.158919\n",
            "[69]\tvalidation_0-merror:0.158205\tvalidation_1-merror:0.15777\n",
            "[70]\tvalidation_0-merror:0.154885\tvalidation_1-merror:0.154082\n",
            "[71]\tvalidation_0-merror:0.154537\tvalidation_1-merror:0.15387\n",
            "[72]\tvalidation_0-merror:0.152993\tvalidation_1-merror:0.152026\n",
            "[73]\tvalidation_0-merror:0.152492\tvalidation_1-merror:0.151675\n",
            "[74]\tvalidation_0-merror:0.151808\tvalidation_1-merror:0.151018\n",
            "[75]\tvalidation_0-merror:0.15132\tvalidation_1-merror:0.150684\n",
            "[76]\tvalidation_0-merror:0.15096\tvalidation_1-merror:0.15054\n",
            "[77]\tvalidation_0-merror:0.15074\tvalidation_1-merror:0.150126\n",
            "[78]\tvalidation_0-merror:0.150441\tvalidation_1-merror:0.149771\n",
            "[79]\tvalidation_0-merror:0.150246\tvalidation_1-merror:0.149674\n",
            "[80]\tvalidation_0-merror:0.150191\tvalidation_1-merror:0.149619\n",
            "[81]\tvalidation_0-merror:0.148982\tvalidation_1-merror:0.148316\n",
            "[82]\tvalidation_0-merror:0.148768\tvalidation_1-merror:0.148122\n",
            "[83]\tvalidation_0-merror:0.148616\tvalidation_1-merror:0.147814\n",
            "[84]\tvalidation_0-merror:0.14803\tvalidation_1-merror:0.147322\n",
            "[85]\tvalidation_0-merror:0.147548\tvalidation_1-merror:0.146945\n",
            "[86]\tvalidation_0-merror:0.146296\tvalidation_1-merror:0.145901\n",
            "[87]\tvalidation_0-merror:0.146193\tvalidation_1-merror:0.145865\n",
            "[88]\tvalidation_0-merror:0.145491\tvalidation_1-merror:0.145185\n",
            "[89]\tvalidation_0-merror:0.144404\tvalidation_1-merror:0.144177\n",
            "[90]\tvalidation_0-merror:0.143696\tvalidation_1-merror:0.143251\n",
            "[91]\tvalidation_0-merror:0.141889\tvalidation_1-merror:0.141313\n",
            "[92]\tvalidation_0-merror:0.141401\tvalidation_1-merror:0.140876\n",
            "[93]\tvalidation_0-merror:0.141267\tvalidation_1-merror:0.140727\n",
            "[94]\tvalidation_0-merror:0.140693\tvalidation_1-merror:0.140136\n",
            "[95]\tvalidation_0-merror:0.140162\tvalidation_1-merror:0.139759\n",
            "[96]\tvalidation_0-merror:0.139478\tvalidation_1-merror:0.139236\n",
            "[97]\tvalidation_0-merror:0.138379\tvalidation_1-merror:0.138095\n",
            "[98]\tvalidation_0-merror:0.137732\tvalidation_1-merror:0.137478\n",
            "[99]\tvalidation_0-merror:0.137513\tvalidation_1-merror:0.137276\n",
            "[100]\tvalidation_0-merror:0.136878\tvalidation_1-merror:0.136643\n",
            "[101]\tvalidation_0-merror:0.136365\tvalidation_1-merror:0.136183\n",
            "[102]\tvalidation_0-merror:0.135437\tvalidation_1-merror:0.135097\n",
            "[103]\tvalidation_0-merror:0.134229\tvalidation_1-merror:0.133763\n",
            "[104]\tvalidation_0-merror:0.133582\tvalidation_1-merror:0.133146\n",
            "[105]\tvalidation_0-merror:0.133936\tvalidation_1-merror:0.133692\n",
            "[106]\tvalidation_0-merror:0.13338\tvalidation_1-merror:0.133295\n",
            "[107]\tvalidation_0-merror:0.132684\tvalidation_1-merror:0.132604\n",
            "[108]\tvalidation_0-merror:0.130725\tvalidation_1-merror:0.130537\n",
            "[109]\tvalidation_0-merror:0.130572\tvalidation_1-merror:0.130339\n",
            "[110]\tvalidation_0-merror:0.129712\tvalidation_1-merror:0.129614\n",
            "[111]\tvalidation_0-merror:0.130078\tvalidation_1-merror:0.12998\n",
            "[112]\tvalidation_0-merror:0.128033\tvalidation_1-merror:0.127921\n",
            "[113]\tvalidation_0-merror:0.12749\tvalidation_1-merror:0.127309\n",
            "[114]\tvalidation_0-merror:0.126946\tvalidation_1-merror:0.126695\n",
            "[115]\tvalidation_0-merror:0.127099\tvalidation_1-merror:0.126802\n",
            "[116]\tvalidation_0-merror:0.126922\tvalidation_1-merror:0.126883\n",
            "[117]\tvalidation_0-merror:0.126214\tvalidation_1-merror:0.125965\n",
            "[118]\tvalidation_0-merror:0.126025\tvalidation_1-merror:0.125784\n",
            "[119]\tvalidation_0-merror:0.124767\tvalidation_1-merror:0.124293\n",
            "[120]\tvalidation_0-merror:0.124291\tvalidation_1-merror:0.124086\n",
            "[121]\tvalidation_0-merror:0.124523\tvalidation_1-merror:0.124374\n",
            "[122]\tvalidation_0-merror:0.124364\tvalidation_1-merror:0.124175\n",
            "[123]\tvalidation_0-merror:0.124303\tvalidation_1-merror:0.12405\n",
            "[124]\tvalidation_0-merror:0.124175\tvalidation_1-merror:0.123867\n",
            "[125]\tvalidation_0-merror:0.123644\tvalidation_1-merror:0.123362\n",
            "[126]\tvalidation_0-merror:0.123443\tvalidation_1-merror:0.1232\n",
            "[127]\tvalidation_0-merror:0.123217\tvalidation_1-merror:0.123009\n",
            "[128]\tvalidation_0-merror:0.12152\tvalidation_1-merror:0.121217\n",
            "[129]\tvalidation_0-merror:0.121489\tvalidation_1-merror:0.121248\n",
            "[130]\tvalidation_0-merror:0.121373\tvalidation_1-merror:0.121128\n",
            "[131]\tvalidation_0-merror:0.12072\tvalidation_1-merror:0.120455\n",
            "[132]\tvalidation_0-merror:0.120537\tvalidation_1-merror:0.120369\n",
            "[133]\tvalidation_0-merror:0.120366\tvalidation_1-merror:0.120322\n",
            "[134]\tvalidation_0-merror:0.120262\tvalidation_1-merror:0.120173\n",
            "[135]\tvalidation_0-merror:0.12025\tvalidation_1-merror:0.12016\n",
            "[136]\tvalidation_0-merror:0.12014\tvalidation_1-merror:0.120115\n",
            "[137]\tvalidation_0-merror:0.119915\tvalidation_1-merror:0.11994\n",
            "[138]\tvalidation_0-merror:0.119982\tvalidation_1-merror:0.120029\n",
            "[139]\tvalidation_0-merror:0.119866\tvalidation_1-merror:0.119958\n",
            "[140]\tvalidation_0-merror:0.119841\tvalidation_1-merror:0.120071\n",
            "[141]\tvalidation_0-merror:0.119548\tvalidation_1-merror:0.119906\n",
            "[142]\tvalidation_0-merror:0.119506\tvalidation_1-merror:0.119864\n",
            "[143]\tvalidation_0-merror:0.119493\tvalidation_1-merror:0.119827\n",
            "[144]\tvalidation_0-merror:0.119457\tvalidation_1-merror:0.119796\n",
            "[145]\tvalidation_0-merror:0.1187\tvalidation_1-merror:0.119134\n",
            "[146]\tvalidation_0-merror:0.11881\tvalidation_1-merror:0.119265\n",
            "[147]\tvalidation_0-merror:0.118596\tvalidation_1-merror:0.118941\n",
            "[148]\tvalidation_0-merror:0.118462\tvalidation_1-merror:0.118844\n",
            "[149]\tvalidation_0-merror:0.118566\tvalidation_1-merror:0.119006\n",
            "[150]\tvalidation_0-merror:0.118517\tvalidation_1-merror:0.118988\n",
            "[151]\tvalidation_0-merror:0.118517\tvalidation_1-merror:0.118928\n",
            "[152]\tvalidation_0-merror:0.118535\tvalidation_1-merror:0.11892\n",
            "[153]\tvalidation_0-merror:0.118407\tvalidation_1-merror:0.118674\n",
            "[154]\tvalidation_0-merror:0.118431\tvalidation_1-merror:0.118687\n",
            "[155]\tvalidation_0-merror:0.118431\tvalidation_1-merror:0.118682\n",
            "[156]\tvalidation_0-merror:0.118437\tvalidation_1-merror:0.118661\n",
            "[157]\tvalidation_0-merror:0.117906\tvalidation_1-merror:0.118035\n",
            "[158]\tvalidation_0-merror:0.117839\tvalidation_1-merror:0.117933\n",
            "[159]\tvalidation_0-merror:0.1179\tvalidation_1-merror:0.117913\n",
            "[160]\tvalidation_0-merror:0.11779\tvalidation_1-merror:0.1178\n",
            "[161]\tvalidation_0-merror:0.117766\tvalidation_1-merror:0.117808\n",
            "[162]\tvalidation_0-merror:0.117632\tvalidation_1-merror:0.11768\n",
            "[163]\tvalidation_0-merror:0.117589\tvalidation_1-merror:0.11764\n",
            "[164]\tvalidation_0-merror:0.117149\tvalidation_1-merror:0.117204\n",
            "[165]\tvalidation_0-merror:0.116881\tvalidation_1-merror:0.116958\n",
            "[166]\tvalidation_0-merror:0.116887\tvalidation_1-merror:0.116947\n",
            "[167]\tvalidation_0-merror:0.116869\tvalidation_1-merror:0.116926\n",
            "[168]\tvalidation_0-merror:0.116838\tvalidation_1-merror:0.116884\n",
            "[169]\tvalidation_0-merror:0.116832\tvalidation_1-merror:0.116887\n",
            "[170]\tvalidation_0-merror:0.115727\tvalidation_1-merror:0.115778\n",
            "[171]\tvalidation_0-merror:0.116051\tvalidation_1-merror:0.11627\n",
            "[172]\tvalidation_0-merror:0.115556\tvalidation_1-merror:0.115841\n",
            "[173]\tvalidation_0-merror:0.115544\tvalidation_1-merror:0.115817\n",
            "[174]\tvalidation_0-merror:0.115477\tvalidation_1-merror:0.115749\n",
            "[175]\tvalidation_0-merror:0.114799\tvalidation_1-merror:0.114985\n",
            "[176]\tvalidation_0-merror:0.115062\tvalidation_1-merror:0.115417\n",
            "[177]\tvalidation_0-merror:0.114519\tvalidation_1-merror:0.114938\n",
            "[178]\tvalidation_0-merror:0.114036\tvalidation_1-merror:0.11447\n",
            "[179]\tvalidation_0-merror:0.114262\tvalidation_1-merror:0.114789\n",
            "[180]\tvalidation_0-merror:0.114067\tvalidation_1-merror:0.114572\n",
            "[181]\tvalidation_0-merror:0.113957\tvalidation_1-merror:0.114389\n",
            "[182]\tvalidation_0-merror:0.113334\tvalidation_1-merror:0.113436\n",
            "[183]\tvalidation_0-merror:0.113328\tvalidation_1-merror:0.113439\n",
            "[184]\tvalidation_0-merror:0.113316\tvalidation_1-merror:0.113387\n",
            "[185]\tvalidation_0-merror:0.113304\tvalidation_1-merror:0.113387\n",
            "[186]\tvalidation_0-merror:0.113176\tvalidation_1-merror:0.113342\n",
            "[187]\tvalidation_0-merror:0.113194\tvalidation_1-merror:0.113337\n",
            "[188]\tvalidation_0-merror:0.113127\tvalidation_1-merror:0.113256\n",
            "[189]\tvalidation_0-merror:0.11309\tvalidation_1-merror:0.11324\n",
            "[190]\tvalidation_0-merror:0.113115\tvalidation_1-merror:0.113277\n",
            "[191]\tvalidation_0-merror:0.113041\tvalidation_1-merror:0.113149\n",
            "[192]\tvalidation_0-merror:0.112846\tvalidation_1-merror:0.112908\n",
            "[193]\tvalidation_0-merror:0.112968\tvalidation_1-merror:0.113026\n",
            "[194]\tvalidation_0-merror:0.112608\tvalidation_1-merror:0.112696\n",
            "[195]\tvalidation_0-merror:0.112553\tvalidation_1-merror:0.112654\n",
            "[196]\tvalidation_0-merror:0.112388\tvalidation_1-merror:0.112563\n",
            "[197]\tvalidation_0-merror:0.112364\tvalidation_1-merror:0.112542\n",
            "[198]\tvalidation_0-merror:0.112339\tvalidation_1-merror:0.1125\n",
            "[199]\tvalidation_0-merror:0.112358\tvalidation_1-merror:0.112531\n",
            "[200]\tvalidation_0-merror:0.112431\tvalidation_1-merror:0.112576\n",
            "[201]\tvalidation_0-merror:0.112315\tvalidation_1-merror:0.112432\n",
            "[202]\tvalidation_0-merror:0.112181\tvalidation_1-merror:0.11228\n",
            "[203]\tvalidation_0-merror:0.112175\tvalidation_1-merror:0.11228\n",
            "[204]\tvalidation_0-merror:0.112138\tvalidation_1-merror:0.112262\n",
            "[205]\tvalidation_0-merror:0.111949\tvalidation_1-merror:0.111841\n",
            "[206]\tvalidation_0-merror:0.111753\tvalidation_1-merror:0.111673\n",
            "[207]\tvalidation_0-merror:0.111882\tvalidation_1-merror:0.111932\n",
            "[208]\tvalidation_0-merror:0.111888\tvalidation_1-merror:0.111943\n",
            "[209]\tvalidation_0-merror:0.111808\tvalidation_1-merror:0.111893\n",
            "[210]\tvalidation_0-merror:0.111759\tvalidation_1-merror:0.111864\n",
            "[211]\tvalidation_0-merror:0.111558\tvalidation_1-merror:0.111587\n",
            "[212]\tvalidation_0-merror:0.111558\tvalidation_1-merror:0.111613\n",
            "[213]\tvalidation_0-merror:0.11143\tvalidation_1-merror:0.111346\n",
            "[214]\tvalidation_0-merror:0.111344\tvalidation_1-merror:0.111338\n",
            "[215]\tvalidation_0-merror:0.111228\tvalidation_1-merror:0.111174\n",
            "[216]\tvalidation_0-merror:0.111064\tvalidation_1-merror:0.11099\n",
            "[217]\tvalidation_0-merror:0.110563\tvalidation_1-merror:0.110478\n",
            "[218]\tvalidation_0-merror:0.110569\tvalidation_1-merror:0.110504\n",
            "[219]\tvalidation_0-merror:0.110569\tvalidation_1-merror:0.110504\n",
            "[220]\tvalidation_0-merror:0.11063\tvalidation_1-merror:0.110543\n",
            "[221]\tvalidation_0-merror:0.110679\tvalidation_1-merror:0.110567\n",
            "[222]\tvalidation_0-merror:0.110673\tvalidation_1-merror:0.110564\n",
            "[223]\tvalidation_0-merror:0.110673\tvalidation_1-merror:0.110588\n",
            "[224]\tvalidation_0-merror:0.110441\tvalidation_1-merror:0.110407\n",
            "[225]\tvalidation_0-merror:0.110417\tvalidation_1-merror:0.110386\n",
            "[226]\tvalidation_0-merror:0.110227\tvalidation_1-merror:0.110258\n",
            "[227]\tvalidation_0-merror:0.110307\tvalidation_1-merror:0.110268\n",
            "[228]\tvalidation_0-merror:0.11024\tvalidation_1-merror:0.110263\n",
            "[229]\tvalidation_0-merror:0.110136\tvalidation_1-merror:0.110132\n",
            "[230]\tvalidation_0-merror:0.110136\tvalidation_1-merror:0.110138\n",
            "[231]\tvalidation_0-merror:0.110215\tvalidation_1-merror:0.110216\n",
            "[232]\tvalidation_0-merror:0.110215\tvalidation_1-merror:0.110221\n",
            "[233]\tvalidation_0-merror:0.110185\tvalidation_1-merror:0.110146\n",
            "[234]\tvalidation_0-merror:0.11016\tvalidation_1-merror:0.11013\n",
            "[235]\tvalidation_0-merror:0.110142\tvalidation_1-merror:0.110111\n",
            "[236]\tvalidation_0-merror:0.109855\tvalidation_1-merror:0.10997\n",
            "[237]\tvalidation_0-merror:0.109794\tvalidation_1-merror:0.109913\n",
            "[238]\tvalidation_0-merror:0.109629\tvalidation_1-merror:0.109549\n",
            "[239]\tvalidation_0-merror:0.10958\tvalidation_1-merror:0.109667\n",
            "[240]\tvalidation_0-merror:0.109367\tvalidation_1-merror:0.109209\n",
            "[241]\tvalidation_0-merror:0.1093\tvalidation_1-merror:0.109081\n",
            "[242]\tvalidation_0-merror:0.109141\tvalidation_1-merror:0.10894\n",
            "[243]\tvalidation_0-merror:0.109092\tvalidation_1-merror:0.108615\n",
            "[244]\tvalidation_0-merror:0.108878\tvalidation_1-merror:0.108445\n",
            "[245]\tvalidation_0-merror:0.10886\tvalidation_1-merror:0.108437\n",
            "[246]\tvalidation_0-merror:0.108872\tvalidation_1-merror:0.108466\n",
            "[247]\tvalidation_0-merror:0.108872\tvalidation_1-merror:0.108482\n",
            "[248]\tvalidation_0-merror:0.108585\tvalidation_1-merror:0.108236\n",
            "[249]\tvalidation_0-merror:0.108585\tvalidation_1-merror:0.108236\n",
            "[250]\tvalidation_0-merror:0.108579\tvalidation_1-merror:0.108233\n",
            "[251]\tvalidation_0-merror:0.108469\tvalidation_1-merror:0.108071\n",
            "[252]\tvalidation_0-merror:0.108207\tvalidation_1-merror:0.10794\n",
            "[253]\tvalidation_0-merror:0.108225\tvalidation_1-merror:0.107856\n",
            "[254]\tvalidation_0-merror:0.108207\tvalidation_1-merror:0.10782\n",
            "[255]\tvalidation_0-merror:0.108219\tvalidation_1-merror:0.10783\n",
            "[256]\tvalidation_0-merror:0.108262\tvalidation_1-merror:0.10788\n",
            "[257]\tvalidation_0-merror:0.10828\tvalidation_1-merror:0.107856\n",
            "[258]\tvalidation_0-merror:0.108274\tvalidation_1-merror:0.10787\n",
            "[259]\tvalidation_0-merror:0.10828\tvalidation_1-merror:0.107877\n",
            "[260]\tvalidation_0-merror:0.108286\tvalidation_1-merror:0.107822\n",
            "[261]\tvalidation_0-merror:0.108274\tvalidation_1-merror:0.107862\n",
            "[262]\tvalidation_0-merror:0.107969\tvalidation_1-merror:0.107642\n",
            "[263]\tvalidation_0-merror:0.107816\tvalidation_1-merror:0.107409\n",
            "[264]\tvalidation_0-merror:0.107658\tvalidation_1-merror:0.107386\n",
            "[265]\tvalidation_0-merror:0.107651\tvalidation_1-merror:0.107388\n",
            "[266]\tvalidation_0-merror:0.10756\tvalidation_1-merror:0.107333\n",
            "[267]\tvalidation_0-merror:0.107529\tvalidation_1-merror:0.107323\n",
            "[268]\tvalidation_0-merror:0.107535\tvalidation_1-merror:0.107323\n",
            "[269]\tvalidation_0-merror:0.10778\tvalidation_1-merror:0.107529\n",
            "[270]\tvalidation_0-merror:0.107816\tvalidation_1-merror:0.107561\n",
            "[271]\tvalidation_0-merror:0.107371\tvalidation_1-merror:0.107297\n",
            "[272]\tvalidation_0-merror:0.107316\tvalidation_1-merror:0.107257\n",
            "[273]\tvalidation_0-merror:0.107157\tvalidation_1-merror:0.107085\n",
            "[274]\tvalidation_0-merror:0.107157\tvalidation_1-merror:0.107093\n",
            "[275]\tvalidation_0-merror:0.107145\tvalidation_1-merror:0.107106\n",
            "[276]\tvalidation_0-merror:0.107139\tvalidation_1-merror:0.1071\n",
            "[277]\tvalidation_0-merror:0.106998\tvalidation_1-merror:0.106985\n",
            "[278]\tvalidation_0-merror:0.107023\tvalidation_1-merror:0.10676\n",
            "[279]\tvalidation_0-merror:0.106711\tvalidation_1-merror:0.106705\n",
            "[280]\tvalidation_0-merror:0.106779\tvalidation_1-merror:0.106768\n",
            "[281]\tvalidation_0-merror:0.106656\tvalidation_1-merror:0.106598\n",
            "[282]\tvalidation_0-merror:0.106638\tvalidation_1-merror:0.10659\n",
            "[283]\tvalidation_0-merror:0.106455\tvalidation_1-merror:0.106266\n",
            "[284]\tvalidation_0-merror:0.106418\tvalidation_1-merror:0.10635\n",
            "[285]\tvalidation_0-merror:0.106473\tvalidation_1-merror:0.106224\n",
            "[286]\tvalidation_0-merror:0.106467\tvalidation_1-merror:0.106229\n",
            "[287]\tvalidation_0-merror:0.106022\tvalidation_1-merror:0.105748\n",
            "[288]\tvalidation_0-merror:0.106034\tvalidation_1-merror:0.105777\n",
            "[289]\tvalidation_0-merror:0.105875\tvalidation_1-merror:0.105819\n",
            "[290]\tvalidation_0-merror:0.105906\tvalidation_1-merror:0.105829\n",
            "[291]\tvalidation_0-merror:0.106064\tvalidation_1-merror:0.105819\n",
            "[292]\tvalidation_0-merror:0.106071\tvalidation_1-merror:0.105842\n",
            "[293]\tvalidation_0-merror:0.105893\tvalidation_1-merror:0.105664\n",
            "[294]\tvalidation_0-merror:0.105826\tvalidation_1-merror:0.105588\n",
            "[295]\tvalidation_0-merror:0.105863\tvalidation_1-merror:0.10563\n",
            "[296]\tvalidation_0-merror:0.105845\tvalidation_1-merror:0.105609\n",
            "[297]\tvalidation_0-merror:0.105808\tvalidation_1-merror:0.105549\n",
            "[298]\tvalidation_0-merror:0.105619\tvalidation_1-merror:0.105353\n",
            "[299]\tvalidation_0-merror:0.105704\tvalidation_1-merror:0.1054\n",
            "[300]\tvalidation_0-merror:0.10571\tvalidation_1-merror:0.105376\n",
            "[301]\tvalidation_0-merror:0.105723\tvalidation_1-merror:0.105384\n",
            "[302]\tvalidation_0-merror:0.105265\tvalidation_1-merror:0.104913\n",
            "[303]\tvalidation_0-merror:0.104813\tvalidation_1-merror:0.104445\n",
            "[304]\tvalidation_0-merror:0.104807\tvalidation_1-merror:0.104448\n",
            "[305]\tvalidation_0-merror:0.104624\tvalidation_1-merror:0.104382\n",
            "[306]\tvalidation_0-merror:0.104508\tvalidation_1-merror:0.104265\n",
            "[307]\tvalidation_0-merror:0.104422\tvalidation_1-merror:0.104184\n",
            "[308]\tvalidation_0-merror:0.104416\tvalidation_1-merror:0.104184\n",
            "[309]\tvalidation_0-merror:0.104306\tvalidation_1-merror:0.104105\n",
            "[310]\tvalidation_0-merror:0.104282\tvalidation_1-merror:0.103951\n",
            "[311]\tvalidation_0-merror:0.104245\tvalidation_1-merror:0.103951\n",
            "[312]\tvalidation_0-merror:0.104062\tvalidation_1-merror:0.10377\n",
            "[313]\tvalidation_0-merror:0.103952\tvalidation_1-merror:0.103621\n",
            "[314]\tvalidation_0-merror:0.103659\tvalidation_1-merror:0.103409\n",
            "[315]\tvalidation_0-merror:0.103531\tvalidation_1-merror:0.103312\n",
            "[316]\tvalidation_0-merror:0.103592\tvalidation_1-merror:0.103271\n",
            "[317]\tvalidation_0-merror:0.103623\tvalidation_1-merror:0.103302\n",
            "[318]\tvalidation_0-merror:0.103653\tvalidation_1-merror:0.103257\n",
            "[319]\tvalidation_0-merror:0.103623\tvalidation_1-merror:0.103213\n",
            "[320]\tvalidation_0-merror:0.103525\tvalidation_1-merror:0.103135\n",
            "[321]\tvalidation_0-merror:0.103275\tvalidation_1-merror:0.102852\n",
            "[322]\tvalidation_0-merror:0.103214\tvalidation_1-merror:0.102763\n",
            "[323]\tvalidation_0-merror:0.103067\tvalidation_1-merror:0.102556\n",
            "[324]\tvalidation_0-merror:0.10325\tvalidation_1-merror:0.102844\n",
            "[325]\tvalidation_0-merror:0.102835\tvalidation_1-merror:0.102428\n",
            "[326]\tvalidation_0-merror:0.10286\tvalidation_1-merror:0.10242\n",
            "[327]\tvalidation_0-merror:0.102848\tvalidation_1-merror:0.102415\n",
            "[328]\tvalidation_0-merror:0.102841\tvalidation_1-merror:0.102418\n",
            "[329]\tvalidation_0-merror:0.102823\tvalidation_1-merror:0.102413\n",
            "[330]\tvalidation_0-merror:0.102835\tvalidation_1-merror:0.10241\n",
            "[331]\tvalidation_0-merror:0.102841\tvalidation_1-merror:0.102415\n",
            "[332]\tvalidation_0-merror:0.102719\tvalidation_1-merror:0.102324\n",
            "[333]\tvalidation_0-merror:0.102719\tvalidation_1-merror:0.102331\n",
            "[334]\tvalidation_0-merror:0.102713\tvalidation_1-merror:0.102321\n",
            "[335]\tvalidation_0-merror:0.102671\tvalidation_1-merror:0.102303\n",
            "[336]\tvalidation_0-merror:0.102689\tvalidation_1-merror:0.102331\n",
            "[337]\tvalidation_0-merror:0.102689\tvalidation_1-merror:0.102316\n",
            "[338]\tvalidation_0-merror:0.102689\tvalidation_1-merror:0.10231\n",
            "[339]\tvalidation_0-merror:0.102689\tvalidation_1-merror:0.102292\n",
            "[340]\tvalidation_0-merror:0.102664\tvalidation_1-merror:0.102292\n",
            "[341]\tvalidation_0-merror:0.102664\tvalidation_1-merror:0.102292\n",
            "[342]\tvalidation_0-merror:0.102664\tvalidation_1-merror:0.102295\n",
            "[343]\tvalidation_0-merror:0.102664\tvalidation_1-merror:0.102287\n",
            "[344]\tvalidation_0-merror:0.102664\tvalidation_1-merror:0.102274\n",
            "[345]\tvalidation_0-merror:0.102622\tvalidation_1-merror:0.102227\n",
            "[346]\tvalidation_0-merror:0.102622\tvalidation_1-merror:0.102219\n",
            "[347]\tvalidation_0-merror:0.102622\tvalidation_1-merror:0.102219\n",
            "[348]\tvalidation_0-merror:0.102609\tvalidation_1-merror:0.102208\n",
            "[349]\tvalidation_0-merror:0.102609\tvalidation_1-merror:0.102206\n",
            "[350]\tvalidation_0-merror:0.102609\tvalidation_1-merror:0.102206\n",
            "[351]\tvalidation_0-merror:0.102634\tvalidation_1-merror:0.102222\n",
            "[352]\tvalidation_0-merror:0.102542\tvalidation_1-merror:0.102313\n",
            "[353]\tvalidation_0-merror:0.102677\tvalidation_1-merror:0.102316\n",
            "[354]\tvalidation_0-merror:0.102677\tvalidation_1-merror:0.102316\n",
            "[355]\tvalidation_0-merror:0.102646\tvalidation_1-merror:0.102305\n",
            "[356]\tvalidation_0-merror:0.102622\tvalidation_1-merror:0.102261\n",
            "[357]\tvalidation_0-merror:0.102603\tvalidation_1-merror:0.102222\n",
            "[358]\tvalidation_0-merror:0.102609\tvalidation_1-merror:0.102232\n",
            "[359]\tvalidation_0-merror:0.102603\tvalidation_1-merror:0.102227\n",
            "[360]\tvalidation_0-merror:0.102597\tvalidation_1-merror:0.102237\n",
            "[361]\tvalidation_0-merror:0.102384\tvalidation_1-merror:0.102104\n",
            "[362]\tvalidation_0-merror:0.10239\tvalidation_1-merror:0.10212\n",
            "[363]\tvalidation_0-merror:0.102414\tvalidation_1-merror:0.102156\n",
            "[364]\tvalidation_0-merror:0.102207\tvalidation_1-merror:0.101991\n",
            "[365]\tvalidation_0-merror:0.10195\tvalidation_1-merror:0.101782\n",
            "[366]\tvalidation_0-merror:0.101669\tvalidation_1-merror:0.101463\n",
            "[367]\tvalidation_0-merror:0.101755\tvalidation_1-merror:0.101536\n",
            "[368]\tvalidation_0-merror:0.101834\tvalidation_1-merror:0.101541\n",
            "[369]\tvalidation_0-merror:0.101706\tvalidation_1-merror:0.101395\n",
            "[370]\tvalidation_0-merror:0.101608\tvalidation_1-merror:0.101384\n",
            "[371]\tvalidation_0-merror:0.101639\tvalidation_1-merror:0.101518\n",
            "[372]\tvalidation_0-merror:0.101621\tvalidation_1-merror:0.101489\n",
            "[373]\tvalidation_0-merror:0.101431\tvalidation_1-merror:0.101272\n",
            "[374]\tvalidation_0-merror:0.10137\tvalidation_1-merror:0.101165\n",
            "[375]\tvalidation_0-merror:0.101376\tvalidation_1-merror:0.10105\n",
            "[376]\tvalidation_0-merror:0.101096\tvalidation_1-merror:0.100762\n",
            "[377]\tvalidation_0-merror:0.100998\tvalidation_1-merror:0.100691\n",
            "[378]\tvalidation_0-merror:0.101004\tvalidation_1-merror:0.100689\n",
            "[379]\tvalidation_0-merror:0.100949\tvalidation_1-merror:0.100644\n",
            "[380]\tvalidation_0-merror:0.100937\tvalidation_1-merror:0.100644\n",
            "[381]\tvalidation_0-merror:0.100906\tvalidation_1-merror:0.10061\n",
            "[382]\tvalidation_0-merror:0.100839\tvalidation_1-merror:0.100513\n",
            "[383]\tvalidation_0-merror:0.100839\tvalidation_1-merror:0.100529\n",
            "[384]\tvalidation_0-merror:0.100845\tvalidation_1-merror:0.1005\n",
            "[385]\tvalidation_0-merror:0.100644\tvalidation_1-merror:0.100424\n",
            "[386]\tvalidation_0-merror:0.100565\tvalidation_1-merror:0.100283\n",
            "[387]\tvalidation_0-merror:0.100571\tvalidation_1-merror:0.100291\n",
            "[388]\tvalidation_0-merror:0.100491\tvalidation_1-merror:0.100257\n",
            "[389]\tvalidation_0-merror:0.100491\tvalidation_1-merror:0.100239\n",
            "[390]\tvalidation_0-merror:0.100418\tvalidation_1-merror:0.100137\n",
            "[391]\tvalidation_0-merror:0.1004\tvalidation_1-merror:0.100029\n",
            "[392]\tvalidation_0-merror:0.100382\tvalidation_1-merror:0.100045\n",
            "[393]\tvalidation_0-merror:0.099979\tvalidation_1-merror:0.099734\n",
            "[394]\tvalidation_0-merror:0.099643\tvalidation_1-merror:0.099268\n",
            "[395]\tvalidation_0-merror:0.099448\tvalidation_1-merror:0.099135\n",
            "[396]\tvalidation_0-merror:0.099307\tvalidation_1-merror:0.099006\n",
            "[397]\tvalidation_0-merror:0.098733\tvalidation_1-merror:0.098593\n",
            "[398]\tvalidation_0-merror:0.098727\tvalidation_1-merror:0.098585\n",
            "[399]\tvalidation_0-merror:0.098434\tvalidation_1-merror:0.09807\n",
            "[400]\tvalidation_0-merror:0.098294\tvalidation_1-merror:0.097973\n",
            "[401]\tvalidation_0-merror:0.098355\tvalidation_1-merror:0.098002\n",
            "[402]\tvalidation_0-merror:0.098343\tvalidation_1-merror:0.097991\n",
            "[403]\tvalidation_0-merror:0.098318\tvalidation_1-merror:0.097978\n",
            "[404]\tvalidation_0-merror:0.098288\tvalidation_1-merror:0.097968\n",
            "[405]\tvalidation_0-merror:0.098324\tvalidation_1-merror:0.097968\n",
            "[406]\tvalidation_0-merror:0.098447\tvalidation_1-merror:0.098057\n",
            "[407]\tvalidation_0-merror:0.098331\tvalidation_1-merror:0.097855\n",
            "[408]\tvalidation_0-merror:0.098337\tvalidation_1-merror:0.097845\n",
            "[409]\tvalidation_0-merror:0.098294\tvalidation_1-merror:0.097753\n",
            "[410]\tvalidation_0-merror:0.09808\tvalidation_1-merror:0.097552\n",
            "[411]\tvalidation_0-merror:0.098062\tvalidation_1-merror:0.09751\n",
            "[412]\tvalidation_0-merror:0.098123\tvalidation_1-merror:0.097604\n",
            "[413]\tvalidation_0-merror:0.098092\tvalidation_1-merror:0.097549\n",
            "[414]\tvalidation_0-merror:0.098038\tvalidation_1-merror:0.097455\n",
            "[415]\tvalidation_0-merror:0.097867\tvalidation_1-merror:0.097243\n",
            "[416]\tvalidation_0-merror:0.097842\tvalidation_1-merror:0.097272\n",
            "[417]\tvalidation_0-merror:0.097842\tvalidation_1-merror:0.097259\n",
            "[418]\tvalidation_0-merror:0.097848\tvalidation_1-merror:0.097288\n",
            "[419]\tvalidation_0-merror:0.097818\tvalidation_1-merror:0.097275\n",
            "[420]\tvalidation_0-merror:0.097769\tvalidation_1-merror:0.097259\n",
            "[421]\tvalidation_0-merror:0.097781\tvalidation_1-merror:0.097248\n",
            "[422]\tvalidation_0-merror:0.097452\tvalidation_1-merror:0.09684\n",
            "[423]\tvalidation_0-merror:0.09725\tvalidation_1-merror:0.096642\n",
            "[424]\tvalidation_0-merror:0.097207\tvalidation_1-merror:0.096574\n",
            "[425]\tvalidation_0-merror:0.097195\tvalidation_1-merror:0.096545\n",
            "[426]\tvalidation_0-merror:0.097268\tvalidation_1-merror:0.09655\n",
            "[427]\tvalidation_0-merror:0.09714\tvalidation_1-merror:0.096435\n",
            "[428]\tvalidation_0-merror:0.097122\tvalidation_1-merror:0.096443\n",
            "[429]\tvalidation_0-merror:0.097122\tvalidation_1-merror:0.096435\n",
            "[430]\tvalidation_0-merror:0.097116\tvalidation_1-merror:0.09643\n",
            "[431]\tvalidation_0-merror:0.097061\tvalidation_1-merror:0.09638\n",
            "[432]\tvalidation_0-merror:0.097061\tvalidation_1-merror:0.096372\n",
            "[433]\tvalidation_0-merror:0.097085\tvalidation_1-merror:0.096377\n",
            "[434]\tvalidation_0-merror:0.097085\tvalidation_1-merror:0.096372\n",
            "[435]\tvalidation_0-merror:0.097073\tvalidation_1-merror:0.096385\n",
            "[436]\tvalidation_0-merror:0.097049\tvalidation_1-merror:0.096354\n",
            "[437]\tvalidation_0-merror:0.097049\tvalidation_1-merror:0.096356\n",
            "[438]\tvalidation_0-merror:0.097055\tvalidation_1-merror:0.096369\n",
            "[439]\tvalidation_0-merror:0.097104\tvalidation_1-merror:0.096354\n",
            "[440]\tvalidation_0-merror:0.097061\tvalidation_1-merror:0.096328\n",
            "[441]\tvalidation_0-merror:0.097061\tvalidation_1-merror:0.096304\n",
            "[442]\tvalidation_0-merror:0.096988\tvalidation_1-merror:0.096288\n",
            "[443]\tvalidation_0-merror:0.096554\tvalidation_1-merror:0.095888\n",
            "[444]\tvalidation_0-merror:0.096573\tvalidation_1-merror:0.095812\n",
            "[445]\tvalidation_0-merror:0.096383\tvalidation_1-merror:0.095752\n",
            "[446]\tvalidation_0-merror:0.096377\tvalidation_1-merror:0.095752\n",
            "[447]\tvalidation_0-merror:0.09631\tvalidation_1-merror:0.095645\n",
            "[448]\tvalidation_0-merror:0.096109\tvalidation_1-merror:0.095548\n",
            "[449]\tvalidation_0-merror:0.096048\tvalidation_1-merror:0.095519\n",
            "[450]\tvalidation_0-merror:0.09598\tvalidation_1-merror:0.095441\n",
            "[451]\tvalidation_0-merror:0.095944\tvalidation_1-merror:0.09543\n",
            "[452]\tvalidation_0-merror:0.095816\tvalidation_1-merror:0.09532\n",
            "[453]\tvalidation_0-merror:0.095901\tvalidation_1-merror:0.095375\n",
            "[454]\tvalidation_0-merror:0.095919\tvalidation_1-merror:0.095388\n",
            "[455]\tvalidation_0-merror:0.095834\tvalidation_1-merror:0.095284\n",
            "[456]\tvalidation_0-merror:0.095938\tvalidation_1-merror:0.095297\n",
            "[457]\tvalidation_0-merror:0.095858\tvalidation_1-merror:0.09515\n",
            "[458]\tvalidation_0-merror:0.095858\tvalidation_1-merror:0.095145\n",
            "[459]\tvalidation_0-merror:0.095864\tvalidation_1-merror:0.095148\n",
            "[460]\tvalidation_0-merror:0.095895\tvalidation_1-merror:0.095111\n",
            "[461]\tvalidation_0-merror:0.095486\tvalidation_1-merror:0.09487\n",
            "[462]\tvalidation_0-merror:0.09548\tvalidation_1-merror:0.094865\n",
            "[463]\tvalidation_0-merror:0.09548\tvalidation_1-merror:0.094857\n",
            "[464]\tvalidation_0-merror:0.095449\tvalidation_1-merror:0.094813\n",
            "[465]\tvalidation_0-merror:0.095681\tvalidation_1-merror:0.094884\n",
            "[466]\tvalidation_0-merror:0.095651\tvalidation_1-merror:0.094821\n",
            "[467]\tvalidation_0-merror:0.095535\tvalidation_1-merror:0.094719\n",
            "[468]\tvalidation_0-merror:0.095504\tvalidation_1-merror:0.094672\n",
            "[469]\tvalidation_0-merror:0.095474\tvalidation_1-merror:0.094651\n",
            "[470]\tvalidation_0-merror:0.095394\tvalidation_1-merror:0.094546\n",
            "[471]\tvalidation_0-merror:0.095364\tvalidation_1-merror:0.094543\n",
            "[472]\tvalidation_0-merror:0.095364\tvalidation_1-merror:0.09453\n",
            "[473]\tvalidation_0-merror:0.095388\tvalidation_1-merror:0.094541\n",
            "[474]\tvalidation_0-merror:0.095346\tvalidation_1-merror:0.094489\n",
            "[475]\tvalidation_0-merror:0.095248\tvalidation_1-merror:0.094421\n",
            "[476]\tvalidation_0-merror:0.095248\tvalidation_1-merror:0.094421\n",
            "[477]\tvalidation_0-merror:0.095248\tvalidation_1-merror:0.094418\n",
            "[478]\tvalidation_0-merror:0.095242\tvalidation_1-merror:0.094405\n",
            "[479]\tvalidation_0-merror:0.095199\tvalidation_1-merror:0.094353\n",
            "[480]\tvalidation_0-merror:0.095181\tvalidation_1-merror:0.094329\n",
            "[481]\tvalidation_0-merror:0.095181\tvalidation_1-merror:0.094326\n",
            "[482]\tvalidation_0-merror:0.095156\tvalidation_1-merror:0.094292\n",
            "[483]\tvalidation_0-merror:0.095132\tvalidation_1-merror:0.094295\n",
            "[484]\tvalidation_0-merror:0.095126\tvalidation_1-merror:0.09429\n",
            "[485]\tvalidation_0-merror:0.094894\tvalidation_1-merror:0.094172\n",
            "[486]\tvalidation_0-merror:0.094802\tvalidation_1-merror:0.094036\n",
            "[487]\tvalidation_0-merror:0.094808\tvalidation_1-merror:0.094039\n",
            "[488]\tvalidation_0-merror:0.094808\tvalidation_1-merror:0.094052\n",
            "[489]\tvalidation_0-merror:0.094802\tvalidation_1-merror:0.094054\n",
            "[490]\tvalidation_0-merror:0.094802\tvalidation_1-merror:0.094054\n",
            "[491]\tvalidation_0-merror:0.094784\tvalidation_1-merror:0.094028\n",
            "[492]\tvalidation_0-merror:0.094754\tvalidation_1-merror:0.094012\n",
            "[493]\tvalidation_0-merror:0.094729\tvalidation_1-merror:0.094015\n",
            "[494]\tvalidation_0-merror:0.094638\tvalidation_1-merror:0.093918\n",
            "[495]\tvalidation_0-merror:0.09465\tvalidation_1-merror:0.093921\n",
            "[496]\tvalidation_0-merror:0.094644\tvalidation_1-merror:0.09391\n",
            "[497]\tvalidation_0-merror:0.09465\tvalidation_1-merror:0.093929\n",
            "[498]\tvalidation_0-merror:0.094656\tvalidation_1-merror:0.093918\n",
            "[499]\tvalidation_0-merror:0.094522\tvalidation_1-merror:0.093842\n",
            "[500]\tvalidation_0-merror:0.09468\tvalidation_1-merror:0.093947\n",
            "[501]\tvalidation_0-merror:0.094686\tvalidation_1-merror:0.093947\n",
            "[502]\tvalidation_0-merror:0.094418\tvalidation_1-merror:0.093769\n",
            "[503]\tvalidation_0-merror:0.094277\tvalidation_1-merror:0.093602\n",
            "[504]\tvalidation_0-merror:0.093966\tvalidation_1-merror:0.093212\n",
            "[505]\tvalidation_0-merror:0.093691\tvalidation_1-merror:0.092963\n",
            "[506]\tvalidation_0-merror:0.093789\tvalidation_1-merror:0.093037\n",
            "[507]\tvalidation_0-merror:0.093722\tvalidation_1-merror:0.092961\n",
            "[508]\tvalidation_0-merror:0.093722\tvalidation_1-merror:0.092956\n",
            "[509]\tvalidation_0-merror:0.093716\tvalidation_1-merror:0.092958\n",
            "[510]\tvalidation_0-merror:0.093698\tvalidation_1-merror:0.092953\n",
            "[511]\tvalidation_0-merror:0.093704\tvalidation_1-merror:0.092963\n",
            "[512]\tvalidation_0-merror:0.093691\tvalidation_1-merror:0.092963\n",
            "[513]\tvalidation_0-merror:0.093679\tvalidation_1-merror:0.09295\n",
            "[514]\tvalidation_0-merror:0.093679\tvalidation_1-merror:0.09294\n",
            "[515]\tvalidation_0-merror:0.093679\tvalidation_1-merror:0.092961\n",
            "[516]\tvalidation_0-merror:0.093667\tvalidation_1-merror:0.092956\n",
            "[517]\tvalidation_0-merror:0.093612\tvalidation_1-merror:0.092916\n",
            "[518]\tvalidation_0-merror:0.093594\tvalidation_1-merror:0.09288\n",
            "[519]\tvalidation_0-merror:0.093594\tvalidation_1-merror:0.092882\n",
            "[520]\tvalidation_0-merror:0.093594\tvalidation_1-merror:0.092888\n",
            "[521]\tvalidation_0-merror:0.093612\tvalidation_1-merror:0.092901\n",
            "[522]\tvalidation_0-merror:0.093618\tvalidation_1-merror:0.092906\n",
            "[523]\tvalidation_0-merror:0.093606\tvalidation_1-merror:0.092888\n",
            "[524]\tvalidation_0-merror:0.093606\tvalidation_1-merror:0.092893\n",
            "[525]\tvalidation_0-merror:0.0936\tvalidation_1-merror:0.09289\n",
            "[526]\tvalidation_0-merror:0.0936\tvalidation_1-merror:0.09289\n",
            "[527]\tvalidation_0-merror:0.093606\tvalidation_1-merror:0.092895\n",
            "[528]\tvalidation_0-merror:0.093612\tvalidation_1-merror:0.092893\n",
            "[529]\tvalidation_0-merror:0.093582\tvalidation_1-merror:0.092859\n",
            "[530]\tvalidation_0-merror:0.093569\tvalidation_1-merror:0.092843\n",
            "[531]\tvalidation_0-merror:0.093569\tvalidation_1-merror:0.092843\n",
            "[532]\tvalidation_0-merror:0.093563\tvalidation_1-merror:0.092843\n",
            "[533]\tvalidation_0-merror:0.093569\tvalidation_1-merror:0.092846\n",
            "[534]\tvalidation_0-merror:0.093575\tvalidation_1-merror:0.092843\n",
            "[535]\tvalidation_0-merror:0.093569\tvalidation_1-merror:0.092833\n",
            "[536]\tvalidation_0-merror:0.093569\tvalidation_1-merror:0.092838\n",
            "[537]\tvalidation_0-merror:0.093575\tvalidation_1-merror:0.092846\n",
            "[538]\tvalidation_0-merror:0.093575\tvalidation_1-merror:0.092846\n",
            "[539]\tvalidation_0-merror:0.093569\tvalidation_1-merror:0.092846\n",
            "[540]\tvalidation_0-merror:0.093569\tvalidation_1-merror:0.092843\n",
            "[541]\tvalidation_0-merror:0.093582\tvalidation_1-merror:0.092851\n",
            "[542]\tvalidation_0-merror:0.093575\tvalidation_1-merror:0.09284\n",
            "[543]\tvalidation_0-merror:0.093575\tvalidation_1-merror:0.092838\n",
            "[544]\tvalidation_0-merror:0.093575\tvalidation_1-merror:0.092838\n",
            "[545]\tvalidation_0-merror:0.093575\tvalidation_1-merror:0.092835\n",
            "[546]\tvalidation_0-merror:0.093575\tvalidation_1-merror:0.092835\n",
            "[547]\tvalidation_0-merror:0.093588\tvalidation_1-merror:0.09284\n",
            "[548]\tvalidation_0-merror:0.093588\tvalidation_1-merror:0.09284\n",
            "[549]\tvalidation_0-merror:0.093588\tvalidation_1-merror:0.09284\n",
            "[550]\tvalidation_0-merror:0.093588\tvalidation_1-merror:0.09284\n",
            "[551]\tvalidation_0-merror:0.093588\tvalidation_1-merror:0.09284\n",
            "[552]\tvalidation_0-merror:0.093588\tvalidation_1-merror:0.09284\n",
            "[553]\tvalidation_0-merror:0.093594\tvalidation_1-merror:0.09284\n",
            "[554]\tvalidation_0-merror:0.093594\tvalidation_1-merror:0.09284\n",
            "[555]\tvalidation_0-merror:0.093582\tvalidation_1-merror:0.092827\n",
            "[556]\tvalidation_0-merror:0.093582\tvalidation_1-merror:0.092817\n",
            "[557]\tvalidation_0-merror:0.093575\tvalidation_1-merror:0.092817\n",
            "[558]\tvalidation_0-merror:0.093575\tvalidation_1-merror:0.092809\n",
            "[559]\tvalidation_0-merror:0.093575\tvalidation_1-merror:0.092809\n",
            "[560]\tvalidation_0-merror:0.093575\tvalidation_1-merror:0.092806\n",
            "[561]\tvalidation_0-merror:0.093594\tvalidation_1-merror:0.092812\n",
            "[562]\tvalidation_0-merror:0.093594\tvalidation_1-merror:0.092812\n",
            "[563]\tvalidation_0-merror:0.093594\tvalidation_1-merror:0.092812\n",
            "[564]\tvalidation_0-merror:0.093594\tvalidation_1-merror:0.092812\n",
            "[565]\tvalidation_0-merror:0.093588\tvalidation_1-merror:0.092809\n",
            "[566]\tvalidation_0-merror:0.093582\tvalidation_1-merror:0.092809\n",
            "[567]\tvalidation_0-merror:0.093588\tvalidation_1-merror:0.092814\n",
            "[568]\tvalidation_0-merror:0.093594\tvalidation_1-merror:0.09282\n",
            "[569]\tvalidation_0-merror:0.093594\tvalidation_1-merror:0.092822\n",
            "[570]\tvalidation_0-merror:0.093594\tvalidation_1-merror:0.092822\n",
            "[571]\tvalidation_0-merror:0.093594\tvalidation_1-merror:0.09282\n",
            "[572]\tvalidation_0-merror:0.093588\tvalidation_1-merror:0.092817\n",
            "[573]\tvalidation_0-merror:0.093588\tvalidation_1-merror:0.09282\n",
            "[574]\tvalidation_0-merror:0.093588\tvalidation_1-merror:0.092822\n",
            "[575]\tvalidation_0-merror:0.093588\tvalidation_1-merror:0.092822\n",
            "[576]\tvalidation_0-merror:0.093588\tvalidation_1-merror:0.092825\n",
            "[577]\tvalidation_0-merror:0.093588\tvalidation_1-merror:0.092825\n",
            "[578]\tvalidation_0-merror:0.093588\tvalidation_1-merror:0.092825\n",
            "[579]\tvalidation_0-merror:0.093588\tvalidation_1-merror:0.092822\n",
            "[580]\tvalidation_0-merror:0.093588\tvalidation_1-merror:0.092822\n",
            "[581]\tvalidation_0-merror:0.093588\tvalidation_1-merror:0.092817\n",
            "[582]\tvalidation_0-merror:0.09363\tvalidation_1-merror:0.092856\n",
            "[583]\tvalidation_0-merror:0.09363\tvalidation_1-merror:0.092856\n",
            "[584]\tvalidation_0-merror:0.09363\tvalidation_1-merror:0.092856\n",
            "[585]\tvalidation_0-merror:0.09363\tvalidation_1-merror:0.092859\n",
            "[586]\tvalidation_0-merror:0.09363\tvalidation_1-merror:0.092859\n",
            "[587]\tvalidation_0-merror:0.09363\tvalidation_1-merror:0.092859\n",
            "[588]\tvalidation_0-merror:0.09363\tvalidation_1-merror:0.092859\n",
            "[589]\tvalidation_0-merror:0.09363\tvalidation_1-merror:0.092859\n",
            "[590]\tvalidation_0-merror:0.09363\tvalidation_1-merror:0.092859\n",
            "[591]\tvalidation_0-merror:0.09363\tvalidation_1-merror:0.092859\n",
            "[592]\tvalidation_0-merror:0.09363\tvalidation_1-merror:0.092859\n",
            "[593]\tvalidation_0-merror:0.09363\tvalidation_1-merror:0.092859\n",
            "[594]\tvalidation_0-merror:0.09363\tvalidation_1-merror:0.092859\n",
            "[595]\tvalidation_0-merror:0.09363\tvalidation_1-merror:0.092859\n",
            "[596]\tvalidation_0-merror:0.09363\tvalidation_1-merror:0.092859\n",
            "[597]\tvalidation_0-merror:0.093637\tvalidation_1-merror:0.092864\n",
            "[598]\tvalidation_0-merror:0.093637\tvalidation_1-merror:0.092864\n",
            "[599]\tvalidation_0-merror:0.093637\tvalidation_1-merror:0.092864\n",
            "[600]\tvalidation_0-merror:0.093643\tvalidation_1-merror:0.092867\n",
            "[601]\tvalidation_0-merror:0.093643\tvalidation_1-merror:0.092867\n",
            "[602]\tvalidation_0-merror:0.093637\tvalidation_1-merror:0.092872\n",
            "[603]\tvalidation_0-merror:0.093637\tvalidation_1-merror:0.092872\n",
            "[604]\tvalidation_0-merror:0.093637\tvalidation_1-merror:0.092869\n",
            "[605]\tvalidation_0-merror:0.093637\tvalidation_1-merror:0.092869\n",
            "[606]\tvalidation_0-merror:0.093649\tvalidation_1-merror:0.092877\n",
            "[607]\tvalidation_0-merror:0.093649\tvalidation_1-merror:0.09288\n",
            "[608]\tvalidation_0-merror:0.093649\tvalidation_1-merror:0.09288\n",
            "[609]\tvalidation_0-merror:0.093649\tvalidation_1-merror:0.09288\n",
            "[610]\tvalidation_0-merror:0.093649\tvalidation_1-merror:0.09288\n",
            "[611]\tvalidation_0-merror:0.093649\tvalidation_1-merror:0.09288\n",
            "[612]\tvalidation_0-merror:0.093649\tvalidation_1-merror:0.09288\n",
            "[613]\tvalidation_0-merror:0.093649\tvalidation_1-merror:0.09288\n",
            "[614]\tvalidation_0-merror:0.093649\tvalidation_1-merror:0.09288\n",
            "[615]\tvalidation_0-merror:0.093649\tvalidation_1-merror:0.09288\n",
            "[616]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092882\n",
            "[617]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092882\n",
            "[618]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092882\n",
            "[619]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092882\n",
            "[620]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092882\n",
            "[621]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092882\n",
            "[622]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092882\n",
            "[623]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092882\n",
            "[624]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092882\n",
            "[625]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092882\n",
            "[626]\tvalidation_0-merror:0.093649\tvalidation_1-merror:0.092885\n",
            "[627]\tvalidation_0-merror:0.093649\tvalidation_1-merror:0.092885\n",
            "[628]\tvalidation_0-merror:0.093649\tvalidation_1-merror:0.092885\n",
            "[629]\tvalidation_0-merror:0.093649\tvalidation_1-merror:0.092885\n",
            "[630]\tvalidation_0-merror:0.093649\tvalidation_1-merror:0.092885\n",
            "[631]\tvalidation_0-merror:0.093649\tvalidation_1-merror:0.092885\n",
            "[632]\tvalidation_0-merror:0.093649\tvalidation_1-merror:0.092885\n",
            "[633]\tvalidation_0-merror:0.093649\tvalidation_1-merror:0.092885\n",
            "[634]\tvalidation_0-merror:0.093649\tvalidation_1-merror:0.092885\n",
            "[635]\tvalidation_0-merror:0.093649\tvalidation_1-merror:0.092885\n",
            "[636]\tvalidation_0-merror:0.093649\tvalidation_1-merror:0.092885\n",
            "[637]\tvalidation_0-merror:0.093649\tvalidation_1-merror:0.092885\n",
            "[638]\tvalidation_0-merror:0.093649\tvalidation_1-merror:0.092885\n",
            "[639]\tvalidation_0-merror:0.093649\tvalidation_1-merror:0.092885\n",
            "[640]\tvalidation_0-merror:0.093649\tvalidation_1-merror:0.092885\n",
            "[641]\tvalidation_0-merror:0.093643\tvalidation_1-merror:0.09288\n",
            "[642]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092911\n",
            "[643]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092911\n",
            "[644]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092911\n",
            "[645]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092911\n",
            "[646]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092911\n",
            "[647]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092911\n",
            "[648]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092911\n",
            "[649]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092911\n",
            "[650]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092911\n",
            "[651]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092911\n",
            "[652]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092911\n",
            "[653]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092911\n",
            "[654]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092908\n",
            "[655]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092908\n",
            "[656]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092908\n",
            "[657]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092908\n",
            "[658]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092908\n",
            "[659]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092908\n",
            "[660]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092908\n",
            "[661]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092908\n",
            "[662]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092908\n",
            "[663]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092906\n",
            "[664]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092906\n",
            "[665]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092906\n",
            "[666]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092906\n",
            "[667]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092906\n",
            "[668]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092906\n",
            "[669]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092906\n",
            "[670]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092906\n",
            "[671]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092906\n",
            "[672]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092906\n",
            "[673]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092906\n",
            "[674]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092906\n",
            "[675]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092906\n",
            "[676]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092903\n",
            "[677]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[678]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[679]\tvalidation_0-merror:0.093649\tvalidation_1-merror:0.092901\n",
            "[680]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[681]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[682]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092903\n",
            "[683]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[684]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[685]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[686]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092903\n",
            "[687]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[688]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[689]\tvalidation_0-merror:0.093649\tvalidation_1-merror:0.092901\n",
            "[690]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[691]\tvalidation_0-merror:0.093643\tvalidation_1-merror:0.092898\n",
            "[692]\tvalidation_0-merror:0.093649\tvalidation_1-merror:0.092898\n",
            "[693]\tvalidation_0-merror:0.093643\tvalidation_1-merror:0.092898\n",
            "[694]\tvalidation_0-merror:0.093643\tvalidation_1-merror:0.092898\n",
            "[695]\tvalidation_0-merror:0.093643\tvalidation_1-merror:0.092898\n",
            "[696]\tvalidation_0-merror:0.093649\tvalidation_1-merror:0.092903\n",
            "[697]\tvalidation_0-merror:0.093649\tvalidation_1-merror:0.092903\n",
            "[698]\tvalidation_0-merror:0.093649\tvalidation_1-merror:0.092903\n",
            "[699]\tvalidation_0-merror:0.093649\tvalidation_1-merror:0.092903\n",
            "[700]\tvalidation_0-merror:0.093649\tvalidation_1-merror:0.092903\n",
            "[701]\tvalidation_0-merror:0.093649\tvalidation_1-merror:0.092903\n",
            "[702]\tvalidation_0-merror:0.093649\tvalidation_1-merror:0.092903\n",
            "[703]\tvalidation_0-merror:0.093649\tvalidation_1-merror:0.092903\n",
            "[704]\tvalidation_0-merror:0.093649\tvalidation_1-merror:0.092903\n",
            "[705]\tvalidation_0-merror:0.093649\tvalidation_1-merror:0.092903\n",
            "[706]\tvalidation_0-merror:0.093649\tvalidation_1-merror:0.092903\n",
            "[707]\tvalidation_0-merror:0.093649\tvalidation_1-merror:0.092903\n",
            "[708]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[709]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[710]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[711]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[712]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[713]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[714]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[715]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[716]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[717]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[718]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[719]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[720]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[721]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[722]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[723]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[724]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[725]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[726]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[727]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[728]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[729]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[730]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[731]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[732]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[733]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[734]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[735]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[736]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[737]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[738]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[739]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[740]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[741]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[742]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092903\n",
            "[743]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092914\n",
            "[744]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092914\n",
            "[745]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092914\n",
            "[746]\tvalidation_0-merror:0.093655\tvalidation_1-merror:0.092914\n",
            "[747]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092914\n",
            "[748]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092914\n",
            "[749]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092914\n",
            "[750]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092914\n",
            "[751]\tvalidation_0-merror:0.093661\tvalidation_1-merror:0.092914\n",
            "[752]\tvalidation_0-merror:0.093667\tvalidation_1-merror:0.092914\n",
            "[753]\tvalidation_0-merror:0.093667\tvalidation_1-merror:0.092914\n",
            "[754]\tvalidation_0-merror:0.093667\tvalidation_1-merror:0.092914\n",
            "[755]\tvalidation_0-merror:0.093667\tvalidation_1-merror:0.092914\n",
            "[756]\tvalidation_0-merror:0.093667\tvalidation_1-merror:0.092914\n",
            "[757]\tvalidation_0-merror:0.093667\tvalidation_1-merror:0.092914\n",
            "[758]\tvalidation_0-merror:0.093667\tvalidation_1-merror:0.092914\n",
            "[759]\tvalidation_0-merror:0.093667\tvalidation_1-merror:0.092914\n",
            "[760]\tvalidation_0-merror:0.093667\tvalidation_1-merror:0.092914\n",
            "[761]\tvalidation_0-merror:0.093667\tvalidation_1-merror:0.092916\n",
            "[762]\tvalidation_0-merror:0.093667\tvalidation_1-merror:0.092916\n",
            "[763]\tvalidation_0-merror:0.093667\tvalidation_1-merror:0.092916\n",
            "[764]\tvalidation_0-merror:0.093667\tvalidation_1-merror:0.092916\n",
            "[765]\tvalidation_0-merror:0.093667\tvalidation_1-merror:0.092916\n",
            "[766]\tvalidation_0-merror:0.093667\tvalidation_1-merror:0.092916\n",
            "[767]\tvalidation_0-merror:0.093667\tvalidation_1-merror:0.092916\n",
            "[768]\tvalidation_0-merror:0.093667\tvalidation_1-merror:0.092916\n",
            "[769]\tvalidation_0-merror:0.093667\tvalidation_1-merror:0.092916\n",
            "[770]\tvalidation_0-merror:0.093667\tvalidation_1-merror:0.092916\n",
            "[771]\tvalidation_0-merror:0.093667\tvalidation_1-merror:0.092916\n",
            "[772]\tvalidation_0-merror:0.093667\tvalidation_1-merror:0.092916\n",
            "[773]\tvalidation_0-merror:0.093667\tvalidation_1-merror:0.092916\n",
            "[774]\tvalidation_0-merror:0.093667\tvalidation_1-merror:0.092916\n",
            "[775]\tvalidation_0-merror:0.093679\tvalidation_1-merror:0.092922\n",
            "[776]\tvalidation_0-merror:0.093679\tvalidation_1-merror:0.092922\n",
            "[777]\tvalidation_0-merror:0.093679\tvalidation_1-merror:0.092922\n",
            "[778]\tvalidation_0-merror:0.093679\tvalidation_1-merror:0.092922\n",
            "[779]\tvalidation_0-merror:0.093679\tvalidation_1-merror:0.092922\n",
            "[780]\tvalidation_0-merror:0.093679\tvalidation_1-merror:0.092922\n",
            "[781]\tvalidation_0-merror:0.093679\tvalidation_1-merror:0.092922\n",
            "[782]\tvalidation_0-merror:0.093679\tvalidation_1-merror:0.092922\n",
            "[783]\tvalidation_0-merror:0.093679\tvalidation_1-merror:0.092922\n",
            "[784]\tvalidation_0-merror:0.093679\tvalidation_1-merror:0.092922\n",
            "[785]\tvalidation_0-merror:0.093679\tvalidation_1-merror:0.092922\n",
            "[786]\tvalidation_0-merror:0.093679\tvalidation_1-merror:0.092922\n",
            "[787]\tvalidation_0-merror:0.093679\tvalidation_1-merror:0.092922\n",
            "[788]\tvalidation_0-merror:0.093679\tvalidation_1-merror:0.092922\n",
            "[789]\tvalidation_0-merror:0.093679\tvalidation_1-merror:0.092922\n",
            "[790]\tvalidation_0-merror:0.093679\tvalidation_1-merror:0.092922\n",
            "[791]\tvalidation_0-merror:0.093679\tvalidation_1-merror:0.092922\n",
            "[792]\tvalidation_0-merror:0.093679\tvalidation_1-merror:0.092922\n",
            "[793]\tvalidation_0-merror:0.093679\tvalidation_1-merror:0.092922\n",
            "[794]\tvalidation_0-merror:0.093679\tvalidation_1-merror:0.092922\n",
            "[795]\tvalidation_0-merror:0.093679\tvalidation_1-merror:0.092922\n",
            "[796]\tvalidation_0-merror:0.093679\tvalidation_1-merror:0.092922\n",
            "[797]\tvalidation_0-merror:0.093679\tvalidation_1-merror:0.092922\n",
            "[798]\tvalidation_0-merror:0.093679\tvalidation_1-merror:0.092922\n",
            "[799]\tvalidation_0-merror:0.093679\tvalidation_1-merror:0.092922\n",
            "[800]\tvalidation_0-merror:0.093679\tvalidation_1-merror:0.092922\n",
            "[801]\tvalidation_0-merror:0.093679\tvalidation_1-merror:0.092922\n",
            "[802]\tvalidation_0-merror:0.093679\tvalidation_1-merror:0.092922\n",
            "[803]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[804]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[805]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[806]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[807]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[808]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[809]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[810]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[811]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[812]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[813]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[814]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[815]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[816]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[817]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[818]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[819]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[820]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[821]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[822]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[823]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[824]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[825]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[826]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[827]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[828]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[829]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[830]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[831]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[832]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[833]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[834]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[835]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[836]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[837]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[838]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[839]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[840]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[841]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[842]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[843]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[844]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[845]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[846]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[847]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[848]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[849]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[850]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[851]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[852]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[853]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[854]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[855]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[856]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[857]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[858]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[859]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[860]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[861]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[862]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[863]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[864]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[865]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[866]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[867]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[868]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[869]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[870]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[871]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[872]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[873]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[874]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[875]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[876]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[877]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[878]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[879]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[880]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[881]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[882]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[883]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[884]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[885]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[886]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[887]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[888]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[889]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[890]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[891]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[892]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[893]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[894]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[895]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[896]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[897]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[898]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[899]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[900]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[901]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[902]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[903]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[904]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[905]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[906]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[907]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[908]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[909]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[910]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[911]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[912]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[913]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[914]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[915]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[916]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[917]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[918]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[919]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[920]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[921]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[922]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[923]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[924]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[925]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[926]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[927]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[928]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[929]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[930]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[931]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[932]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[933]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[934]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[935]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[936]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[937]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[938]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[939]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[940]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[941]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[942]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[943]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[944]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[945]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[946]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[947]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[948]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[949]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[950]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[951]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[952]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[953]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[954]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[955]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[956]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[957]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[958]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[959]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[960]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[961]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[962]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[963]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[964]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[965]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[966]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[967]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[968]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[969]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[970]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[971]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[972]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[973]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[974]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[975]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[976]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[977]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[978]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[979]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[980]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[981]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[982]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[983]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[984]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[985]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[986]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[987]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[988]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[989]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[990]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[991]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[992]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[993]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[994]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[995]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[996]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[997]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[998]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n",
            "[999]\tvalidation_0-merror:0.093685\tvalidation_1-merror:0.092927\n"
          ]
        },
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "X_train, X_validation = train_test_split(poly_data, train_size=0.7, random_state=0)\n",
        "X_train = pd.DataFrame(X_train)\n",
        "X_validation = pd.DataFrame(X_validation)\n",
        "X_train, y_train = X_train.drop(labels=[col for col in X_train.columns if str(col).find(\"gt\")!=-1], axis=1), X_train[\"gt\"]\n",
        "X_validation, y_validation = X_validation.drop(labels=[col for col in X_validation.columns if str(col).find(\"gt\")!=-1], axis=1), X_validation[\"gt\"]\n",
        "\n",
        "evallist = [(X_validation, y_validation), (X_train, y_train)]\n",
        "\n",
        "print(\"Parameter optimization\")\n",
        "params = { \n",
        "    'eta': 0.1,\n",
        "    'gamma': 0.1,\n",
        "    'maxDepth': 10, \n",
        "    'lambda_': 1.0,\n",
        "    \"n_estimators\": 1000,\n",
        "    'subsample': 1.0,\n",
        "    'numRound': 100,\n",
        "    'numWorkers': 1,\n",
        "    'verbosity': 1\n",
        "}\n",
        "xgb_model = xgb.XGBClassifier(**params)\n",
        "\n",
        "xgb_model.fit(X=X_train, \n",
        "              y=y_train, \n",
        "              eval_set=evallist,\n",
        "              early_stopping_rounds=10)  # merror = 1 - accuracy\n",
        "\n",
        "plot_importance(xgb_model, max_num_features=14) # top 10 most important features\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "zX4DPfsLYr9a",
        "outputId": "ceed059e-6a6d-45f4-cccb-60a264330df6"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "array([0.05268458, 0.00998434, 0.00945802, 0.08155405, 0.        ,\n",
              "       0.04686033, 0.0134109 , 0.00784575, 0.00904694, 0.01885984,\n",
              "       0.00291689, 0.16654849, 0.        , 0.008858  , 0.02953519,\n",
              "       0.01867567, 0.03194933, 0.01048754, 0.10777926, 0.        ,\n",
              "       0.01064978, 0.05821737, 0.00670422, 0.        , 0.00498638,\n",
              "       0.        , 0.01101717, 0.00240076, 0.00589516, 0.01581816,\n",
              "       0.        , 0.01174534, 0.03120974, 0.01312835, 0.02842448,\n",
              "       0.        , 0.        , 0.        , 0.        , 0.03983697,\n",
              "       0.00866731, 0.0232415 , 0.0222871 , 0.04596944, 0.03334566],\n",
              "      dtype=float32)"
            ]
          },
          "execution_count": 32,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "xgb_model.feature_importances_"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "7W8jOgXr0E8c",
        "outputId": "612be2d7-57af-4294-dd66-03593ad61118"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "import matplotlib.pylab as plt\n",
        "from matplotlib import pyplot\n",
        "from xgboost import plot_importance\n",
        "plot_importance(xgb_model, max_num_features=10)\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "r4OBUB0_0Vxa",
        "outputId": "2c8d6b65-ed13-4c6e-b2a8-b9d56c6d7dce"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "plot_importance(xgb_model, max_num_features=15)\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "2QTArJph0XYx",
        "outputId": "76606e47-cf44-432d-b1a6-a0cbf6239700"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "plot_importance(xgb_model, max_num_features=20)\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "AkRJ8YFW0buh",
        "outputId": "51ea6bf6-3abc-4cd5-a58c-fb9e18b1ea41"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "plot_importance(xgb_model, max_num_features=25)\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "euDCvw2f0dTh",
        "outputId": "da450561-5f10-4102-acfe-e00c84924f94"
      },
      "outputs": [
        {
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAf0AAAEWCAYAAABsT07JAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOydd3hVxdaH30UNRRCkGkqUGilGULhYIIjCtaKIIupFiF4L6sWCBfUqooANKSKCoqBIEVTAa0EUiKB8lyYBpF9NkI6EIgECCVnfHzMn7BxOkpPKIZn3efaTvWdmz6x98HH2nlm/tURVcTgcDofDUfQpcboNcDgcDofDUTi4Sd/hcDgcjmKCm/QdDofD4SgmuEnf4XA4HI5igpv0HQ6Hw+EoJrhJ3+FwOByOYoKb9B0OhwMQkWdFZPzptsPhKEjE6fQdDkdeEZEEoCZwwlPcWFV35LHPe1X1h7xZd+YhIgOBhqp61+m2xVG0cF/6Docjv7hBVSt6jlxP+PmBiJQ6nePnljPVbseZgZv0HQ5HgSEilUXkAxHZKSLbReQVESlp6xqIyHwRSRSRvSIyWUTOtnWTgHrAf0QkSUSeEpFoEdnm13+CiFxlzweKyGci8omI/AX0zmr8ALYOFJFP7HmEiKiI9BGRrSKyX0QeEJFLRGS1iBwQkdGee3uLyM8iMlpEDorIBhHp5Kk/V0S+FJF9IvI/Efmn37heux8AngV62GdfZdv1EZH1InJIRH4Xkfs9fUSLyDYReUJE9tjn7eOpLyciw0Rki7XvJxEpZ+v+JiKL7TOtEpHoXP1jO84I3KTvcDgKkolAKtAQuAjoDNxr6wQYCpwLRAJ1gYEAqvoP4A9Orh68HuR4XYHPgLOBydmMHwxtgUZAD2AE8BxwFdAMuE1EOvi1/Q2oBrwIfCEiVW3dNGCbfdbuwBARuTITuz8AhgCf2me/0LbZA1wPVAL6AMNFpJWnj1pAZSAcuAd4R0Sq2Lo3gdbApUBV4CkgTUTCga+BV2x5f+BzEameg9/IcQbhJn2Hw5FfzLJfiwdEZJaI1ASuBR5V1cOqugcYDtwOoKr/U9XvVfWYqv4JvAV0yLz7oPg/VZ2lqmmYyTHT8YPkZVVNVtW5wGFgqqruUdXtwCLMi4SPPcAIVU1R1U+BjcB1IlIXuAx42vYVB4wHegWyW1WPBjJEVb9W1d/U8CMwF7jC0yQFGGTH/wZIApqISAkgBuinqttV9YSqLlbVY8BdwDeq+o0d+3tguf3dHEUQt3fkcDjyi5u8Tnci0gYoDewUEV9xCWCrra8JjMRMXGfZuv15tGGr57x+VuMHyW7P+dEA1xU919s1o2f0FsyX/bnAPlU95Fd3cSZ2B0RErsGsIDTGPEd5YI2nSaKqpnquj1j7qgFhmFUIf+oDt4rIDZ6y0sCC7OxxnJm4Sd/hcBQUW4FjQDW/ycjHEECBFqq6T0RuAkZ76v2lRYcxEx0Adm/efxnae0924+c34SIinom/HvAlsAOoKiJneSb+esB2z73+z5rhWkTKAp9jVgdmq2qKiMzCbJFkx14gGWgArPKr2wpMUtV/nnKXo0jilvcdDkeBoKo7MUvQw0SkkoiUsM57viX8szBL0Aft3vKTfl3sBs73XG8CwkTkOhEpDTwPlM3D+PlNDeBfIlJaRG7F+Cl8o6pbgcXAUBEJE5GWmD33T7LoazcQYZfmAcpgnvVPINV+9XcOxii71fEh8JZ1KCwpIu3si8QnwA0i0sWWh1mnwDo5f3zHmYCb9B0OR0HSCzNhrcMs3X8G1LZ1LwGtgIMYZ7Iv/O4dCjxvfQT6q+pBoC9mP3w75st/G1mT1fj5zRKM099eYDDQXVUTbV1PIALz1T8TeDGb+AMz7N9EEfnFrhD8C5iOeY47MKsIwdIfsxWwDNgHvAaUsC8kXTFqgT8xX/5P4uaGIosLzuNwOBx5RER6YwIJXX66bXE4ssK9zTkcDofDUUxwk77D4XA4HMUEt7zvcDgcDkcxwX3pOxwOh8NRTHA6fUfIcfbZZ2vDhg1PtxmZcvjwYSpUqHC6zciUULcPQt9GZ1/eCHX7IPRtzI19K1as2KuqWYZQdpO+I+SoWbMmy5cvP91mZEpsbCzR0dGn24xMCXX7IPRtdPbljVC3D0LfxtzYJyJbsmvjlvcdDofD4SgmuEnf4XA4HI58JCYmhho1atC8efP0soEDBxIeHk5UVBRRUVF88803ACQmJtKxY0cqVqzIww8/nN4+OTmZ6667jqZNm9KsWTOeeeaZ9LotW7bQqVMnWrZsSXR0NNu2ZRej6iRu0nc4HA6HIx/p3bs3c+bMOaX8scceIy4ujri4OK691iQyDAsL4+WXX+bNN988pX3//v3ZsGEDK1eu5Oeff+bbb79NL+/VqxerV6/mhRdeYMCAAUHb5ib9EEFEbhIRFZGmObhncS7HihCRXzOp6yIicfZIEpGN9vxjEXlARHoFus/hcDgchvbt21O1atWg2laoUIHLL7+csLCwDOVhYWF07NgRgDJlytCqVav0L/p169Zx5ZVXAtCxY0dmz54dtG1u0g8degI/2b8ZEJFSga5V9dL8NkJVv1PVKFWNwuTVvtNe91LVsar6cX6P6XA4HMWB0aNH07JlS2JiYti/P/gs0gcOHOA///kPnTp1AuDCCy/kiy9MqoqZM2dy6NAhEhMTs+oiHRecJwQQkYrARqAj8B9VbSIi0cDLmOQaTYH7vNeq2lhEklS1oohMw6TH/Nr2NxH4CjNpTwJ8uo+HVXWxiEQAX6nqyQ2nwHbFAv1Vdbm9Hggkqeqbtm4lJhd6BUxikwFAC+BTVX3e3nMXJlFIGUxCkr6qeiLAWPfZZ6RateqtXxjxfrA/X6FTsxzsPnq6rcicULcPQt9GZ1/eCHX7oOBsbBFeGYBdu3YxYMAAJkyYAMC+ffuoXLkyIsKHH35IYmIiTz/9dPp9c+bMYePGjfTr1w+ApKQkKlasyIkTJ3j22We55JJL6N69OwB79+5l1KhR7Ny5k5YtW7Jw4UImTJjADTfcsEJVL87SQFV1x2k+gDuBD+z5YqA1EI3JInaeLc9wbcuS7N+bgY/seRlMpqxymNzjYba8EbDcnkcAvwZhVyxwsed6IOYlwFf3mj3vh8keVhuT/nMbcA4mteh/gNK23RigV3bjNm7cWEOZBQsWnG4TsiTU7VMNfRudfXkj1O1TLXgb4+PjtVmzZkHXTZgwQR966KH0a599ffr00UceeSTTcQ4dOqTh4eGqqur7f3xWh1veDw16AtPs+TROLvEvVdV4Tzv/ax/fAh1tfuxrgIWqehQoDbwvImswqTovyGe7fak91wBrVXWnqh4DfgfqAp0wLzDLRCTOXp8fsCeHw+EowuzcuTP9fObMmRk8+zPj+eef5+DBg4wYMSJD+d69e0lLSwNg6NChxMTEBG2HC85zmhGRqsCVQAsRUaAkoJj84of9mvtfA6CqyXa5vQvQg5MvEI8Bu4ELMf4byfls/jH7N81z7rsuBQhmBSJ419JixtatW+nVqxe7d+9GRLjvvvvo168fcXFxPPDAAyQnJ1OqVCnGjBlDmzZt2L9/PzExMfz222+EhYXx4Ycfpv/PIyYmhq+++ooKFSoQHx/o3dDhcBQGPXv2JDY2lr1791KnTh1eeuklYmNjiYuLQ0SIiIhg3Lhx6e0jIiL466+/OH78OLNmzWLu3Ln8+eefDB48mKZNm9KqVSsAHn74Ye69915iY2MZMGAAIkL79u155513grbNTfqnn+6Y/fj7fQUi8iNmrzwnfArcC1wM9LZllYFtqpomIndjXigKk3nAbBEZrqp77AvOWaqabdSo4kKpUqUYNmwYrVq14tChQ7Ru3Zqrr76ap556ihdffJFrrrmGb775hqeeeorY2FiGDBlCVFQUM2fOZMOGDTz00EPMmzcPMDKhhx9+mFtuueU0P5XDUbyZOnXqKWX33HNPpu0TEhJOKduzZ49vW/UUunfvnr6/n1Pc8n425EZKZ+8LVk7XE5jpuS8Cs/9+ihc/UN0rpwPK++R0QH3gauAHVT1u248B7haRVRhnwIArBdnwpYhU81w3EJGvgrlRVdcBzwNzRWQ18D1m399hqV27dvpb/FlnnUVkZCTbt29HRPjrr78AOHjwIOeeey6QUarTtGlTEhIS2L17N5AzmZDD4SieuC/97PFK6V70rxSRUqqa6n+tQcrpVLVjgOJ96udZr6qxQFvPOLF4POstY/zu2Qy09BQ9bcsTgGw3lFQ1WkQSPNcDraqgv6pG+9kW673Pc/4pZhUiHRERjHIkLTsbihMJCQmsXLmStm3bMmLECLp06UL//v1JS0tj8WLzDumT6lxxxRUsXbqULVu2sG3bNmrWrHmarXc4HGcCbtLPAiuluxwrpcNO+v5yOis388rrci2ny4OtAykgOV0WY3YARtpLBdqr6iEReRK4DePJP1NVX7QrGN/ZcVoD1wJbPH15JXu8PTn4YBOFTc1y5It9PmkPwNGjR+nXrx/33nsvv/zyC6NGjeKee+6hQ4cOLFiwgG7dujFs2DAuu+wyRo8eTcOGDTn//PNp2LAhK1eu5NChQ4CRCaWlpREbG5tn+wqSpKSkkLbR2Zc3Qt0+CH0bC8y+7Nz7i/NBACmdBpDP+V/rmSOn+y8QZ4+9mEl4pt84CUA1z3U0RuMP5kXoMnteEfMS2Rl4D+PEVwLzgtPePlca8Lfsnq24SfaOHz+unTt31mHDhqWXVapUSdPS0lRVNS0tTc8666xT7ktLS9P69evrwYMH08vi4+M1IiIiX+0rCEJd0uXsyxuhbp9q6NuYG/twkr08k5mUDoqGnK6ep30iMEFVb/brJ5Ania/sZ+AtEfkXcLaabY7O9lgJ/IJZ+Whk229R1f/m7bGKFqrKPffcQ2RkJI8//nh6+bnnnsuPP/4IwPz582nUyPyEBw4c4Phx47Ixfvx42rdvT6VKlQrfcIfDcUbilvczITMpnV26huIjp0sEqmBWAgCq+s5V9VUR+RqzVP+ziHSx/Q5V1XHeTuzyfm4cCUOWzOR2+/bto0ePHiQkJBAREcH06dOpUqUKYHJkP/roo6SkpFCtWjUGDx7MpEmTqFChAmPHjgXg7rvv5v3336dfv36kpqYSFhbGe++9B8D69eu5++67ERGaNWvGBx98kG6PTyb0559/psuEsvIYdjgcxQ836WdOfknp4MyR021V1Qp+7WKBfwAviEhJ4C5gFoCI3AHcoarXi8glmK/674CXRWSyqiaJSDiQUkjPVKhkJrebOHEinTp14plnnuHVV1/l1Vdf5bXXXuPAgQP07duXOXPmUK9ePfbs2UONGjXYsWMHO3fuzNBPv379WLFixSljtmvXjk2bNgW0xycTio2NJTo6uiAf3eFwnKG45f3MySCls3xOYClddswFOpD/crpco4HldBKg6ctAQ2vnSuB/wCe2rjvQ3t6fAnyrqnOBKcD/2a2Lz4CzCvRhThOZye1mz57N3XffDZiv9lmzZgEwZcoUunXrRr16ZlelRo0aWfbjcDgc+Y370s8EDSClU9VRnstYT3ms99qWVfScp2CWxb31eZbT+V0PDFTnb5tmIaez2n+fOmEgZhm/ObACiFJVFZG/A+tE5AhGyrjQfulXAN4VkeYYf4XnVXW2iIzEZOobJCJPiMhCIFqLmFzPK7fbvXs3tWubcAS1atVK19Fv2rSJlJQUoqOjOXToEP369aNXr16Z9uNwOBz5jZv0HZlxEdAM4/n/M3CZiCwH3sf4OvyPjPr754D5qhojImcDS0XkB4xUcJmILAJGAdcGmvDPNMmeV0rjL7dLTU3NUH/ixAliY2PZsmULGzduZNiwYRw/fpyHHnoIEaFu3boB+8ktoS5FgtC30dmXN0LdPgh9G51kr5gdGMe/OL9jZgGP6ZMZRgPfe8rfxezlR2G+7H3lN3JSvrcc+NVj6x9ApK27FEgFHgnGjjNJshdIbte4cWPdsWOHqqru2LFDfc8zdOhQfeGFF9LbxcTE6PTp0zPtJz/sC1VC3UZnX94IdftUQ99GJ9krZqjqd6oa5Xf4y+kKEq/H/wmyXxUS4BaPrfVUdb2ta4FRAZxbAHaeNlQDy+1uvPFGPvroIwA++ugjunbtCkDXrl356aefSE1N5ciRIyxZsoTIyMhM+3E4HI78xk36jpywAYgQkQb22uvU+B3wiA2xi4hcZP/WB57AbBdcIyJn5GZ1TEwMNWrUyJAOs1OnTkyaNIlx48ZRpkwZypUrxzfffEODBg0YPHgwZcuWZfDgwbzxxhvExcURGRlJ7dq1qVixIlWrVqVmzZo0b96cn3/+mUmTJjF//nyioqKIiorim2++OY1P63A4iipuT98RNGpiDtwHfG0d+RZx0jP/ZWAEsFpESgDxInID8AEmUuAOEbkHmCgil6hqfsclKFB8Gey8jnfz589PP3/iiSeoXLky1157LQAPPPAAAGvWrOGmm24iKiqKxMREFi5cyNatW6levTp333038+bNo1OnTr7tEofD4ShQiuWkLyK1MBPUJcABTJCcR1U1sAA6+H7PxujWx9jrc4FRqpq7HIin9r8EE0q3KiaUr0/X1Rd4PK/jqFUcqPX498Tzf9jTbCwm/O9ev3uPAvdzKld52qzALPWfcbRv3z5g+kswy/zTp0/P8BLgY+rUqdx+++0A/P777zRq1Ijq1asDcNVVV/H555/TqVOnArPb4XA4vBS7Sd8uP8/ERKO73ZZdCNQENnnaZcieFyRnYybgMQCqugOjZc8XVLWtta03ZuL1Tsa5TtZT2OTytw1ZFi1aRM2aNdND5Xr59NNPmT3bKBEaNmzIxo0bSUhIoE6dOsyaNSs9pK7D4XAUBsVu0sdkzEtR1bG+AlVdBQGz50UCr2K82csC76jqOJt9bzYmPG26Jt22bSAicZhgN+9gvNubi0gYxgv+Yown++OqusBO4DdiEvA0wHjoP5WTB7Ihbn3j9AZuwmTXawS8iUn08w+Mc961qrrP7su/A1QHjgD/VNUNORizAjAdqIOJJviyqn4qIq2BtzAJePYCvVV1pw1FHIfJWjgVGObXX8hK9nzZ8Hbt2sXhw4dPkdIMHz6cNm3anCKvWbduHarK3r170+v69u3LNddcQ4kSJWjWrBn79+/Pd1lOqEuRIPRtdPbljVC3D0LfRifZyz9Z2r+A4ZnURZMxe959mAkdzKS/HDgP87JUyZZXw2jWBb8Med5rjDPbh/a8KUbSFoYJy/s7JixvGCbTXd1snqE3MDqTcXpbe87CTOgHgQds3XDMNgaYMLyN7HlbjMbef5yB2Mx9nrIE+8y3AO97yitjXoAWA9VtWQ/PM8cCY4L5NwpVyV58fLw2a9Ysg5QmJSVFa9SooVu3bj2l/aOPPqqDBw/OtL9x48bpk08+me92hroUSTX0bXT25Y1Qt0819G0sKMlecfzSzw5vtrzOQEsR8S3RV8Z8PW8DhohIe0wCm3DM9kBWXA68DaCqG0RkC9DY1s1T1YMAIrIOqI9JwZtbFqjqIeCQiBzEpMAFk3GvpV2puBSYYZ3twbzU+JOZd5navoaJyGuYVYZFNhpfc+B7229JYKfnvk9P6ekM54cffqBp06bUqVMnQ3laWhrTp09n0aJFGcp98fb379/PmDFjmD59emGa63A4ijnFcdJfS9b77N4Y+IIJKPOdt4FdQq8OtFbVFBFJwHyl55acauJz0p83y54vw14J4ICqRmXTTyJQ26/sLHtvooi0wmTYe0VE5mF8JdaqartM+jsjs+zFxMQwZcoUTpw4AcCtt95KREQEKSkp/PHHH4gIUVFRxMXFpd8zY8YMdu3axRdffEH//v0BmDNnDrfddhvJyclUqVKFkSNH0rhx44BjOhwOR0FQHHX684Gydg8ZABFpKSKBsud9BzwoIqVtu8Z2L7sysMdO+B0xX+YAh8g8ucwi4E5fP5hc9hvz44Fyiqr+hZHU3WrtEevM6M9C4EYROcu26wasUtUTVplwRFU/Ad4AWmGep7qItLPtS4tIs0J4pAKld+/eLF68mCZNmpCSksKMGTNYtmwZcXFx7Nu3j969e9OtW7cM98yYMSND2YkTJ3jooYeIi4sjKSmJWrVq0bJlS/+hHA6Ho0Apjl/6XYE2wD4ReRqTxz4BeBSzTO9lPGa//Bfr+LYE4yQ3GfiPzSK3HBO0Bvv1+7OI/Ap8i3GUK22vL8YkpFmDceTrjfEhGASUs6sH24Fa9p6mmEn143x+/ptF5GrMV/8EEXkX43A3BVjlbaiqq0VkNPCTiCiwB5MiGIz07g37u6wDHlTV43YrZJSIVMb89zUCs7pyxpJTud6sWbM477zzqFDhZJbipUuX0rBhQ84//3wAbr/9dmbPns0FF1xQoLY7HA6Hl+I46ffEZIdbqqrX+NVtFpGffBdqct2/oKrPBugn4BK2qt7hvRaRLpg972SgT4Bb6tl2sRinueXZPYCqTgQmeq4TsJn5AtRFeO+zLxf9VXW5iJQBhmLkf4MyGWscMC5A+XeYlRD/8jigfYDy6Oye60zEX66XlJTEa6+9xvfff8+bb76Z3m779u3piXUA6tSpw5IlSwrdXofDUbwpVpO+dWC7HCPb+w/woi2PJqNU7z7vNdBYRJJUtaKITAMmqerX9t6JwFeYL/5JGKkcwMOqmmvtvCcwzpv2hWAlcIXtvxcme10L4FNVfd7ecxdGnVAGsyrRV1VPZDaG/TJ/CvifiFyoqqsC9QH8E2igqk/acXpj4wT4fhdb/jQmMU8a8K2qPhOsNDBUJXv+cr3Y2NgMUhp/ud67775L586dWb58OQkJCZQrV47Y2FjWrl3Lzp0709utX7+e7du3F4gkJ9SlSBD6Njr78kao2wehb6OT7OWPXO9O4AN7vhjjiAenSvUyXNsyXwa6mzGBfcBMjFsx0fHKA2G2vBFWOoGfjC8L22IxE6nvehtmuT8OSMJEDWwB9MOku62N8bjfBpwDRGJeZErb+8cAvbIbx5bNwsjrAvaBmbD/52n/LXC53+9yjf1Ny9vrqhqkNND/CEXJnk+up3pSShNIrnf55Zdr/fr1tX79+lq5cmWtUqWKvv3227p48WLt3LlzershQ4bokCFDCsTWUJciqYa+jc6+vBHq9qmGvo1Ospc/9ARG2vNp9nqFvfZK9QJd+/gWGCkiZYG/Y1LNHrV72KNFJArjgZ9Xt+zxZPzSf05V14hIdYyH/E4AEfkdqItZwWiNyV0P5kVkT5Bj+XR7nQL1oap/isjvIvI3YDNm9eNnvz6uAiao6hEANQGAgpUGnpEEkut5JXoDBw6kYsWKPPzww6SmprJ582bi4+MJDw9n2rRpTJky5XSY7XA4ijHFxntfRKoCVwLjrcTuSeA2OTkb+cvJAsrL1OzNx2Ly3ffgpPb8MczX+IUYp70y+Wg+ZJTd+UvySmEm7o/0ZGrbJqo6MLtORaQkZgVhfTZ9TANuwwTlmWnfKrMjXRroOSKDuC9PbNy4MT1bXVRUFJUqVWLEiBEAvP322zRt2pRmzZrx1FMnAx+uXr2adu3a0axZM1q0aEFy8sl8QD179qRdu3Zs3LiROnXq8PXXXwMwbdo0evbsSTCUKlWK0aNH06VLFyIjI7ntttto1uyMFzY4HI4zjOL0pd8dsxefnhRGRH7E7JPnlE8xXuwXY7zwwcj4tqlx/rsbE5imMJkHzBaR4aq6x77knKWqWzK7wUoRBwNb1Xjqp2bRx0zgOUyK3KcDdPc98IKITFbVIyJS1X7tx4vIrao6w75gtVQb9rigaNKkSbpm/sSJE4SHh3PzzTezYMECZs+ezapVqyhbtix79piFkNTUVO666y4mTZrEhRdeSGJiIqVLl07vb+rUqRn69+2zTZw4MUs7Bg4cmOH62muvTc/C53A4HKeDYvOlj1nKn+lX9jmenPAicpOVptXLpq+5QAfgBzXOcIsx+993i8gqzPJ3toFoRCTCyvkC0QB4zMbxvwL4zJ4PAOqJSC9vY1VdBzwPzBWR1ZhJ2D+wjo/Jts2vGMfArrZ8DPBhoD5UdT9mNaC+qi719FXB1s8BvgSWWzv72/o7gXvs77LWM1ahMG/ePBo0aED9+vV59913eeaZZyhb1uww1KhRA4C5c+fSsmVLLrzQhCo455xzKFmysN/ZHA6Ho+ApNl/6qtoxQNkoz2WsiHyKkfM1UNXr/dpWFJsdTlVTgKoiUsrWXWqbeaOtPG3rErByumzsi/a7/ofvPCs5n/c+Vf2UbELd+o8TgO9VdWgm914foPiwp/5VTNIh7z3xGN+H04J3CX7Tpk0sWrSI5557jrCwMN58800uueQSNm3ahIjQpUsX/vzzT26//fYMS/8Oh8NRVCg2k352ODnfKWMkYZwerweOAl1VdbeInIcJ5OPLNOi950nMvn9ZzL7/iyJyM/AwxtGvFvAj0F5Vd/ndmy+SPZ/EDiAlJYXPP/+c66+/ntjYWA4ePMiaNWt49dVX2bBhAzfeeCNTpkxh48aN/PDDD4wdO5ayZcvyxBNPULJkSVq3bh1wjGIr9clHQt1GZ1/eCHX7IPRtdJK9M1/Odz1GfrcOo1ePwwS3yY2cbyA2+52te82e55ucD5NU5wZ7/jonsw1+6bsXeMjz7J2B9zDOgCUwLzvtbd0nmIn/K6Bnds+bX5K9WbNm6dVXX51+3aVLF50/f3769fnnn6979uzRqVOnaq9evdLLBw0apK+//nqm/RZFqU9hE+o2OvvyRqjbpxr6NhaUZK847elnR0+MhzqclPP5yImcr6OV812DlfNhUs72wDj3HQdQk+zm/gB95IYv7d81WDmfqh7DpOytS0YpXpy9Pj+bPo9jJmkwssYIe34Z4PNsm+Rp39keK4FfMKsgjWzdI5jVh2OqmtErrgCZOnVqBu/6m266iQULFgBmqf/48eNUq1aNLl26sGbNGo4cOUJqaio//vijC4/rcDiKJG55nwxyvhbWka8koHa5GnIg57PL7T45n+8lwivnK4GJ95+fBCvnG5CDPlPsmyOcmvkvkFxPgKFqwvb6U8faUlNESqhqWg7syDEbN26ke/furF+/nrVr19KvXz8GDRpE3759iTBq8QoAACAASURBVImJoVatWuzevZsvvvgCEWH8+PEkJiZStWpVAI4dO+ZS3jocjiKJ+9I3+OR89VU1QlXrAvHkXs7Xx947x5ZVBnbaye4fQEkROQF8AzQUkVUi8oSI5PbfI1JERmVRPw/oLiI1wLzkiEj9AO2aYl5+suJn4HZ7fidQXkSqYbYqYqxvBCISLiI1rLPjh5iVk/XA48E+VG5p0qQJa9asITU1lbi4OMqXL8/NN99MmTJlGDp0KBdeeCH16tXjiivMP++TTz7Jli1bSE5OZsaMGXTs2DH9BcDhcDiKEm7SN2Qr58sBGeR8tiyQnO8oJhf9/4CrMdsBL+ZiPID1qvqvzCo1Z3K+7OgHPGSzBaZnJVTVuRgHv/+zdZ9h0gw/CyxS1Z8wE/69IlLgAXp8eCV7AI899hivv/46ngiBGfDfEnA4HI6ihFveJzg5n6c81nttyyp6zlOAqn71m/GT81mP/wSsnM96ry+znvklMNK3aIxD3jtqsuL51AED7T0TgdFARRH5SlWvF5HlIjIBEzhIgfoY9cB+zItGWcxef6D4AHNsHcBeEXkJuAHjk3CrLf8LOARUwnjjb/Xcn2jHKAOsxqQs/hb4QETCMEv8JyjEwEVeyd7s2bMJDw9P1+P7c+TIEebMmcPo0aMLyzyHw+EoVOTktq2jMPFmp/OUHQCaYALY1FDVV6xT4M+YSTcKuElV77ZpcX/DxPhvi/Hmv15EXgPKquqjts8qmEn2C+AaVT1ss+GVVb90uj6Joap+ZkMVD1PVt0WkL9BKVe+12wh7VXWQiFyHcfarbo/XgW6qmiIiY4D/qurHIvIKEIZRMmzTAHEA/CR7rV8Y8X6ufld/yV737t2ZMGEC5cuX57HHHuONN96gYsWK3H777YwbN47KlU+2nz9/Pj/88ANDhgzJcoykpCQqVqyYZZvTSajbB6Fvo7Mvb4S6fRD6NubGvo4dO65Q1YuzbJSde787CkwimOQ574KR8J3AfIEfwHxNx9kjHuMZHwb8gfla7wpMtvdHYyZrMJ72jfzGuh7Y6+lvHVae6NduItDdnicA4fa8LWa7Anv/+Z579gHVMJK8HZ4xNgIDbZsywCpMfICS2f02BSHZW716tVavXj09A17JkiW1bt26unPnzvT2N910k06ePDnbfoui1KewCXUbnX15I9TtUw19G12WvSKMqn4nIpuBZZigOp8B76nqd/5tM1EHZIdgIu3ldLPapwTw997PbIzMFALnYIL5lMa8uGQbojg/8O7Pt2jRIj3WPkBERATLly+nWrVqABw8eJAff/yRTz75pDBMczgcjtOCc+QLAWy63LHAaPu29h3woE2Ig4g0FhFfNL9A6gAv32OC5vj6rgL8F7hMRBrasgoiktvUvwuBO2w/1wBVbHlWCoFxwL+BycBruRw3aDZu3EiLFi347LPPeOuttzJk2QMYNmwYW7ZsITExEYDJkyfTvHlz0tLSuPrqq1m1qkDzATkcDsdpw33pnz7K2UA5pYFUTKCbt2zdeEwwnF9sZrqmmK9lMOqAScBsPakO8PIKJjnPX5htgZdU9QsR6Q1MtT4CYLz5N2Vj41MisoeMjos3Ab+JSE9M5MI/wCgERMSnECgBpGC8/DtgNP9TbBrfxSJyparOz2bsXOOT7EHGLHsAW7duZe7cudSrV49zzjE/6Xnnncfq1aupUqUK3377Lffddx9LliwpKPMcDofjtOEm/dOEqmbqwa5Gz/+sPXxOfwdtXSB1QCx2YlbVJBF5FevY52kzH7gkG5t6e84jrJIANYl+om1VGnCrqu4NcH+ghD//BT629SeAtr5ERYVBZpK9rl1PJvu79NJL08//9re/sW3btsIyz+FwOAoVt7x/BiEi0SISKyKficgGEZlsVwIQkb/bsl+Abp57KojIhyKyVERWikhXWz5SRF6w511EZGFOggPZfr+2gYV+FZEetry1iPwoIitE5DsRqW3LY0VkhIgsx2j9C4WcSPYAPvjgA6655prCMs/hcDgKFSfZOwOQk1n8ojGZ7ZphPOV/Bp7E6PA3Y6Lp/Q/ztV1ejYRvCLBOVT8RkbOBpcBFGA3/MozX/VjgWlX9zW/cgdhsfp6yBEwMgA7A31X1n7a8MiaR0I+YjHx/2heBLqoaYx0Q16lq30ye8bRL9lauXMmIESMYNWpUhnJ/iqLUp7AJdRudfXkj1O2D0LfRSfaK8cHJTHbRGC98X/m7wF0Y/f5CT/mNnJTwLcfIAH1Suj+ASFt3Kcaf4JFMxn0Rm83PU5aA8S9obM9fA66wdc0xwXt8Y60B5tq6WKBDMM97OiR7q1at0vPPP183btyYbb9FUepT2IS6jc6+vBHq9qmGvo1Osufw4U2oE6yU7hZV3RigrgUmit65mdybyKnhes8CDqhqooi0woQSfkVE5mFCGa9V1XaZ9FcoUj0fwUr2/vjjD7p168akSZNo3Di3ogaHw+EIfdyeftFgAxAhIg3stVeP/x3wiGfv/yL7tz7wBGap/xoRaRug34XAjSJylr2nG7BKVU+IyLnAEVX9BHgDaIUJyFNdRNrZ9qVFpFl+PODWrVvp2LEjF1xwAc2aNWPkyJEA/Pvf/6Zly5ZERUXRuXNnduzYAcDhw4f59ttveeutt2jWrBkdOnTItO9BgwaRmJhI3759iYqK4uKLs14dczgcjjMV96VfBFCT0vc+4GsROQIswnyRA7wMjABWW0e9eBG5AfgAs3S/Q0TuASaKyCWqmuzpd7WIjAZ+EpNyeA9wr61uAbwhImkYed6DqnpcRLoDo+wefyk79tq8PmOpUqUYNmwYrVq14tChQ7Ru3Zqrr76aJ598kpdffhmAUaNGMWjQIMaOHUtKSgrh4eF8/fXX1KtXL8NXPkBCQkL6+fjx4xk/fnxeTXQ4HI6Qx036BYyI1MJMfJdgwuvuBh5V1ew08umojdGvVppnHfLuUNWH7RjnAveqatMA9x4F7g/Q7VWeNiswk7jP5iWYUL9VMfHyt9uqgZj4+t3VRAs8JWKgqsYB7QOUR2f/pJlTu3Ztatc2Ow1nnXUWkZGRbN++nQsuuCC9zeHDh9Oz502ZMoVu3bpRr149AGrUqJGX4R0Oh6NI4Cb9AsQuqc/EhKe93ZZdCNTEBsYRkVKqmprDrs8G+mJS9qKqO4Du+WW3qra1tvUGLva9XFgW59c4uSUhIYGVK1fStq3ZkXjuuef4+OOPqVy5MgsWLABg06ZNpKSkEB0dzaFDh+jXrx+9evU6nWY7HA7HacdJ9goQEbkSk3SmvV95NGbZfT8m2l4kp6bSHSciFTESvSqYyH3Pq+psMSl2u2L20L8H3sF46zcXk8L2XYysLhV4XFUX2An8RqA80ACYqapPZWN/bzyTvohEeMbpjYnOVwFoBLyJSazzD4yz4bWqus/6GbyDycJ3BPinqm4IMFaWkj2fFO/o0aP069ePu+66i/btMy4oTJ48mePHj9OnTx9GjhzJxo0bGTZsGMePH+ehhx5i6NCh1K1bN6tHDoqiKPUpbELdRmdf3gh1+yD0bXSSvTPwAP4FDA9QHo3xZD/PXt+HmdDBTPrLgfMwKzGVbHk1jAZfMCF6f/X0l36Ncc770J43xUj0woDewO9AZXu9Baibjf29MfkAAo3T29pzFmZCPwg8YOuGY7YwwMTkb2TP2wLzs/vdMpPsHT9+XDt37qzDhg0LWL9lyxZt1qyZqqoOHTpUX3jhhfS6mJgYnT59esD7ckpRlPoUNqFuo7Mvb4S6faqhb2NBSfac9/7pY6mqxtvzzkAvG4t/CUYH3wgzwQ8RkdXAD0A4ZmsgKy4HPgFQ80W9BaOpB5inqgfVOOutA+oH7iJoFqjqIVX9EzPp/8eWr8GoCSpiYgHMsM82jlMlgEGhqtxzzz1ERkby+OOPp5dv3rw5/Xz27Nk0bWrcGrp27cpPP/1EamoqR44cYcmSJURGRuZmaIfD4SgyuD39gmUtme+1ezXrggmQk8Exzi6hVwdaq2qKjYYXlgd7cqrxz0l/aZ7rNNt3CYymPyqP4/DFF18wadIkypYty9ixYznnnHN4//33eemll1i1ahXHjh3j8ssvZ9o0k2340KFDrF+/Pn157I477qB58+Z5NcPhcDjOaNyXfsEyHyhr96sBEJGWmLS4XjJLpVsZ2GMn/I6c/DI/xElJnj+LgDt9/QD1MHv/hY6q/oWRCN5q7RHryJhjLr30UlasWEFycjJ//vknFSpUICIigokTJ7Jq1So6dOjA8OHDCQ8PB6B58+b88ccfJCcnEx8fz1dffUVqak79JR0Oh6NoUaCTvojcJCIqIqdIybK5L1ce4iISISK/ZlLXRUTi7JEkIhvt+cci8oCIFIRrdylMfPthInLMaujHA7v82o3HLLf/Yu0fZ++dDFwsImuAXpggPKhqIrBXRNaKyBu2j/pWyjcGKGHv+RTorarHyAEi8o5djh8MPOD73YBrPOMEy53APSKyCrPy0TWb9gGpXbs2rVq1AjJK9iIjI2nSpMkp7cuXL0+pUmYhIzk5OV3K53A4HMWZgl7e7wn8ZP++6F/pL1fzXavqpf5t84p6dOU2+Ut/NSljC5KXMV/k1VT1mIjUxMSfnw543dNFVdNT6fqRWUjbczCe9b4Ut94v/z7+jVV1IjDRc329fxtP3UOQwVvfuzz/rv3r319EoLGs38LfMxsrN/hL9jJjyZIlxMTEsGXLFiZNmpT+EuBwOBzFlQL70rdOXJcD9wC3e8qjRWSRiHwJrPO/tm2S7N9pInKd596JItLdftEvEpFf7JGnlwQRGSgi/e15rIgMF5HlIrJeRC4RkS9EZLOIvOK55y4x6WrjRGSciJT067M88E/MXv0xAFXdbSd87GrDMPsF3C6z/kTkXWvLWhF5yZb9CxMvf4GILLBlCSJSzZ4/Libd7a8i8qgti7DP877ta66IlMvFb5UgItVsfxvsv8kmMWl+rxKRn+1v1ca2D5jaNyuOppwg4pmv0w8vSUlJ3HLLLYwYMYJKlSpl2U/btm1Zu3Yty5YtY+jQoSQnJ2fZ3uFwOIo6Bfnp0xWYo6qbRCRRRFqrifwGJk57c1WNF6NZT7/26+NT4DZMeNkyQCfgQYzj29Vqws82AqZidOn5xXFVvVhE+mF08q2BfcBvIjIcqAH0AC6z++1jMMvYH3v6aAj8Yfe1A1EBWKKqT4hIJPB0Jv09p0bvXhKYJyItVXWUiDwOdPR86QMmnz3mS78t5ndaIiI/YmICNAJ6quo/RWQ6cIuIPIKRCXr5h6quCeJ3agjcCsRgtjHuwLzo3YhZtbgJeA4j04ux2wJLReQHVc2QfEc8Ov3q1asz/e8V0utiY2MBSE1NZcCAAbRt25aqVaumlwMcOHCAFStWkJSUFNDQ1NRUPvroo4BbATklKSkpw9ihRqjbB6Fvo7Mvb4S6fRD6NhaUfQU56fcERtrzafbaN+kv9Zvg/a99fAuMFJGymCXihap6VExc99EiEoXxQs/v1Ghf2r9rMFnjdgKIyO9AXczE1hpYZveKy2Hi0ueEE8Dn9rxTFv3dZifEUhi52wXA6iz6vRwTeOewtfkLjOPgl0C8mjC5YP4tItRG38sl8b6XAxFZi5EEqvUniLBtOmOS9vS312EY58L13o5U9T3gPYAmTZpodHQ0fvXcfffdXHbZZYwYMeIUQ84++2xat26dniwnPj6eunXrUqpUKbZs2cKuXbu45ZZbqFatWh4e1xAbG4u/faFEqNsHoW+jsy9vhLp9EPo2FpR9BTLpi0hV4EqghZhELSUBFZEnbRP/FKsBU67aL/lYoAvmy3qarXoME8P+QswWRX6v23qlZ/6ytFKYL+iPVHVAFn38D6gnIpUy+dpPVtUT9jxgfyJyHtAfuERV94vIRPJXspfj5f0s+gsk2YOsU/sGRUxMDDNnzuTAgQO0aNGCqKgodu3aRUpKCpUqVWL79u2oKtdddx1NmjRh2bJlVK9enT179lCxYkXq1avHmDFj8mXCdzgcjjOZgtrT7w5MUtX6qhqhqnWBeE6VqgXDp5jl6iuAObasMrBTVdMwYV9LZnJvQTEP6C4iNcC85IhJVZuOqh7BZLIbabcmEJHqYuVrQfZXCfNCdNA6AV7juScz2d4i4CYRKS9G9nezLcstpeVURcRTGH+FYAiY2jcn9O7dm3nz5tGsWTNWr15NXFwcDzzwAAMGDCA+Pp7jx4+TkpLC7t27+fjjj2nQoEG6XG/v3r388ssv3HTTTTkd1uFwOIocBTXp98QkmvHyORnzvAfLXKAD8IOqHrdlY4C7rRNcUzJZKSgoVHUd8DwwV0y0vO8JHGnueeBPjMPir8BXwClf/Zn1p6qrgJUYqd4U4GfPbe8Bc3yOfJ6+fsF4zi/FRPcbr6orc/+0eeZlTN6A1XYL4GURydEKU/v27alatWqBGOdwOBzFiQJZ3lfVjgHKRnkuYz3lsd5rW1bRc56CSfHqrd8MtPQUPW3LE4Bsw66pX5pXVR0YqM7fNr+6TzGrEFmNcxzzVXxKYhvvM2bVn6r2zqTvt4G3PdcRnvO3gLf82ifg+W1U9c2sbPfdIyJdMC8r6eOIyEAgCXhPRNZhEvusU9XP7OrCC8AREVmJSTh0v5jogt2AipiVjQ7ZjZ8do0eP5uOPP+biiy9m2LBhVKlSBTD7+RdddBGVKlXilVde4YorcrPA5HA4HEUPF5HPkReeAS5S1ZbAA7bM563fBugIvGFfBMCoNLqrapYTvleylxkPPvggv/32G3FxcdSuXZsnnngCMEF8/vjjD1auXMlbb73FHXfcwV9/ZSagcDgcjuJFkYxWYr9OX/MrjlfVm0+HPaGMiLwDXOZXPFJVJ9jzzHIvK0ZFMFlEZgGzbHlm3voA36vqvkzsCCjZ80lWdu3axeHDhwNKWFq0aMGUKVMC1p1zzjlMnTo1X6R6Poqr1Cc/CXUbnX15I9Ttg9C3scDsyy4NnzuK94FZjt/uVzYKuBvjQNkRs5WwHvMSuQJoEqCf3njS9GZ1BEqtGx8fn542V1V1x44d6edvvfWW9ujRQ1VV9+zZo6mpqaqq+ttvv+m5556riYmJp/SXF4piSs7CJtRtdPbljVC3TzX0bXSpdR2nBVVNAnaKyJWQLsf8Oya8cl1VXYDxqaiMeUHIs7e+l+HDh1O5cmUaNWrEunXrCA8P54orruCCCy4gLCyMypUr8/333zN8+HAAFi5cSMuWLYmKiqJ79+6MHTvWOQE6HA6HpUgu7zvynV7AOyLicw58CfgDEwa4MkaLP0pVD4jIy8AIjLd+CYxUM9M4/1mxfft2Ro0axa5duyhXrhy33XYb1157Ld26dUsPwfv4449To0YNatc24olbbrmFW265JS/P6nA4HEWWYjHpi0gtzER0CXAAE9jnUVXdlMd+zwbuUNUx9vpczOTXPY8m+/pfggmRWxUTSGe7reoLPJ5f43jG641J4vOwt1yNpPAURQZwuQ0Y9JWqfmbbHgXu92+ofgl/giU1NZWjR49SunRpjhw5wrnnnps+4asqR48edRn0HA6HI0iK/PK+XWaeCcSqagNVbQ0MAGr6tcvNC9DZmAkYAFXdkZ8Tsaq2VZPh7gXgU1WNssfi/J7wQ5Hw8HD69+9PvXr1qF27NpUrV6Zz584A9OnTh1q1arFhwwYeeeSR02ypw+FwnBkU+Ukf84WaoqpjfQWqukpVF8mpGf9KisgbIrJMRFaLyP1gMgaKyDwxGf3WeDLFvQo0EJMZ7w2bee5Xe0+YiEyw7VeKSEdb3ltM1r45Nhvd6zl9IL9xeovILBH5XkwGvIfFZNlbKSL/tXvwiEgDO+YK+8xNsxljooiMEpHFIvK7iHS35SIio0Vko4j8gEk+5LuntYj8aMf4TkRqi0hl27aJbTNVRLKM5ueT7O3fv5/Zs2cTHx/Pjh07OHz4MJ988gkAEyZMYMeOHURGRvLpp1mGS3A4HA6HpTgs7zfnZKKfQHgz/t0HHFTVS8Qk+flZROYCW4GbVfUvMelr/2tfFJ6x90ZBev55Hw8Bqqot7AQ7V0R8iYGigIswseo3isjbqro1j894EUYe9z/gaVW9SExGwF6YrY33gAdUdbOItMVENbwym35rYxL4NMUk7PkME9a3CSbxT01MOuQPRaQ0JlhQV1X9U0R6AIPVZNd7GJgoIiOBKqr6vv9AgSR7I0eOJCwsjLVr1wIQGRnJjBkzqFOnTvp9TZo04b333uO8887L8Y+WW4qt1CcfCXUbnX15I9Ttg9C38UzMsnem4M3w1xlo6fuqxXikNwK2AUNEpD0mmUw4ftsDAbgcGzFPVTeIyBZOZgOcp6oHAWxEu/qYF4vcskBVDwGHROQg8B9bvsY+T0XgUmCGZ//bP51uIGapyW+wTkzsf4D2wFQ1yYJ2iMh8W94E8/LxvR2jJLATQFW/F5Nz4B1MkqRT0ABZ9sqVK8eMGTNo06YN5cqVY8KECVx11VXUqVOHhg0boqp89dVXXHbZZYWaLau4ZufKT0LdRmdf3gh1+yD0bTyjsuyFGGsxCYAywxu3X4BHVPU7bwPr4FYdaK0m330C+ZvtLq//DtlluysBHPCtSOSy3+y85QSThrjdKRXGiz8SOAJUwbxEZUvbtm2pU6cOVapUQUSoVasWo0aNomHDhvz111+ICOHh4SxdujTIx3E4HI7iTXHY058PlLXLxwCISEsRCRSQ/TvgQbtUjYg0tiFkKwN77ITfEfNlDplnugOT2e5OXz+YqHS5Ti+bF9Sk9o23X9u+ffkLvb4BHtrIyWh6gVgI9LD+D7U56dW/EaguIu3sGKVFpJmtewwTvOcOYILv982O7du3s2nTJg4cOEBycjJt2rRh5syZfPTRRyQnJ3P06FHatGnD9OnTg/odHA6Ho7hT5Cd9G6XoZuAqEflNTKa3ocCuAM3HY/aof7GT4TjMl/Jk4GIRWYPZI99g+07E7Pv/KiJv+PU1Bihh7/kU6K2qxzh93AncIyYz4VqgazbtM2MmsBnzO30M/B+kJxfqDrxmx4gDLrUOfPcCT6jqIsxLw/PBDuaT7KWmpqZL9q699lpEBBGhTZs2bNsW1MKBw+FwFHuCWlYWkQbANlU9JiLRmAx3H6vqgYI0Lr9Q1R3AbQGqNpMxi14a8Kw9/Dll2drec4dfUXNbngz0CdB+Ih69uqpmG7gmwD0JnnH86yIC3Wf9Fv7u7dfP8RBVnegrU9XeIvIvOZlFz5dpr7w9DgLVMFn0ZsvJLHqpmK2E9KQ61oFvkd3rrwwkZPfMkFGyV65cOTp37pwu2QNISUlh0qRJjBw5MpjuHA6Ho9gT7Jf+58AJEWmIcbaqi8nv7ija5EsWPVUda/0JLsHs52dI++tPMJI9gL59+9K+fXuXOtfhcDiCRMzqdzaNRH5R1VYi8iSQrKpvi8hKVc1TXHXHSeRk9D0v/1DVNQU4Zn3ga1Vt7ikbCBxS1WEiMgdIwmTQm6WqSSKyHOPEmGpvqQp0AdoCHVT1lNUNT99jgD9V9cUAdV7JXuvp06cTGxvL0qVLeeqppwD47rvvWLduHY899hgfffQRmzdvZtCgQZQoUbi7VElJSVSsWLFQx8wJoW4fhL6Nzr68Eer2QejbmBv7OnbsuEJVL86yUXYZeexLwRKgJ/ArcJ4t+zWYe90RugdZZNCz5/mWRc/Wfw2UyM4uX5a9//73v3rBBRfo4cOHNS0tTXv16qWjRo3S999/X9u1a6dHjhzR00FRzM5V2IS6jc6+vBHq9qmGvo2nO8teH8ye9mA1QWzOAyYFea8jRNEsMuhZmV2+ZNETkdZAf+AuNX4TQXH22WeTmJhI1apVKV++PFOmTOHYsWPcf//9LF++nPLly9O4cWMGDRqUwyd3OByO4klQjnyquk5EnsbIzlDjFPZaQRrmKFisw95XGAdHXwa9WpjAQb9ZWd0nkj9Z9B7GbAMssO8Ky1X13uxsbNKkCbt2GZHFiRMnCA8P59Zbb+W6666jRIkS3H///bz55ptcfHHWq1kOh8PhMATrvX8D8CZQBjhPRKKAQap6Y0Ea5yh41JNBz+7nJ9nyFExUQf/2WWbRE5FSqprqV5fpPn+wzJs3jwYNGlC/fv3sGzscDocjIMEu7w8E2mDS0qKqccD5BWSTIwTwyfXEJB6aZssqiMiHIrJUTEKfrra8t4h8aUPyzvPr5xLbR5i9f62INA8wZJZMmzaNnj175suzORwOR3El2PCvKap6UDLmLQ96b9ZxRvIMxmnzmIicbct8cr0YW7ZUTKY9MHK9lqq6z9uJqi6zyYleAcoBn6iqfxTADBxNOZHh+vjx43z55ZcMHTo070/lcDgcxZhgJ/21InIHUFJEGgH/AhYXnFmOQiAzraavfDUwWURmYSR7YBIS3egJ0xuG9fMAvvef8D0MApYByZj/dk7BP8ueN7vUTz/9xHnnncf69etZv359evmBAwdYsWIFSUlJmT5kQVBcs3PlJ6Fuo7Mvb4S6fRD6NhaYfdm59xsVAOWBwZj/cS/DfLWFBXOvO0LzoHDlerWB3zCheytkZ5tPsuejR48e+uGHH6o/HTp00GXLlp1SXtAURalPYRPqNjr78kao26ca+jaeNsmeiJTEBHB5TlUvscfzasLMOs5QtJDkepZxwL8xOQxypPrYvn07M2fOZMiQIURGRvJ///d/9OjRg5IlS7Jw4ULatWtH69atc9Klw+FwFFuynfTV5E1Ps9ItR9GiF/BvEYnDZCN8SVV/w3zlf2KTBa3EyvWAl4HSGLneWnudJSLSC+MTMgV4FbjE96IRDM8++yzvvPMOmzdvZtWqVURGRhIZGclrr71GWloaKSkprFixIqfP7XA4HMWSYPf0k4A1IvI9nvzzqhpwf9ZxZqAeuZ5fea7kepmM8TEmG5/vBbJtsPYdPHiQhQsXMnGi6bpMmTKUKVMm2NsdDofD4Uewkr0vbhrNmgAAIABJREFUMMuzCzH7ur7DUcwJRpInIg+ISJw94kVkQTB9x8fHU716dfr06cNFF13Evffey+HD5p1z9OjRtGzZkpiYGPbv318AT+ZwOBxFj6AmfVX9KNBR0MY5Qh9VXQb4JHmvE0CSp7nIsgeQmprKL7/8woMPPsjKlSupUKECr776Kg8++CC//fYbcXFx1K5dmyeeeKIgHs3hcDiKHMFm2YsngMRLVV2AHgciUoaTkrxL7TJ+oHY5yrK3b98++vbty7Rp0wBYvXo1U6ZM4dVXX02/b9euXQwYMIAJEybk+3NlRlHMzlXYhLqNzr68Eer2QejbeLqz7J3jOcKBRzFheE+79Mwdp/8gCEkeuciyp6p6+eWX64YNG1RV9cUXX9T+/fvrjh070uvfeust7dGjhxYmRVHqU9iEuo3OvrwR6vaphr6NBSXZCzbhTqJf0QgRWQG8ENTrh6Oo45PknYeR5D3srfRk2btCc5BlD2DIkCH87W9/48iRI5QpU4bPPvuMzp07s3nzZkSEihUrMnfu3Px6DofD4SjSBLWnLyKtPMfFIvIAwXv+O0IAEclR2DoRiRaRr4JoF4wkz5tlL05Exgdrx/jx43njjTc4duwYiYmJtG3blp9//pnk5GSOHj3Kv//9b8aNG5eTR3M4HI5iS7AT9zDPeSomnept+W+O40xDg5DkaS6z7AUj2Tt8+DB+OSEcDofDkQnBSvbuUdWO9rhaVe8DjhekYY6CwX7Bx4rIZyKyQUQmeyLs/d2W/QJ089yTWXa9kSLygj3vIiILbTQ/73jZtsmMrCR7zz33HHXr1mXy5MkMGjQoP34ah8PhKPIEO+l/FmSZ48zgIowz5gWYFMmXiUgY8D5wA9AaqOVp78uu1wYTzOcNEakADAB6iEhHTNz+PgH27INpk4HsJHsAgwcPZuvWrdx5552MHj0697+Ew+FwFCOyXN4XkaZAM6CyiHTzVFXCZFhznJksVdVtADYEbwQm6mK8qm625Z9gJXRkkl1PVdeLyD8xQZseUxPCNwOqeiS7Nna8U7Ls7du3j2rVqnH06FFiY2Np0KABU6ZMoVOnTun3nX/++TzzzDP/396Zh0lVnO37fmSPyGqIKCqigIgYNkWNskVBlJ8aJVFjPoEPk2iin8QlaIgLJN9n3CIuSYiigrjhimtcIgyaRBRBFFQQFYwYBWIEGRhghPf3R1VD03TP9GzNcea9r+tcfU5VnTpPF0zXOXXqqZcBA3ZYWLDGqLPRuaqRpGt0fVUj6fog+RprSl957/Q7A0OBFoQnwBRrgR9XuxqnUGxM299M+f8PBJxqZouz5HUDPgf2LOP8csuY2W3AbQCdO3e2/v37A3DjjTfStm1bOnfuTFFREUcffTR77bUXHTt2BOCWW26hV69epMoXgqKiooJer6IkXR8kX6PrqxpJ1wfJ11hT+sr8sTezx4HHJR1hZq9U+9WdJLEIaC9p//g0fkZaXiq63vlmZpJ6mNkbkvYFLiK8LnhG0nQzezW90nzKlEU2y96FF17IzJkzKS0tpVmzZom+W3ccx0kS+b7Tf0PSzyX9MU7oulPSnTWqzCkoFkIl/wR4Ok7kW5mWvUN0vTj57w7gYjP7FzAKmBTnBgCQT5nyyGbZ69KlC7/+9a/ZuHEjF110EXfffXcVv73jOE7dIN9OfyphYtdgYBbQjjDEnxgk7SHpAUkfSJor6RlJnaqh3haSfpZ2vKekapvEKOnV6F3/p6RVaYFpjqzO65hZaj3H9kCfOAt/CdAReC+WedbMDjSznmZ2gZkNjeklZvZTM+tmZl3NbGhcAOoYoLekY8xsbszfkHZNM7NjzOyJeLxDmbJIWfZGjRoFBMteixYtePzxxxk+fDgAw4cPZ/r06dXRRI7jOLWefDv9A8zscmCdhUA7J1CBEKk1TXyifAwoMrP9zawXYdb4tzLKVWZBoRbA1k7fzP5lZsOqojcdM+tjIRjNFcA0M+set39U53UymGZmPcysI2FBnUcldalMRWZ2hZn9tXrlBXJZ9lasWEHbtm0B2GOPPVixYkVNXN5xHKfWkW+nXxo/V8ewqc2BNjUjqVIMIKwKNzGVYGZvmtnL0Zf+sqQngHck1ZN0naQ5MSTsTwEkNZX0oqR5khakvOiETnH/+PR9naT2khbGcxpLuiuWfyPa0pA0QtKjkp6VtETStRX9QhnXGSFpuqQXJC2TdJ6kC+M1Z0tqFcvtH685N37nA8u7jpnNJEyg+0muOiQ1l/RRyl8fffsfS2ogabKkYTH9UEn/kPSmgqd/t1ztXRb5WPbS2skX53Ecx8mTfJ98b5PUkrC++hNAU5K17v7BwNwy8nsCB5vZ0mgNW2Nmh0pqBPxd0vPAx8D3zOxLSbsDs+ONwqXx3O4QOuO0en9OGMXuFjvY59NeKXQnTF7bCCyWdIuZfVzF79iDYJd7HxhjZj0k3QicBUwgdN7nmNkSSX2APwKZS+JmYx6Q6ox3qMPMBipY+/oBMwmOjufMrDTV4SpE2psGnGZmcyQ1A0oI7/F3aG8zW5ouQBWw7DVr1oxHHnmE1q1b8/nnn7PbbrsVdDJfXbX6VCdJ1+j6qkbS9UHyNe4syx4AZpZaK30WYTGXrxuvpXUyg4BDUk+nhFGLjoQ47/8nqS+whRBN8Fs71LQ9RwG3AJjZIkkfAalO/0UzWwMg6R1gX8KNRWWZaWZrgbWS1gBPxvQF8fs0BY4EHkp78m2UZ92pFfnKqmMacBqh0z+dcEORTmfgUzObA2BmX8Y6c7X3dp1+RSx7AEuWLOHUU0/ld7/7Haeffrpb9tJIuj5IvkbXVzWSrg+Sr3GnWPZSSPoW8H/AnmY2RNJBwBFmdke1K6ocbwNlvf9el7Yv4Hwzey69gKQRwDeBXvEJdhlVW4Cool74itS3Je14S6x7F2B1akSigvQA3i2njicIN0WtCCv2zciz7qztnS8TJkygR48e1K9fn4EDBzJixAiuuOIKli5dyrhx4+jZsydPPvlk+RU5juM4eb/Tn0zwaqcWV3mPsIxrUpgBNIpDxABIOkTS0VnKPgecK6lBLNdJYUnZ5sDK2OEPIDyZQ3Ap7Jbjui8DZ6bqAfYBsi1gU+PEJ+ulkr4f9UjSt9PLKEukPUn9CMPqt2epo7+kWbH+YmAOcBPwVAyuk85ioK2kQ+O5uylMnMzV3nkxa9YsTj75ZPr27cv06dP55S9/ybRp01i7di033HADHTt2pFWrVvlW5ziOU6fJt9Pf3cweJDxVYmZfEZ5eE4GZGfA94BgFy97bwNXAZ1mKTwLeAebFiXJ/Jjwp30uwny0gvCNfFOv+nPAeeqGk6zLq+iOwSzxnGjDCzDay8zgTGCXpTcLox0k5yp0WJya+B/yKsNreu1nqmMz2a/BPA34UP7fDzDYRhv9viee+QBgpydXe5bJ8+XKefvppzj777K1pkvjyyy+BYOnbc8+yFgJ0HMdx0sl3yHmdpNaAAUg6HFhTY6oqQVz8JVu43yVAUVq5LYSO7ldZyh6Ro+4fZiQdHNM3ADuEjTWzyYQOM3U8tCztOc5ZlnadzLz22c6L8xaOK+868dXFVYQJgQcDP5P0j3jz1Jng5V9PGNLvAGHGPnA84Wl/gqSr4oqNawiBewBaEaIvHmlpQXUkdWfbTWJP4GRgSpkNAowePZprr72WtWu3LQkxadIkjj/+eJo0aUKzZs2YPXt2edU4juM4kXyf9C8kdAD7S/o7IX76+TWmyikEBYu0Z2bHx3kCo4CPgDJX0ykp3cxTTz1FmzZt6NWr13Z5N954I8888wzLly9n5MiRXHjhhZX9/o7jOHUOhYe7HJnSPmb2z7hfn/AUKGCxmZXmPNHJiqRX2XFG/X+Z2YICXb/YzJpK6g+MNbNjY/qfgL8DC4GbzaxvTD8R+ImZDZX0OmG4/qtYXStgsIVIe0eyLYreLTmuvXss8wMzW5glP92y1+vYY4/l+eefp169emzatIn169fTvXt3Pv74Y+69914AVqxYwZgxY5g8eXK1tE++FBcX07Rp0/IL7iSSrg+Sr9H1VY2k64Pka6yMvgEDBsw1s95lFjKznBswL23/kbLK+pb8DSiOn/0Jk/FS6bcCIwhrC7yUln5iqhxhHYTOOer9KbACuDpHfj3gr8Dp+ejs1KmTpTNz5kw74YQTrLS01Fq3bm2LFy82M7NJkybZKaecYoVm5syZBb9mRUi6PrPka3R9VSPp+sySr7Ey+oDXrZzf13xCqqb4OvrznYpRI5H2CKsavmVmD1RU0ObNmzn77LNZu3Yt9evX56CDDqJbt25IomXLloleXMNxHCdplPdO33LsO7UQq4FIe5GLgUHaFkzoxHw13XTTTRx22GEceuihAFx66aVs2LCBkpIS+vfvz4wZ+S4X4DiO45T3pP9tSV8SnvibxH3isZlZsxpV51QrFiPtmVkR2zsazkvbfxbYYc1+Myth21K96RyTVmYu0C3LuZVaHD9l2Rs7diy///3vATj++OO35h922GEsX768MlU7juPUScp80jezembWzMx2M7P6cT917B2+g6TxkkanHf+vpAsyypyT9pS/VNLMfOpOWfZ22WXH/6alpaVMnTqV444r06HoOI7jpJGvZc9xcnEnYTEjFKLwnQ7ck17AzCZasOwdSohx8PuyKizLspfiZz/7GX379t26Fr/jOI5TPmVa9hwnHyS9APySEKDobDPLGgdB0h+BVWZ2ZZa8ci17Rx99NGPHjmXKlCksWbKE8ePHZx0FqGlqo9Wn0CRdo+urGknXB8nXuFMse775ls9GWH73JsLyvMfnKDMCeBrYpbz6cln2zMxuv/12O+KII2z9+vW2s6iNVp9Ck3SNrq9qJF2fWfI17izLnuPkw2PAeMLs/swli5HUizCD/2jLWK0vFxs2bKBv375s3LiRNWvW0LBhQwB+/OMf07BhQ1q1asVXX33FAQccwLvvvltObY7jOA5UPdyr42Bmm+LkvNW2Y/Q9gPMIK/jNDC4/Xjezs7OU20qjRo2YMWMGTZs2pbS0lKOOOorZs2enRg0AOPXUUznppFwxhRzHcZxMvNN3qkycwHc48P1s+Wa2Q1CiPOrc+j6rtLSU0tJS4g0DAF9++SUzZszgrrvuqpxox3GcOojP3ncqjaRDJS0GPgBmAdMlHZxRplxLXy42b95M9+7dadOmDcceeyx9+vTZmjd9+nS++93v0qyZO0cdx3HyxTt9p9KY2RzgIeARwqjRPbZjMJ1yLX2ZlJSGNwT16tVj/vz5LF++nNdee42FC7dVff/993PGGWfkqsJxHMfJglv2nCohqSEwB9gAHJntnX4+lr5My96DDz64Xf6UKVNo3Lgxp512GmvWrOGss87ioYce2jrBr5DURqtPoUm6RtdXNZKuD5Kv0S17viVyA9oShvffAXbNUaZcS1/61qlTJ1u5cqV98cUXZma2fv16O+qoo+zJJ580M7M//elPdtZZZ9nOojZafQpN0jW6vqqRdH1myddYU5Y9H953qsqfgcuBe4FrcpR5DDiOsCLfc/lUumzZMvbaay+aNGlCy5YtqVevHkOHDmXEiBGMHj2al19+me7duzN//vxq+RKO4zh1Ae/06wiSiitYvr+kp8opcxZQamb3EcLnHippYGY5M9sEzAQetOyWvh3o3bs3K1asoKSkhLVr11JSUsLs2bMBuOeee/jwww+ZP38+3bt3r8jXchzHqdO4Zc+pNGZ2N3B33N8M9MlWrjxLX45zyrTsOY7jOBXHn/TrGPEJvkjSw5IWSbpXsTeVdFxMmwecknbOrpLulPSapDcknRTTb5J0RdwfLOml2MGnzttF0kfAh8CLwAeS3pf0zXy05rLsjR07lkMOOYRf/OIXbNy4sZpaxnEcp/bjnX7dpAcwGjgI6AB8R1Jj4Hbg/wG9gD3Syo8FZpjZYcAA4DpJuwKXAadJGgDcDIy0tGV24/6dwAQzuwg4BnjTzFaVJa4sy97VV1/NokWLmDNnDv/5z3+45ppc0wgcx3GcTNyyV0eQVGxmTSX1B8aa2bEx/U/A34GFwM1m1jemnwj8xMyGSnodaAx8FatrBQw2s3clHQm8BPzCzG7Jct29gcfNrKekBwhe/h3mClTEspdi/vz5TJs2jauvvroKLVNxaqPVp9AkXaPrqxpJ1wfJ11hTlj1/p183SR8T30z5/w8EnGpmi7PkdQM+B/bMdqKZfSxpRZzgdxhwZo5ytwG3AXTu3Nm6du1KgwYNaNGiBSUlJVx++eWMGTOGzp0707ZtW8yM6dOn069fP/r371+O/OqlqKio4NesCEnXB8nX6PqqRtL1QfI11pQ+7/SdFIuA9pL2N7MPgPTl7p4Dzpd0vpmZpB5m9oakfYGLCK8LnpE03cxezVL3JMIqfFPznb3/6aefMnz4cDZv3syWLVv4wQ9+wNChQxk4cCCrVq3CzOjevTsTJ06s2rd2HMepQ/g7fQcAM9tAGF5/Ok7kW5mW/RtC2Ny3JL0N/CZO/rsDuNjM/gWMAibFuQGZPAE0BfKOjtOpUycaNGiAJMyMzZvDvcKMGTNYsGABAwcOZPr06YkennMcx0ka/qSfICRtBhYQOtivCHa4Gy3PGPQZdfUGzjKz/wEws6bxswgoSpUzs/PS9p8FDsyop4jQsf80HrcHnoqrPx2Tdu5cwlB/Nr5NmMC3KF/92ULrDhkyhMMPP5zXX3+dL774It+qHMdxnIg/6SeLEjPrbmZdgWOBIcCVlanIzF5Pdfg7E0mXEgLyXJaRXq+c87L69Ddv3swll1zCtddeW2OaHcdxaive6ScUM1tJGG4/T4F6kq6TNEfSW5JST94PSDohdZ6kyZKGpa+oJ6mppLskLYjnnhrTB0l6RdI8SQ9JqtBYuaSu0bs/P9bbMab/KJUO7Ad0MLO/SSqWdIOkN4EjctWbsuxl8+nfeuutnHjiibRt27YiUh3HcRzcspcoUra6jLTVQGfgJKCNmf1WUiOCze77QHfgZDMbHiPefQB0IqyOd3G03F0DNDKz0bHOlkA94FFgiJmtkzQmlhmfcf2iWM/r8bg9YXj/YEm3ALPN7N547XpAe+Ba4BQzK5X0x1jmbkkGnGZm2/vxKNuyV1xczOWXX86IESOYNGkSEyZMoF69egwZMoS//OUvlWrrqlAbrT6FJukaXV/VSLo+SL5Gt+w5g4BDJKXC0jYHOgJ/AW6KNwLHAS+ZWUnGkrXHEOLYA2BmX0gaSlic5++xbEPglSzXzXZXmEp7BRgrqR3wqJktkfRdwuI+c2K9Tdg2KXAzYah/xwozLHuZVpV58+axevVqVq1axahRowDYuHEjZ599Nu+//362KmuMumr1qU6SrtH1VY2k64Pka3TLXh1EUgdCR7mS4JU/38x2iFIXn8YHE0LYPpBv9cALZnZGOeU+B1qmHbcC/g1gZvdJehU4gWDZ+2msd4qZXbZDTbAhX8veqlWrtvPpv/DCC4wZM4bPPvtsa5mmTZsWvMN3HMf5OuPv9BNKXJ9+InBrnCn/HHCupAYxv1NcChdCnPqRwNHAs1mqewH4eVrdLYHZhOV3D4hpu0rqlOXcIuBH2jZ0MJwQMS91U/Khmd0MPA4cQlhjf5ikNrFMq+jnrxC5QuueeeaZdO7cmYMPPpiNGzdSWlpa0aodx3HqLN7pFwBJJ0sySQeWU7RJnBT3NvBXwrv5cTFvEvAOME/SQkIc+9RIzfNAP+CvZrYpvndP98T/FmgpaaGkD4F5hBuBbwILJJUAy4HRCuFy07kNWAu8GSfgNQWuj3k/ABbGCXsHA3eb2TvAr4HnJb0Vr1PhWXe5QuueeeaZLFq0iAULFjBs2DAmTZpU0aodx3HqLD68XxjOAP4WP3ew4Emqb2ZfmVm99OP0MtGr/6u4kZFXShh2T2edmQ2N+cWEJ/TM6xaRNkkvG2a2CTgvR97vgN9lSZ9GGH3ITM97Vkouy97xxx+/tcxhhx3G8uXL863ScRynzuNP+jVMtMEdRVix7vS09P6SXpb0BPBO5nEsUxw/c9ny2sdz5sXtyCpqvUrSxXG/SNKNkl6X9K6kQyU9KmmJpN+mnbPVnifpz8rw30tqLmmxpM7x+H5JPy5LR1mWvRSlpaVMnTqV4447ripf2XEcp07hT/o1z0nAs2b2nqTPJfWKq9cB9AQONrOlCtHvth5n1DGNMJT+dLTGfRc4lzBp7lgz2xA98vcDZdo1KsgmM+st6QLCO/tewH+ADyTdCLQhTB78Tpo970zCSoIAmNkaSecBkyXdBLQ0s9szL5Rh2aOoqAiACRMmbLXsHXjggey3334AXH/99XTo0IHNmzdvLVsoiouLC37NipB0fZB8ja6vaiRdHyRfY03p806/5jkDuCnuPxCPU53+axkdfOZxily2vObArZK6E2b5Z5uIVxWeiJ8LgLfN7FOAOC9gb8IIRi573lbM7AVJ3wf+QFiSdwfysex9/vnnjBw5knHjxlG/fn0efPBBdtml8INVddXqU50kXaPrqxpJ1wfJ1+iWva8hkloBA4FucWGaeoBJuiQWWZdxSuYxEILh5LDl/QJYQehIdwE2VOsX2BaCdwvbh+PdQvi/U5Y9byuSdgG6AOsJ9r9yX8TnsuxNmjSJ5557jhdffHGndPiO4zhfZ/xXs2YZRggnu6+ZtTezvYGlBGtdRclmy2sOfBon+f0X4aaikORrz/sF8C7wQ+CulO2wLHJZ9s455xzmzZtHixYtaNmyJVdeWanQBI7jOHUS7/RrljOAxzLSHmH7WPX5sp0tL6b9ERgerXQHkmOkIBvR1ndoRnL/eI28yMeeFyfwnQ1cZGYvAy/Fc8okl2XvlFNOYfLkyWzcuJHTTjuNPfbYI1+5juM4dR4f3q9BzGxAlrSb0w6L0tKL0o9jWtO0/R1seWa2hLAgTooxMX0ZwTdfHksz7HpFQHGso38ubRl529nzJG33f8rMFhOG9lPHF+ahK6dlb8aMGdx3330ADB8+nKuuuopzzz03nyodx3HqPP6k72RF0v9Iekchet4DMW1XSXdGi94bkk6K6SMkPSFpBmHIP72e8ZJGpx3/b3QD5CSXZW///fenRYsW1K8f7ivatWvHJ598Uq3f23EcpzbjT/q1GEmDgWsykpea2ffyOP1SYD8z2yipRUwbC8wws/+Oaa9J+mvM6wkcYmb/yajnTkI0vwlxQt/pwGFZtJZr2WvXrh0lJSVb81auXMm6devcspdB0vVB8jW6vqqRdH2QfI1u2XMqTAzOs0OAnlR2OelvAfdKmg5Mj2mDgBNTC/gAjYF94v4LWTp8zGxZXJ+gB/At4A0z+zxLuXItexs2bGDjxo0cddRR1K9fn1deeYVOnToV3HZTV60+1UnSNbq+qpF0fZB8jTWlz4f36y6Z0fMgLYIeIXLeHwhP8HPiu3oBp5pZ97jtY2bvxvJlTSKcBIwguA/uzEfcqlWrWL16NcBWy16XLl0YMGAADz/8MABTpkzhpJNOyqc6x3EcB+/06yxxPf5PJQ2ErWsKHAf8LQ7D721mMwmTA5sTAu08B5yfirgXn97z4bFY96HkHnnYjjfeeIO9996bxo0b07JlS5o2bcrQoUO55pprGDNmDA0bNuS+++7z0LqO4zgVwIf36zZnAX+Q9Pt4PM7MPog++nviin8Cbjaz1ZJ+A0wA3oo3BkuBoeVdJEb+mwmsNrPN+Qjr1q0bs2bNomfPnqxdu5ZevXrxzjvvsGLFCjp27Mh7771Ho0aNWLlyhwUAHcdxnBx4p1+HiT77bLbCUsISu5npJcBPs6RPBibnuk68QTgc+H6+2tq2bUvbtsHyv9tuu9GlSxc++eQTbr/9di699FIaNWoEQJs2bfKt0nEcp87jw/tOjSLpB0AJwef/L0lvS8pnDYGtLFu2jDfeeIM+ffrw3nvv8fLLL9OnTx/69evHnDlzakK24zhOrURmuSZxO071EEPxNiYE5FluZldnKZNu2ev14IMPAmES3wUXXMCPfvQj+vbty8iRI+nRowfnn38+ixYtYvz48dx3333EaQYFobi4eOvCQUkk6fog+RpdX9VIuj5IvsbK6BswYMBcMys70qqZ+eZbjW5AQ+BN4FWgXnnlO3XqZGZmmzZtskGDBtkNN9xgKQYPHmwzZszYetyhQwdbuXKlFZKZM2cW9HoVJen6zJKv0fVVjaTrM0u+xsroA163cn5ffXjfKQStCbP/dyM88ZeLmTFq1Ci6dOnChRduW7n35JNPZubMmQC89957bNq0id13370GJDuO49Q+vNN3CsGfgcuBe9lxhcCsPProo0ydOpWJEyfSuHFj9tprL5555hn++c9/csstt9CkSRN69+7NDTfcUNChfcdxnK8z3uk7NYqks4BSM7sP+B1waGptgLI48sgjmTt3Lhs2bGDVqlXsuuuutG/fnksvvZQvvviCkpISfvvb3zJjxowa/w6O4zi1BbfsOTWKmd0N3B33NwN98jkvl2XvoIMO2lpm3bp1/pTvOI5TAbzTd2ocSecA58TD5sAyyxJ2OBfplj2AsWPHcvfdd9O8efOt7/cdx3Gc8nHLnlMw4kp/M4BrzezJjLy8LHvp3HvvvWzatImRI0cWRH+K2mj1KTRJ1+j6qkbS9UHyNbplz7ev/Qb8kbDUb5nlyrLspfPRRx9Z165ds+bVJLXR6lNokq7R9VWNpOszS75Gt+w5X2skjQD2BcblU94su2VvyZIlW/cff/xxDjzwwGpW6jiOU3vxd/pOjSOpF3AxcLSZbcnnnJRlr1GjRkycOJHWrVtz++23M3r0aD766CN22WUXmjdv7rP3HcdxKoA/6TuF4DygFTBT0nxJk8o7IZdl79Zbb2XdunWUlJQwfPhwpkyZUvPqHcdxagn+pO/UOGZW4Zl2uSx7gwYN2lrm8MMP5+GHH64+oY7jOLUcf9J3ahRJ4yWNTjv+X0kXVKSOTMv1m80yAAARkUlEQVReijvvvJMhQ4ZUk1LHcZzaj1v2nBpFUnvgUTPrKWkXYAlwmJl9nlGuQpa9e+65h8WLFzN+/PiCL9BTG60+hSbpGl1f1Ui6Pki+Rrfs+fa13YAXgB7AccDD5ZUvz7J311132eGHH27r1q2znUFttPoUmqRrdH1VI+n6zJKvsaYse/5O3ykEk4ARwB7AnfmcYJbdsvfss89y7bXXMmvWLL7xjW/UiFjHcZzainf6TiF4DBgPNAB+mM8JQ4cO5ZlnnqFRo0YUFRUBMGrUKC655BK2bNnCvvvuS7t27Rg4cCATJ06sMeGO4zi1CZ/I59Q4ZrYJmAk8aCHoTrmMGTOGuXPncsABBzB//nzmz5/Pk08+yWOPPcamTZt4+OGH2XPPPb3DdxzHqQDe6RcISSdLMkkVWkJO0j8qeb32khbmyBsc/fLzJRVLWhz375Z0TgyHW23ECXyHA3fke07fvn1p1apVZj18+eWXAKxZs4Y999yzOmU6juPUenx4v3CcAfwtfl6ZmSmpvpl9lXlsZkdWtxAzew54Ll6nCLjYzF6v7uvE+g8CngIeM7Ml5ZUviwkTJjB48GAuvvhitmzZwj/+Uan7IcdxnDqLW/YKgKSmwGJgAPCkmXWO6f2B3wBfAAcSLGtbj82sk6RiM2sq6QFgqpk9Hc+dTOhMXwemArvGy51nZv+IVrmnzOzgcrQVkdbpS7oKKDaz62PeG8DRsf6zgMuAbsA0M/t1POdHwP8ADYFXgZ+lD+NL6k2YzAdQDzjYzLbz2WWz7H322Wdcdtll3HXXXQDcfPPNfPvb36Zfv37MnDmTp556ihtuuKGsr1cj1EarT6FJukbXVzWSrg+Sr9Ete1/jDTgTuCPu/wPoFff7A+uA/bIdx7Ti+Pk9YErcbwh8DDQBvgE0jukdiZYNoD2wMA9tRUDvtOOrCDcBqbxr4v4FwL+AtkAjYDnQGugCPAk0iOX+CJxVxvWuA64rS1PKsrd06dLtoug1a9bMtmzZYmZmW7Zssd122812BrXR6lNokq7R9VWNpOszS75Gj7L39eYM4IG4/0A8TvGamS0t4zjFX4ABkhoBQ4CXzKyEMCP+dkkLgIeAg6pZ+xPxcwHwtpl9amYbgQ+BvYHvAr2AOZLmx+MO2SqSdBrQE7i0MkL23HNPZs2aBcCMGTPo2LFjZapxHMeps/g7/RpGUitgINBNkhGGt03SJbHIuoxTMo8BMLMNcbh9MHAa224ifgGsAL5NmJi5oVq/AGyMn1vS9lPH9QERRiAuK6sSSQcTRhH6Wh4z+Dt06MBHH33Eli1baNeuHePGjeOHP/whQ4YMYcOGDXTt2tWD7TiO41QQf9KveYYR3sXva2btzWxvYCnhPXlFmQaMjOc+G9OaA59aCFn7X0C9+MT9PLC/pIvi7PnKsBshJG5ZvAgMk9QGwk2OpH3TC0hqAdxPGPZflc+FJ0+ezJw5c+jatSvLly9n1KhRDBs2jPnz59OvXz8mT55Mr169KvGVHMdx6i7e6dc8ZxAWp0nnEbYf4s+X54F+wF8teN8hvEMfLulNwmRAzKw7MAhYRngVsINbIE/WAteXVcDM3gF+DTwv6S3CkrttM4qdBOxLeA0xP96UlEk2y16XLl3o3LlzBeQ7juM46fjwfg1jZgOypN2cdliUll6UfhzTmqbtlxLi0qfnLwEOSR1L+nlMXwZ0kdSB8L79KsJN3u8IEwYbAX8ws/6SHpA01cyeNrOrJE2WNIwwHH8VMJTgElgV5w4YMM7MXpc0CBhNGO5fAow0s+I0PfWB84ATzaxI0tWxrOM4jlNgvNOv5ZjZh5LqAW0IT9xrzOzQOCHw75KeJ7w2+AHwtKSGhMl45wLpsWwvj+d2A5DUUtLuhKf8Y8xsnaQxwIWEJXdT1/9K0gjgYUnnE4LubB8jlx0sexQVFfHZZ5+xbt26rcvwpli9ejVz586luLg4s5qCUFxcvIOmJJF0fZB8ja6vaiRdHyRfY43pK296v29fr41o8Us7HgxsBhYCqwkT/VYD8wlzCwYBjYF/Ep7+TwLujef2J3j9AeYCHTPqHgr8O9Y1H3iHaE3MoutX8do9yvsOuSx7Kfr162dz5szZIb1Q1EarT6FJukbXVzWSrs8s+Ro9yp5TWZYQOvluwMPAbRZW5NuOHM6A8hDwgpnlMz+hW9TRJs+6HcdxnGrGJ/LVYiR9E5gI3BrvAp8DzpXUIOZ3kpRayS+bMyCdF4Cfp9XdEpgNfEfSATFtV0mdsug4hTAXoS9wS5zNXyZnnHEGhxxyCG+//TYNGjTgsMMOY9q0abRr145XXnmFE044gcGDB+fdFo7jOI53+rWRXSWVSNpAGLLfwLZ37JMIQ/DzYjCeP7NtXkc2Z0BnoGvc/y3QUtLC6BQYYMF+NwK4P87cf4XoIEgR3/v/DjjbzN4DbgVuKu9LXH/99bRu3Zr169dTWlpK+/btKSkpYfny5WzcuJEVK1bw3HM7DFg4juM4ZeDD+7WPdRZn/Efv/H3AFcCVFrz8v4rbdlh2Z8CfCTcGWJiRPzzLeTOAQ3OJMbN/A53Sjm/OVTaTr776ipKSEho0aMD69es9qp7jOE4V8Sf9WoyZrSTMiD9PgXqSrpM0R9Jbkn4KEC17J6TOS1n2JPWX9FRMayrpLkkL4rmnxvRBkl6RNE/SQzG4EGl17ZkWxne+pM2Zi/dkY6+99uLiiy9mn332oW3btjRv3pxBgwZVZ/M4juPUOTzKXi0jFZUvI201Yaj+JKCNmf02ZdkDvg90B042s+HRsvcB4em8DyH4zlBJ1wCNzGx0rLMlYUnhR4Ehts2y18jMxpOFuIZAPzP7QZa87aLs3XHHHVx55ZVcccUVNG3alKuuuop+/fpx7LHHVr2RqkhtjM5VaJKu0fVVjaTrg+RrrKkoez68X7cYBBwSF96BsIRvR0Iwn5vijcBxxGA+0nbRb48BTk8dmNkXkoYSAvz8PZZtSHivvwOSvgP8GDgqW76Z3QbcBtC5c2fbsGEDPXr04OSTTwbgX//6F7Nnz6Z///6V+d7VSlFRUSJ05CLp+iD5Gl1f1Ui6Pki+xprS551+LSeuyLcZWEmw2J1faMuepLbAHYRV+fJaUWefffZh9uzZrF+/niZNmvDiiy/Su3fZYaIdx3GcsvF3+rWYJFj24rUeAsbE2ft50adPH4YNG0bPnj3p1q0bW7Zs4Sc/+Um+pzuO4zhZ8E6/9tEkTph7G/grwYo3LuZV1LKXTqUse8CRQG9gXNpkvrym4Y8bN45FixaxcOFCpk6dSqNGjfJrAcdxHCcrPrxfyzCzemXkVdSyV0QMAFQFy94swjK/juM4zk7Gn/Qdx3Ecp47glj0ncUhaCyze2TrKYHdCoKGkknR9kHyNrq9qJF0fJF9jZfTta2bfLKuAD+87SWRxeV7TnYmk111f1Ui6RtdXNZKuD5Kvsab0+fC+4ziO49QRvNN3HMdxnDqCd/pOErltZwsoB9dXdZKu0fVVjaTrg+RrrBF9PpHPcRzHceoI/qTvOI7jOHUE7/Qdx3Ecp47gnb6TGCQdJ2mxpPclXVrA6+4taaakdyS9LemCmN5K0guSlsTPljFdkm6OOt+S1DOtruGx/BJJO6xgWEWd9SS9IempeLyfpFejjmkxLDKSGsXj92N++7Q6LovpiyUNrmZ9LSQ9LGmRpHclHZGkNpT0i/jvu1DS/ZIa7+w2lHSnpJVxWexUWrW1maRekhbEc26Wtg+dWUl918V/47ckPSapRVpe1rbJ9bedq/2roi8t7yJJJmn3eJyI9ovp58c2fFvStWnpNd9+Zuabbzt9A+oBHwAdCCF63wQOKtC12wI94/5uwHuEkMHXApfG9EuBa+L+8YRwxAIOB16N6a2AD+Nny7jfshp1XgjcBzwVjx8ETo/7E4Fz4/7PgIlx/3RgWtw/KLZrI2C/2N71qlHfFODsuN8QaJGUNgT2ApYCTdLabsTObkOgL9ATWJiWVm1tBrwWyyqeO6Qa9A0C6sf9a9L0ZW0byvjbztX+VdEX0/cmBBj7CNg9Ye03gBAXpVE8blPI9qvxH1TffMtnA44Anks7vgy4bCdpeRw4lrAqYNuY1pawaBCEQEVnpJVfHPPPAP6clr5duSpqage8CAwEnoo/Qv9O+/Hd2n7xx+6IuF8/llNmm6aXqwZ9zQmdqjLSE9GGhE7/4/jDXj+24eAktCHQPqNTqJY2i3mL0tK3K1dZfRl53wPujftZ24Ycf9tl/R+uqj7gYeDbwDK2dfqJaD9CR31MlnIFaT8f3neSQupHOcXymFZQ4jBuD+BV4Ftm9mnM+gz4VtzPpbUmv8ME4JfAlnjcGlhtZl9ludZWHTF/TSxfk/r2A1YBdym8gpikELY5EW1oZp8A1wP/BD4ltMlcktWGKaqrzfaK+zWp9b8JT8CV0VfW/+FKI+kk4BMzezMjKynt1wk4Og7Lz5KUClhWkPbzTt9xIpKaAo8Ao83sy/Q8C7fSO8XfKmkosNLM5u6M6+dJfcIw5p/MrAewjjA0vZWd3IYtgZMINyd7ArsCx+0MLRVhZ7ZZeUgaC3wF3LuztaSQ9A1CFNErdraWMqhPGHE6HLgEeLCicwWqgnf6TlL4hPAeLkW7mFYQJDUgdPj3mtmjMXmFpLYxvy2wshytNfUdvgOcKGkZ8ABhiP8moIWkVPyM9Gtt1RHzmwOf16A+CE8Zy83s1Xj8MOEmIClteAyw1MxWWQgj/SihXZPUhimqq80+ifvVrlXSCGAocGa8MamMvs/J3f6VZX/Cjd2b8e+lHTBP0h6V0FdT7bcceNQCrxFG73avhL7KtV9F30/45ltNbIS73w8Jf7CpySpdC3RtAXcDEzLSr2P7CVXXxv0T2H5C0GsxvRXhvXbLuC0FWlWz1v5sm8j3ENtP4vlZ3P85209CezDud2X7iUIfUr0T+V4GOsf9q2L7JaINgT7A28A34jWnAOcnoQ3Z8Z1vtbUZO05EO74a9B0HvAN8M6Nc1rahjL/tXO1fFX0ZecvY9k4/Ke13DjA+7nciDN2rUO1XbT9GvvlW1Y0wu/Y9wkzVsQW87lGEIdS3gPlxO57wzuxFYAlhtm3qh0DAH6LOBUDvtLr+G3g/biNrQGt/tnX6HeKP0vvxjz81G7hxPH4/5ndIO39s1L2YCs5EzkNbd+D12I7T4w9oYtoQGAcsAhYCU+OP605tQ+B+whyDUsIT4KjqbDOgd/y+HwC3kjHRspL63id0VKm/lYnltQ05/rZztX9V9GXkL2Nbp5+U9msI3BPrnQcMLGT7+TK8juM4jlNH8Hf6juM4jlNH8E7fcRzHceoI3uk7juM4Th3BO33HcRzHqSN4p+84juM4dYT65RdxHMepHUjaTLBrpTjZzJbtJDmOU3Dcsuc4Tp1BUrGZNS3g9erbtrXRHWen48P7juM4EUltJb0kab6khZKOjunHSZon6U1JL8a0VpKmx9jssyUdEtOvkjRV0t+BqZK+KekRSXPi9p2d+BWdOo4P7zuOU5doIml+3F9qZt/LyP8hITzp/0qqB3xD0jeB24G+ZrZUUqtYdhzwhpmdLGkgYSnn7jHvIOAoMyuRdB9wo5n9TdI+hJCpXWrwOzpOTrzTdxynLlFiZt3LyJ8D3BkDME03s/mS+gMvmdlSADP7Tyx7FHBqTJshqbWkZjHvCTMrifvHAAelBVJrJqmpmRVX39dynPzwTt9xHCdiZi9J6ksIzjJZ0u+BLypR1bq0/V2Aw81sQ3VodJyq4O/0HcdxIpL2BVaY2e3AJEJ44NlAX0n7xTKp4f2XgTNjWn/g32b2ZZZqnydE9Etdo6yRBsepUfxJ33EcZxv9gUsklQLFwFlmtkrST4BHJe1CiG9/LCF88J2S3gLWA8Nz1Pk/wB9iufrAS4Twqo5TcNyy5ziO4zh1BB/edxzHcZw6gnf6juM4jlNH8E7fcRzHceoI3uk7juM4Th3BO33HcRzHqSN4p+84juM4dQTv9B3HcRynjvD/Ae7GGkBPHyLCAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "plot_importance(xgb_model, max_num_features=30)\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "onOs__yR0ei-",
        "outputId": "0214b3a6-b051-4d21-eb0d-d0cc9601e5a8"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "plot_importance(xgb_model, max_num_features=35)\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "CdTlAYWc0gRf",
        "outputId": "68d0c1c1-8c43-466e-b186-7bfb7c260d79"
      },
      "outputs": [
        {
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAf0AAAEWCAYAAABsT07JAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOydeZyPVfvH39fY9+WxZJ8oY6wTSqkYeVCiKE+lPEI9PYXSoqQFrSgqotKjFKnQYpCkHyalFDG2YbTYd2Md2wxz/f445zu+853tO6uvmfN+ve7X3Pc55z7nur+ep3Pf53yu6xJVxeFwOBwOR/4n6EIb4HA4HA6HI29wk77D4XA4HAUEN+k7HA6Hw1FAcJO+w+FwOBwFBDfpOxwOh8NRQHCTvsPhcDgcBQQ36TscDgcgIs+IyOQLbYfDkZuI89N3OBzZRUS2AlWBc17F9VV1dzb7vF9V/y971l18iMgI4DJV7XWhbXHkL9yXvsPhyCm6qmppryPLE35OICKFL+T4WeVitdtxceAmfYfDkWuISDkR+UBE9ojILhF5WUQK2bp6IrJYRGJF5KCITBeR8rZuGlAbmCsicSLylIiEi8hOn/63isg/7fkIEflCRD4RkWNAn/TGT8XWESLyiT0PFhEVkb4iskNEDovIgyJypYisFZEjIjLB694+IrJMRCaIyFER2SQi7b3qq4vIHBE5JCJ/ish/fMb1tvtB4BngTvvsa2y7viKyUUSOi8jfIvJfrz7CRWSniDwhIvvt8/b1qi8hImNFZJu17ycRKWHrrhaRn+0zrRGR8Cz9YzsuCtyk73A4cpOPgLPAZcAVQEfgflsnwEigOhAK1AJGAKjqv4HtnF89eM3P8W4FvgDKA9MzGN8fWgGXA3cCbwHPAv8EGgF3iEhbn7Z/AZWA4cBXIlLR1n0O7LTP2gN4VURuSMPuD4BXgRn22ZvZNvuBLkBZoC/wpog09+rjEqAcUAO4D5goIhVs3RigBdAaqAg8BSSKSA3gG+BlWz4Y+FJEKmfiN3JcRLhJ3+Fw5BSz7dfiERGZLSJVgc7Ao6p6QlX3A28CdwGo6p+q+r2qnlHVA8AbQNu0u/eLX1R1tqomYibHNMf3k5dU9bSqLgROAJ+p6n5V3QX8iHmR8LAfeEtVE1R1BhAD3CwitYBrgSG2ryhgMtA7NbtV9VRqhqjqN6r6lxp+ABYC13s1SQBetOPPB+KAEBEJAvoBg1R1l6qeU9WfVfUM0AuYr6rz7djfAyvt7+bIh7i9I4fDkVN08xbdichVQBFgj4h4ioOAHba+KjAOM3GVsXWHs2nDDq/zOumN7yf7vM5PpXJd2ut6lyZXRm/DfNlXBw6p6nGfupZp2J0qInITZgWhPuY5SgLrvJrEqupZr+uT1r5KQHHMKoQvdYB/iUhXr7IiwJKM7HFcnLhJ3+Fw5BY7gDNAJZ/JyMOrgAJNVPWQiHQDJnjV+7oWncBMdADYvXnfZWjvezIaP6epISLiNfHXBuYAu4GKIlLGa+KvDezyutf3WZNdi0gx4EvM6kCEqiaIyGzMFklGHAROA/WANT51O4BpqvqfFHc58iVued/hcOQKqroHswQ9VkTKikiQFe95lvDLYJagj9q95Sd9utgH1PW63gwUF5GbRaQI8BxQLBvj5zRVgEdEpIiI/AujU5ivqjuAn4GRIlJcRJpi9tw/SaevfUCwXZoHKIp51gPAWfvV39Efo+xWx4fAG1ZQWEhErrEvEp8AXUWkky0vbkWBNTP/+I6LATfpOxyO3KQ3ZsKKxizdfwFUs3UvAM2Boxgx2Vc+944EnrMagcGqehToj9kP34X58t9J+qQ3fk7zK0b0dxB4BeihqrG2ricQjPnq/xoYnkH8gVn2b6yIrLIrBI8AMzHPcTdmFcFfBmO2AlYAh4DRQJB9IbkV4y1wAPPl/yRubsi3uOA8DofDkU1EpA8mkNB1F9oWhyM93Nucw+FwOBwFBDfpOxwOh8NRQHDL+w6Hw+FwFBDcl77D4XA4HAUE56fvCDjKly+vl1122YU2I01OnDhBqVKlLrQZaRLo9kHg2+jsyx6Bbh8Evo1Zse/3338/qKrphlB2k74j4KhatSorV6680GakSWRkJOHh4RfajDQJdPsg8G109mWPQLcPAt/GrNgnItsyauOW9x0Oh8PhyEH69etHlSpVaNy4cVLZiBEjqFGjBmFhYYSFhTF//nwAYmNjadeuHaVLl2bgwIFJ7U+fPs3NN99MgwYNaNSoEU8//XRS3bZt22jfvj1NmzYlPDycnTszCldxHjfpOxwOh8ORg/Tp04cFCxakKH/ssceIiooiKiqKzp1NTqPixYvz0ksvMWbMmBTtBw8ezKZNm1i9ejXLli3j22+/TSrv3bs3a9euZdiwYQwdOtRv29ykHyCISDebv7tBJu75OYtjBYvI+jTqOolIlD3iRCTGnk+1+cR7p3afw+FwOAxt2rShYsWKGTcESpUqxXXXXUfx4sWTlRcvXpx27doBULRoUZo3b570RR8dHc0NN5jMzO3atSMiIsJv29ykHzj0BH6yf5MhIoVTu1bV1jlthKp+p6phqhqGSbF5j73urarvqerUnB7T4XA4CgITJkygadOm9OvXj8OH/U8oeeTIEebOnUv79u0BaNasGV99ZaJWf/311xw/fpzY2Nj0ukjC+ekHACJSGpN7ux0wV1VDRCQceAkTZ7sB8ID3tarWF5E4VS0tIp9jMmV9Y/v7CJiHmbSnAR4J6EBV/VlEgoF5qnp+wyl1uyKBwaq60l6PAOJUdYytW41Ji1oKE+N8KNAEmKGqz9l7emFihhfFxCbvr6rnUhnrAfuMVKpUucWwt/7n78+X51QtAftSzXgeGAS6fRD4Njr7skeg2we5Z2OTGuUA2Lt3L0OHDmXKlCkAHDp0iHLlyiEifPjhh8TGxjJkyJCk+xYsWEBMTAyDBg0CIC4ujtKlS3Pu3DmeeeYZrrzySnr06AHAwYMHGT9+PHv27KFp06YsXbqUKVOm0LVr199VtSXpoaruuMAHcA/wgT3/GWgBhGMSilxqy5Nd27I4+7c78LE9L4pJmlECk4a0uC2/HFhpz4OB9X7YFQm09LoegXkJ8NSNtueDMIlEqmEyge0E/oHJMjYXKGLbvQP0zmjc+vXrayCzZMmSC21CugS6faqBb6OzL3sEun2quW/jli1btFGjRn7XTZkyRQcMGJB07bGvb9+++vDDD6c5zvHjx7VGjRqqqur5b3x6h1veDwx6Ap/b8885v8T/m6pu8Wrne+3hW6CdTZV5E7BUVU8BRYD/icg6TNauhjlstyfL1zpgg6ruUdUzwN9ALaA95gVmhYhE2eu6qfZUQNmxYwft2rWjYcOGNGrUiHHjxgEQFRXF1VdfTVhYGC1btuS3334D4PDhw3Tv3p2mTZty1VVXsX79eWmGRzHct2/fC/IsDocjbfbs2ZN0/vXXXydT9qfFc889x9GjR3nrrbeSlR88eJDExEQARo4cSb9+/fy2w/npX2BEpCJwA9BERBQoBCgm1egJn+a+1wCo6mm73N4JuJPzLxCPYfJyN8PoN07nsPln7N9Er3PPdWFAMCsQ/ktLCxiFCxdm7NixNG/enOPHj9OiRQs6dOjAU089xfDhw7npppuYP38+Tz31FJGRkbz66quEhYXx9ddfs2nTJgYMGMCiRYsAoxgeOHAgt99++wV+KoejYNOzZ08iIyM5ePAgNWvW5IUXXiAyMpKoqChEhODgYCZNmpTUPjg4mGPHjhEfH8/s2bNZuHAhBw4c4JVXXqFBgwY0b94cgIEDB3L//fcTGRnJ0KFDERHatGnDxIkT/bbNTfoZICLdMPmvQ1V1Uybu+1n9E9r1wOzH/9feFwxswOyV+1LZfjEDXAaUtNdrgWPAMKAq0Me2KQfsVNVEEbkX80KRWeaISFNVPWiv64nIPD/vXQREiMibqrrfvuCUUdUMA0gUFKpVq0a1aia9e5kyZQgNDWXXrl2ICMeOHQPg6NGjVK9eHTCqXY+/boMGDdi6dSv79u2jatWqtGnThq1bt16Q53A4HOf57LPPUpTdd999abZP7f+3+/fv92yrpqBHjx5J+/uZxU36GeOtqh/uWykihVX1rO+1nxO+p//RPmVHbflfPuUHVLWLHScSuEqNyh4RKYL5qo9Q1Xjb/h3gS+tmt4A0VgpyC1WNFpHngIUiEgQkAANEZDtGRJqYl/YEOlu3bmX16tW0atWKt956i06dOjF48GASExP5+WfjnelR7V5//fX89ttvbNu2jZ07d1K1atULbL3D4bgYcOr9dEhNVW/LwykgynoR2YoR8x30evbBqtpFRNoC42xTBdqo6nEReRK4AyPq+1pVh9vn+s6O0wLo7P3FXxDV+x6VL8CpU6cYNGgQvXr1ok2bNowfP55mzZrRtm1blixZwrx58xg7diwnTpxgwoQJ/PHHH9StW5ft27czePBgPLkK9u7dy5AhQ/j444+zb2Au4lEmByrOvuwR6PZB4NuYFfvatWvn1PvZOUhFVa+pKOl9rzUfKeuBrUAlr+twzEsJ9v5r7XlpzMpRR+B9zH5+EOYFp419rkTg6oyeraCp9+Pj47Vjx446duzYpLKyZctqYmKiqqomJiZqmTJlUtyXmJioderU0aNHjyaVbdmyRYODg3PUvtwg0NXdzr7sEej2qQa+jVmxD6fezzZpqeohfyjrWwPHROQU0A94TUS+9ukntaUgT9ky4A0ReQQor2abo6M9VgOrMCshl9v221R1eY48XQCQlvL+0KFDdOjQgcsvv5wOHTokC8IRGRlJWFgYjRo1om3btqgqd911Fxs2bGDy5MlJ/VSvXp0ffvgBgMWLF3P55eYnPHLkCPHxZvdm8uTJtGnThrJly+bxkzscjosVN+mngZeqfrJd4n4SuENExDbxW1mP+fL2KOtn2CpvZX1LzCpATuKPsv59VS1hj+JAaVXt7tNPLFDB67oi4BH1Lcd4BJQAltkQwgKMVBvVT1UvU9UPbPs81RTkNh7lfXR0NMuXL2fixIlER0czatQo2rdvzx9//EH79u0ZNWoUYCbs/v37M2fOHDZs2MCsWbNYtmwZX331FSVKlKBo0aIUKlSI1157jWeeeYYnnniCZs2a8cwzz/D+++8DsHHjRho3bkxISAjffvtt0osGGMXwNddcw44dO6hZsyYffPBBqnY7HI6CixPypU0yVT2AiPxA6qr6jJgB3I+Z3PvYspxQ1meH1JT1kkq7SODfwDARKQT0AmbbuurAcVUdLSJXYr7qvwNeEpHpqhonIjUwAr58R1rK+4iICCIjIwG49957CQ8PZ/To0Xz66afcdttt1K5dG4AqVapQpUqVFArdW2+9lUsuuYTff/89xZjXXHMNmzdvTtUej2I40FOGOhyOC4f70k+bnhhXPW++JJXY+H6wEGgL/J8mV9bfKyJrMJNlnivrAY+yfi3wPXbSF5FwEYkUkS+AW4De1s7VwDngGRFZBQwB2tj7E4FuwMtADWC93bpYiRERApQWkaVWyZ+v8Fbe79u3L+ll4JJLLmHfvn0AbN68mcOHDxMeHk6LFi2YOjVlGgPvfhwOhyOncep9RxJeHgfhQATQCCMCXIbZ3lgJ/IHZ9vgTs4JRUo2S/1UgWlU/EZHywG/AFZj9/xXAQOA9jGrf1xXxolPvV6mYtvK+S5cuzJt3PpRB165dmTt3LuPGjSMmJoaxY8cSHx/PgAEDGDlyJLVq1Uq1n6wS6KpkCHwbnX3ZI9Dtg8C30an33ZHrB+c9DsKB773K38Us64dhhIie8ls4r+RfCawHouyxHRPQCIxg8CzwsD92XEzq/dSU9/Xr19fdu3erquru3bvV8zwjR47UYcOGJbXr16+fzpw5M81+csK+QCXQbXT2ZY9At0818G106v0Chk9ee8/hu92Qm3iL/86Rsf5DgNv1vICvtqputHVNMILA6rlgZ57giWvvHS/7jjvuoGrVqqxatYrx48cTFhYGwGWXXUbz5s0JCwujefPmbN68maioKG699Va+/PJLmjRpQmhoKHPmzCE0NBRV5b777iM0NJTHH3/8Qj2iw+EoALhJP0BRr7z2Xoevsj6v2QQEi0g9e+2tb/gOeNjj3SAiV9i/dYAnMEv9N4nIRblZ3adPHxYsWJCs7JFHHuHw4cNUq1aNkydPsm/fPubPn8/UqVNp2LAhJ06coE6dOtSpU4ewsDCqVKnCzp07iY+Pp1ChQgQHB7Nv3z6WLVvGtGnTWLx4MWFhYYSFhTF//vwL9KQOhyM/UyDV+yJyCfAWcCVwBOM696iqpi6L9r/f8sDdqvqOva4OjFfVrAVJTtn/r5gAOxUxbnK7bFV/4PGcGsdrvBFAnOdaTWKfYsACETkO/AiUsdUvYX7TtVaot0VEugIfYIIG7RaR+4CPRORKNa6MFw2pxbW/7rrrkpbMateuncyf3pME55lnnsHj5fn333/TokWLpLpp06bx5Zdf8s4773i2SxwOhyNXKXCTvv0S/RoTJe8uW9YMk6hms1e7ZDH1/aQ8ZgJ+B0BVd2Nc/3IEVW1lbeuDicY30Kv65xzov7T9GwlE2kkfn3FOAdfo+QQ8nntPAf8lJf/0avM7Jptgvvrf3Y8//kjVqlWTJnxvZsyYQUREBGCW/WNiYti6dSs1a9Zk9uzZSYF2HA6HIy/IV//x9ZN2QIKqvucpUNU1kDKmvoiEAqMwwrZiwERVnWRj8kdggtYUAZ5T1Qjbtp7NfPc9MBEbR19EimMEcS0xorbHVXWJncBvwYTlrYeJVf9UZh7IO16/7a8bJub+5cAYTOCff2P26Tur6iG7RD8RqAycBP6jmcsiWAqYCdTExBh4SVVniEgL4A1MWN6DQB9V3WPzAUQB1wGfAWPT6vtUwjmCn/7G7+fPbbaOujnd+s8++4yePVN6cv7666+ULFkySQdQoUIF3n33Xe68806CgoJo3bo1f/2VwpHB4XA4co2COOk3BlJGPTlPc6Cxqm6xbmRHVfVKu6y9TEQWYuLnd1fVYyJSCVguInOAp+29nsx3wV79DgBUVZvYyHULRaS+rQvD7HmfAWJE5G1V3ZHNZ7wCKI5xrRuiqleIyJuY5DtvYeLjP6iqf9h99ncwrnj+ciOwW1VvBhCRcjbT39vArap6QETuBF7BhPgFKKppuJN4u+xVrlyZmTeWSq3ZBcETaGfv3r2cOHGCuLi4pLJz584xY8YMJk2alFTmYeLEibRq1SpZeZkyZRg92iRVnDt3LsWLF09xX3bxti9QCXQbnX3ZI9Dtg8C3MbfsK4iTfkZ4x9DvCDQVEc8SfTnM1/NO4FURaYMJSlMDsz2QHtdhJkRUdZOIbAM8k/4iVT0KICLRQB3Mi0VWWaKqx4HjInIUkxgHTBz+pnalojUw63xUYYql0k9aG81q+xorIqMxqww/ikhjzAvH97bfQsAer/tmpOjJ06Hq+5gXEUJCQjSQIsr169ePefPmUb58eUqVKkXp0qV59913iYmJ4dixY5w8eZJXXnmFqKiopHu2bt3Kl19+yZAhQ5Ki4y1YsIABAwYgItxzzz0sWrSImTNnUr9+/TRGzhoXQ0S+QLfR2Zc9At0+CHwbc8u+gjjpbyD9fXbvyHiC8S3/zruBXUKvjMm6l2Bj8xfPhk2ZdY/LTH/esfc9cfeDgCOeFYl0iMVk5/OmjL03VkSaA52Bl0VkEUYrsUFVr0mjv4sy9n6fPn3YvXt3kgDvX//6F6NGjWLGjBn06dOHnTt3pgimc++991KhQgX+8Y9/AGZFYMCAATRs2JA///yT0aNH8/LLL+f4hO9wOBzpURBd9soA14vIME+BiDQVkdRi6n8HPCQiRUTkZxGpb/eyywH77YTfDvNlDnCc82p2D0VEZD1G6X6PHa8+UNseLwJ3ikiciMRgYvsPFZEHRaR3jj31ebpjIuSVFZFdIjJBRMpbMaMvS4FbRKSMtfs2YI2qnrOeCScxwsXXMdsiMUBlEbnGti8iIo1y4RnylDZt2vDee+8REhJCQkICs2bN4r777gNgypQpxMTEJNvTnz17NldddRUPP/xwUtlvv/3GZZddxty5c9m4cSPDhw8nISFfpiRwOBwBTEGc9HtiQsTeKSJ/icgGYCSw19Z7J52ZDGzEpIgtC0zCfClPB1ra2PK9Mf7rqGosZt9/vYi87jPuO0CQvWcGRuA2Dxhmr1diXgp+xGSpe09VUwZnzxnuwSzDrwfuwLj+3erbSFXXAhOAn6w48UFM4iAwAXd+wwgQhwMvq8kr0AMYbWP1R2G2EfItvsr9uLg4Ro8ezfDhw5O127VrV1K4XYCaNWuya9cuHA6HIy8pUMv7di/7OoyCf66q1rPl4cCHGNX+5b4qflWtb+PSNxaRzzHZ9zxfsx8Bz9vzHzGq+XiMCn+riHTC7HmfBvr62qSqH2F81yPtdRfb1whMWNwxtm41ZhWgFOZF4xYR+QOYoarPAY1FpBfmy7uoiEwC+qtqsPdYdmsCq1voJCZz3p8YbwRsH49gFP+/2v4EqKeqT9o2fTAug0098fpt+RBMuN5E4FtVfVpE6onIAswKyJsikqGXQCCp9zOr3B8xYgSPPfZYQMf0djgcBZcCNeljvmYXqOpmEYkVkRbWdxySq/bDva99+piB+Tr+RkSKAu2BhzATYwcbwOZyjFta+okPMke8qrYUkUGYCboFcAj4y6ryqwB3AtfabYd3MF/06a4W2KX6NRgXxfg0+vgS+AWTdAfb5hXvfkTkJszv20pVT4pJ1Qt+egkEqnrfV7kfGRmZpKpNTbm/cOFCPvnkEx555BHi4uIICgpix44d1K9fnzVr1iS1W7p0abL+c5JAVyVD4Nvo7MsegW4fBL6NuWZfRsH589MBzMNMzGC+Zsfo+QQzS7zaJbvW5MloimOSyRTDTHLTbXk5YBpG1R4FnLTlwcB6P2yLxHw9e653YpbdozBR8f7ALKnfQPJkOEsxLn8DMRnxPAlvYoARGY1jyyIwE3mafWDSA18N/APYwvkMjZ7fZSzG19+739KYYD5RXsfGjH6L7Cbc2bRpkzZr1izpKFOmjL755puqqjp+/HgNCQnRhg0b6pNPPpl0z5o1a/Tqq6/Whg0bauPGjfXUqVPJ+tyyZYs2atRIVc8nwvj222+1TZs2adoxfPhwff3111VVNSEhQS+99FL9+++/9cyZM9q0aVNdv359tp4zLQI9kYhq4Nvo7MsegW6fauDbmFsJdwrMl7798rwBExFOMe5kKiKer1dfZXmqSnM1X/KRQCfMRPm5rXoME863GUYrkd0ws5NJvrw/WFXX2VUIX3V+YcxKw8eqOjRFT+lgl/ebYLQLVdLp43PMCscmzNaFP3Fj/fUSyFFCQkKS3OfOnTtHjRo16N69O0uWLCEiIoI1a9ZQrFgx9u/fD8DZs2fp1asX06ZNo1mzZsTGxlKkSJGk/nr27ElkZCQHDx6kZs2a9OzZk/DwcD7//PNUg/KkRuHChZkwYQKdOnXi3Llz9OvXj0aNLnqNo8PhuMgoMJM+RmA2TVWTQsWKyA+YfXLPdTeM29m9GfQ1AyNoawn0EZGfMfvfO1U1UUTuxbxUpIt3JL1UqusBN9g99ibAFyLiyRNQTUR6a3Kh3yIgQkTeVNX99iWnjKpuS2f8Iphl+h2qulZEJgOXpNHH18CzmKA/Q7y68azDfw8ME5Hpapf31UT+2yIi/1LVWWKc95uqjYCYFyxatIh69epRp04dnnzySZ5++mmKFTMhCapUqQKYJfmmTZvSrJlxYPC42Xn47LPPkl17ltw++uijdMceMWJEsuvOnTvTuXPnLD6Jw+FwZJ+CpN7viZm4vPmS5JniegI/kUZkOjkfM34h0BYTVCdeVVtj9qrv9eyPk32f9L+AN+1X8nHgeXs+EtjuM+GjqtHAc5hIf2sxk7Cvj72H6bbNesyk7VHun8SECk7Rh6oexqwG1FHV33w7VNUFwBxgpVX6D7ZV9wD32d9lA6l4CeQm3l/jmzdv5scff6RVq1a0bduWFStWJJWLCJ06daJ58+a89tpreWmiw+Fw5BkF5ktfVdulUjbecy4iKzF72B5lf4gtD8co+RcD0VZw9hLm5aC1bROnqqVtNL1pqvoNMMQq++cBXbyU/QADVTVZghxVDfe5HuF1GYWZcFHVSBEJF5HBqjrG1t8jIu9xXtk/FLM60AVYbm30VuVHYpT951L5qb5X1ZEiEgeMAyaJyClMaN0uInKpiPyC2a+PIPnLzTl7XQzwOKGHYfIThAGXAD+IyPuqupc0yI5631ttHx8fz5w5cxg5ciRglvEPHTrE8uXLWbFiBXfccQd///03Z8+e5aeffmLFihWULFmS9u3b06JFC9q3b58lGxwOhyNQKTCTvh84ZX9ySgHLVfVZEXkN+A/wMuZF4F1VnSoiAzyNRaQjJkTxVfZ554hIG1X9WkRux+QeuBEYntqEn1PqfW+1608//cSll17Kxo0b2bhxIyVLlqRu3br88MMPgHkpiIiI4NixY9SvX5/169cDEBoayqxZsyhUKPUdmgKr+s1BAt1GZ1/2CHT7IPBtdOr9i1/Z/38YJfspjPguChPxLyvK/hEYYZ+n7lp7nmPKfoxY0KPQvxOYbM9jgSL2vKzXs48BtnqN8Sdwn62rgPFE+NKff4ucUu+XL19ea9WqlaTef/fdd/X555/XMWPGKKDVq1fXxMREHTFihJYoUUKbNGmiDRs2VEA/++yzNPvPj6rfvCbQbXT2ZY9At0818G106v1cJI+U/VGYBD4eZX8TzFJ7XbvfPRWzh5+YhUeoKyaj3VdkT9nfAPM7rMSkH/Yo9H3zAXgr90uKyTQomEiCk1Lpt6a1paqIBGXxGf0mJCSEZcuWUbt2baKiomjYsCHdu3enWrVq3HXXXSxcuJAiRYowYcIERIThw4dTr149Ro4cSVxcHLVq1eKuu+7KTRMdDofjglCQhHzp4VH211HVYFWthfFFTy0ef0bMwETeux5YYMvKAXvsZPdvzEvFKUyymj+BDsBNmHC2WWGjqj6STv0ioIeIVAHzkiMiddJpnx7LAM+MeI9X+XdAPxv1EBGpISJVrPjxQ4xIcunocZQAACAASURBVCPweBbHzRSlSpUiNjaWlStXJqn3ixYtSlBQEMuWLaN69epcf/35f95evXqxYcMGrr32Wp5//vm8MNHhcDjyHDfpG/xR9vuLR9n/f2pi0UMGyn5V3Y/Zzx4ohkIi8rqIrLAq+moAIvK5iCQp1axQsDLQQkTm2eJCIjJFTIz/lsANapT9M4EtVpT3F1A3g+coKSIviMgqTOCdcrZ8ODBBRE5bmz3PsBDYDOy3Y6yy90zABChaifEueEVEumYwdo7hrd6PiIigRo0aSa55vpw8eZIFCxZw++2355V5DofDkae45X0yVvZj9ro95ZHe17astNd5AlDRp/4PoKlX0RCr+N+KSXyDqv5tA+VUwegBjqrqlSJSDPN1HYsVCqrqvVYo+BdQH2gFdFWj7L8JKKaqTQBEpIJdfr8eqKKqJ8TEyL8eWOLz2AuAv+35duCgqjYXkf4Y8SKYVYw3VPVF+wIyz44TCpQHyul5seA1qvqgiBzEiABLYLQEc31/b29yQ71/8uRJXn31VRYuXJjmvXPnzuXaa6+lYsWKabZxOByOixk36QcmHYGmItLDXpfDKOO/BcbZF4EbgaWqesrEvEnin5xffkdVD4tIF6AhJgMgGC3BL37Y8ZX9+ztwmz1v4zlX1W9E5LAtb4/xGlhhxygB7Ld1L2LS+Z7GiCRTkNvq/b///pvNmzcTEhICwIEDB2jUqBHvvvtu0iQ/YcIE2rZtm6FitsCqfnOQQLfR2Zc9At0+CHwbnXo/nx1Y1bs97wREA2cxgr8jwLI07psK3AJ8Ctxiy8Ixkf3ATNCX+9zTFfjMD5s+AnrY861AJXveEoi051FAXa97DgGVgIcxQr7U+q2GWZWIBkplZEd21fse7rzzTv3www9TratTp44eOHAg6frIkSNaoUIFjYuLy7Df/Kj6zWsC3UZnX/YIdPtUA9/G3FLvuz39wGAVJsHOK2qi7j0FHLBhchGR+iLi+fRNTSjozfcYn3jsvRUwAXquFZHLbFkpEamfRVuXAnfbfm7CuONB+mLBSZj0w9OB0Vkc129iYmJo0qQJX3zxBW+88QZly5blrbfeSqofO3Ys27ZtIzY2FoDp06fTuHFjEhMT6dChA2vW5FmUYIfD4chT3PL+haOEDVdbBPOFPw14w9ZNxojfVtl49Q0w2e3ACAWnARF6XijozcuYOP3HMB4IL6jqVyLSB/jMbg2AEdVtzsDGp0RkP8k1DN0wQX96Aj9j9v5R1WgR8YQBDsJE5BsgIm0x7n+fWs3CzyJyg6ouzmDsLBMSEsK6deuA5Al3AHbs2MHChQupXbt2Uoz9Sy+9lLVr11KhQgW+/fZbHnjgAX799dfcMs/hcDguGG7Sv0CoapoJedS49j1jD0+Y36O2LjWhYCR2YlbVOBEZhQne08WrzWLgygxs6uN1HiwiI+z5SswWAhh/+3+p6sFU7p+BWYnwZjk28p+asL+tvHIY5DreCXcAHnvsMV577TVuvfV8CoDWrVsnnV999dXs3Lkzr8xzOByOPMVN+hcRNgTwCOAgRvX/O9BLVVVEbgTewiTN+cnrnlLA27Z9EYx6PkJExgGxalT4nTAZ9MLVz8A5tt+ZmMA7hYCXVHWGiLTArFiUtnb2UdU9NmhRFHAdJgzx2LT6zin1PmTOZQ/ggw8+4KabbsrS2A6HwxHoeMKsOgIYr4Q+4ZjY+o0wYXWXAU9ifOD/wETT+xPztV1STYKcV4FoVf1ERMoDv2HS4ypGUT8QeA/orKp/+Yw7AiM4HONVthUj7GsL3Kiq/7Hl5TAvHD9gkvMcsFECO6lqPzvpR6tq/zSe0Vu932LmzJnZ+MUMCQkJ9OjRgylTplCyZEkee+wxXn/9dUqXLs1dd93FpEmTKFeuXFL71atX89ZbbzF+/Phk5b7ExcVRunTpNOsvNIFuHwS+jc6+7BHo9kHg25gV+9q1a/e7qqaf1yUjpZ87LvzB+fj24SSPrf8u0AsTX3+pV/ktnFfzr8Sk0PXExN8OhNq61hg9wcNpjDscG+Pfq2wrRl9Q356PBq63dY2BY15jrQMW2rpIoK0/z5tT6v3Zs2drhw4dVFV17dq1WrlyZa1Tp47WqVNHCxUqpLVq1dI9e/aoquqaNWu0bt26GhMTk2G/+VH1m9cEuo3OvuwR6PapBr6NLva+w4N3bH3fmPipIcDtqhqTSl0TTNCf6mncG4uNBuhFGeCIqsaKSHNMKOGXRWQRJqrhBlW9Jo3+Us1Z4A87duygd+/e7Nu3DxHhgQceYNCgQTz//PNEREQQFBRElSpV+Oijj6he3TzOuHHj+PPPP2nUqBGVKlVi//79Sf0FBwezcuVKKlWqxPbt27ntttuYNm0a9etn1anB4XA4Ah/nspfLiMglNnzuXyLyu4jMz4a7nKfP8jZKnoejQEsRqWevvcMHfwc8bL0AEJEr7N86wBOYpf6bRKSVV/+/Ws+CZzHRA9eKSJSIDAXOqeo5EakOnFTVT4DXMRH7YoDKInKN7aeIiDTKzrN6KFy4MGPHjiU6Oprly5czceJEoqOjefLJJ1m7di1RUVF06dKFF198EYBdu3axdOlS5s+fz4YNG5g1a1aafb/44ovExsbSv39/wsLCaNkyJ7MeOxwOR+DgvvRzETvRfo3JcHeXLWsGVMW6y4lIYVU9m8muywP9Oa/Y3yIitwHfiMhJ4EfMFznASxiB31rrSrfFxr7/ALN0v1tE7gM+EpErVfW0qraytvUB7uW8t0A74Gp73gR4XUQSMe55D6lqvI0iON7u8Re2Y2/I5POloFq1alSrZhYdypQpQ2hoKLt27aJhw4ZJbU6cOIEnOmFERARPP/00jRs3BqBKlSrJ+tu6dWvS+eTJk5k8eXJ2TXQ4HI6Ax036uUs7jI/6e54CVV0jIuEi8iNwGGhg49aPwuzZFwMmquokMRnrIoDNNoHOc2rEeZ8D9TD78adEJBgYo6oNRKQ4Zq+/joisBh5X1f/aCfwWoCTmheNrVZ1jbfodM4mnxgZVHQhgx5mH2buvhonTXwoTIvhqEWmDySJ4BminqodEpJ6ILMC8hLwpIv9R1U3p/Wipqfe9Vflbt25l9erVtGplFieeffZZpk6dSrly5ViyxKQT2Lx5MwkJCYSHh3P8+HEGDRpE79690xvW4XA48j1OvZ+LiMgjwKWq+phPeTjwDdDYfqU/gEmG87JXgp1/ATswKvxjNmnOcswEWwcj1Gts+wv2XIvIE0AjNYr5BphgPvUx8fiHYZbzz2CW4q9T1R3p2N8HaOk76dtx+mAC/FwBFMd4DQxR1fdE5E1gm6q+Zff6H1TVP+wWwkhVvSGVsfxS7586dYpBgwbRq1cv2rRpk6xu+vTpxMfH07dvX8aNG0dMTAxjx44lPj6eAQMGMHLkSGrVqpXW4/pNflT95jWBbqOzL3sEun0Q+DbmlnrffelfOH5T1S32PK0EOzuBV+0XdCJQA7M1kB7XYfzyUdVNIrINM+kDLFIb5EdEojEvD2lO+n6wRFWPA8dF5CjgyZ63zj5PaYyHwCw5nxSoWMpuQFXfB94HCAkJ0fDw8BRtEhIS6NKlCw8++CCPP/54ivq6devSuXNnPv74Y5YvX07Tpk2TfO7nzJlD8eLFSa3fzBIZGZkj/eQWgW4fBL6Nzr7sEej2QeDbmFv2uUk/d9kA9EijzlvJLhi3ue+8G9iv6cpACzXpardivqqzSmaV/5npL9HrOtH2HYRR+odlcxy2b9/OVVddRXx8PDt37qRQoUIMGjSI8ePHM2nSJDZu3MgTTzxBgwYNABNa9/7772f27NkkJiZy9OhRHnvssQxGcTgcjvyNU+/nLouBYnbpGgARaYpJluPNd8BDqSTYKQfstxN+O8yXOcBxzgv1fPkRuMfTD1Abs5Sf56jqMYxw8F/WHrFCxkyzatUq9u3bR82aNSlUqBBPPfUU7733HvPmzSMhIYGSJUuyfPlyxo0bB0DXrl157rnnSEhIICEhgUOHDiW9EDgcDkdBJVcnfRHpJiJq95Yzc9/PWRwvWETWp1HXybqdRYlInIjE2POpIvKgiOSGyqswJurdWBE5Y5X1k4G9Pu0mY9LOrrL2T7L3Tse44q0DegObAFQ1FjgoIhtE5HXbRx0bce8dIMjeMwMTBvcMmUBEJlqXvVeABz2/G3CT1zj+cg9wn4iswax83JpB+1Tp1q0bqsratWtZu3YtN954I/Xq1WPhwoVs3ryZli1b8uabb1KjRg0ASpYsyZAhQ4iOjubbb7+lVKlSGYzgcDgc+Z/cXt7viYkD3xMT3S0Zvu5qnmtVbe3bNrvYpfPv7DiRGHe1lTk9jg8vYb7IK6nqGRGpiolKNxP4n1c7UdWkBDs+pBXo5h8YkZ0n8Y33l39f38aq+hHwkdd1F982XnUDIJlwz3t5/l3717e/4NTGsrqFG9MaKzV81fu+8fR91ftp8euvv9KvXz+2bdvGtGnTKFzY7WY5HI6CTa6p962IKwbjtjZXVUNseThmMjyMSRn7gPe1qtaX87HmPwemqeo39t6PMC5jKzHpZT2fbwNV9WdvdXkGtkXiNemLV4x5W7caswRfCvOFPRTj0jZDVZ+z9/QCHgGKAr8C/dVkkfOMURIjkrvULnP72hCH+aL/JzAAk0o3RX8i8i4mO14J4AtVHW69AsbY3/egqraz+/0tVfWgiDwO9LNDTbYq+mDgW8xLWGtgFyZG/ql0fqcUv6ecj71fGliA8ShojVnRmAK8AFQB7lHV3ySNhD+pjJUt9f6jjz7KQw89REhISIp7tm3bxqhRoxg3bhxFixZN63H9Jj+qfvOaQLfR2Zc9At0+CHwbL7rY+5hl3Q/s+c8YMRoYX/QTmMkwxbUmjzXfHRPYBsxkuAMz+ZUEitvyy7HxhjET53o/bIvETJCe6xHYGPO2brQ9H4RJbFMNozrfifnCDsUo1YvYdu8AvX3GaAqsTscGBe6w52n2B1S0fwtZ25rq+Rj4lbz62wpUAlpg1POlMBPzBoxbXTDGrz/Mtp+Jidv/K+dj5XuOJmn9nl7jePprgtkm+h34ECNKvBWYbdu/iskECCao0GagVHr/PmnF3o+Pj9eOHTvq2LFjU9S1bdtWV6xYkep9qqrt2rVLtz4z5MeY3XlNoNvo7MsegW6fauDbeDHG3u8JjLPnn9vr3+21t7taatcevgXGWd/1GzFJZU7ZaG8TRCQMo0LP6YDpc+zfdZjgNHsARORvoBbGLa4FsMK6opUA9qfST3qcA7605+3T6e8O+xVcGPPy0RBYm06/12EC75ywNn+FWbWYA2xR1Sjb7ncgWG30vSyyRVXX2XE2YFwC1eoJgm2bjsAtIjLYXhfHiAs3+jtIv379mDt3LgkJCfTp04fHH3+cESNG8L///Y/KlSsDkJhoMgJv3bqV0NBQgoODKVq0KNdccw1Dhw5l06ZNBAcHpzOKw+Fw5H9yZdIXkYqYNK9NREQxX6kqIk/aJr6JV1JNxKKqp+1yeyfgTszLA8BjwD6gGeYr83SOPkBy1zNft7TCmK/Zj1V1aDp9/AnUFpGymsryPnBaz28HpNqfiFwKDAauVNXDdnsjJ132SvhxTxERWa/Jt0yewoTezchlD9JP+OMXffr04dprr+X+++9n8eLFhIWFsXfvXm6++Wa6dOnCww8/zIEDB7j55psJCQmhXr16DBkyhFGjRrF8+XK6d+/OO++8Q6VKlbJqgsPhcOQLcku93wOzF19HVYNVtRawhZSuav4wAyNMux6zhwzGlW2PqiZiwr4WygGbM8MioIeIVAHzkiMmgU0SqnoSE99+nIgUte0qe9zX/OyvLOaF6KgVAd7kdU9abns/At1EpKTdT+9uyy4UKRL+iEimXjbbtGlD+/btadSoUVJynQcffJDQ0FC6d+/Ozp07OXPmDPv27WPq1KkA/Pvf/2bDhg1ERUWxatUqunXrlvNP5nA4HBcZuTXp98QkmvHmS5Jnf/OXhUBb4P9UNd6WvQPca93AGpCNlK1ZQVWjMSFoF4rIWuB7UqagxbY5AERbV7x5mHzzfvWnqmswosJNwKeY8Lwe3gcWiMgSn75WYZTzv2H26yer6uqsP226VBSRaGtzWwD7ovEaUFdM7P/VGAHfdhE5Zp9tUXqdetT7vvH3fZkwYQJNmzalX79+HD58OKl8y5YtXHHFFbRt25Yff7yQ7zsOh8MRWLjY+450SUPBPwKIAx7HCDDPiEh5VT0iIq8C0ar6ifXn/w0jJPwX8DJGiHgolXHSVe/v3buXoUOHMmXKFAAOHTpEuXLlEBE+/PBDYmNjGTJkCPHx8Zw6dYpy5coRExPD888/z5QpU3LUTz8/qn7zmkC30dmXPQLdPgh8Gy869b478seBiQLoq+AfATyB2W75AuMFUNrWrQTWc94TYDvGO6EPMMWfMVNT72/ZskUbNWqUojyjuoxU/VkhP6p+85pAt9HZlz0C3T7VwLfxYlTvXzBEpBMw2qd4i6p2vxD2BDIiMhG41qd4nKpOseexQAWf+ooYjcbNQBugK/CsiDQhDeGemAx7md6GefPNN5k8eTIJCQns37+f06dPM2DAAH755RcKFy5M/fr1admyJY0bm4WIAwcOULFiRQoVKsTff//NH3/8Qd26dTM7rMPhcORL8uWkr17R9xzpozb6Xjr1cSKyR0RuUNXF1jPjRow7Zi1VXSIiP2FS95bmvHDvYVVVEblCs6gp2LVrF+PHj+fKK6/kxx9/5NixY1SvXp0XX3yR+Ph4oqKi+OGHH4iOjmbRIiMTWLp0KcOGDaNIkSIEBQXx3nvvUbFixawM73A4HPmOfDnp+yIilwBvYSLbHcG4+z2qqpuz2W954G5VfcdeVwfGq2pamfUy2/+vmKBAFTHudbtsVX/g8Zwax2u8PpigRQN9qnoDE0XkDXv9AmbZfomINMJ4EoxXs6f/Eua3XisiQZgVgTRD/mbE2bNnee+99yhbtizdunXjkUceoWPHjgwcOBBVpX///gQHB1OtmtFR3n777dx+++1ZHc7hcDjyNfl+0reuYl9j/ODvsmXNMHnpN3u1S5YHwE/KYybgdwBUdTdpp9LNNGoD56QxGWcpKVEW7YjGhFP25TpPaGRV/cK2PQX8N5U+PsIrVn96eNT7W0fdzODBg6lduzYlSpSgY8eOdOzYEYC+ffsyf/58GjZsyNixY7P0XA6Hw1HQyPfqfRG5ARPvvU0qdeEkzwMQCozChAYuBkxU1Uk2j0AEZm+7CPCcqkbY3AC3YmLgfw9MxCrdRaQ4JjlNS0y42sftUngf4BZMKOF6mOh5T2XwDH3wmvS9FfW2rhsm7O7lmJj8RTHxC84AnVX1kIjUs/ZVBk4C/1HVTWmNYyfzY9b+S4CnVPUL+xL1NtABExY5HvjQ1rUA3sAs8x/EiPdOYhT8t6hqjIh8BixW1f/5jJ1CvX/8+HGGDx/OsGHDKF26NCNGjKBt27Z06NABgHPnzjF+/HgaNGjATTd5hzDIXfKj6jevCXQbnX3ZI9Dtg8C30an3s64+fwR4M426cJLnAXgAM6GDmfRXApdiVkTK2vJKmGh7gk9seu9rjLr9Q3veALMcXhwzEf6NCTBUHNiG2RtP7xn6ABPSGKePtacMZkI/Cjxo697EbGOA8Y2/3J63wky8aY6D+SqfhYnl0BD405bfhnnBKQRUx2yX9MC8DP0MVLbt7vR6/g7AL5h9/wUZ/Zt51PszZ87Ufv36qYePP/5YH3roIfXmhx9+0Jtvvlnzkvyo+s1rAt1GZ1/2CHT7VAPfRqfezz284/53BJqKiGeJvhzm63kn8KqItMGEmK2B2R5Ij+swX8So6iYR2cb5HAGLVPUogIhEY9zidmTjGZao6nHguIgcxSTvAZM7oKldqWgNzLKB8cC81GTEbDVRD6NtREAwav3P1IQQ3i0ii215CCaT3vd2jELAHgBV/d5GIpyICZ3sF7Vr1+abb74hNDSUoKAgTp48ycMPP8wtt9xCTEwMhQsXRkSSlvwdDofDkT4FYdLfQPr77N5uZAI8rEb9f77QLHtXxmQKTLDpZXMyBn52/x0yioEfBBxR1TDvm9JIRXyVV3Ic736F9BFMcqJrUlQYQV8oZqm/AuYlKkNq1qzJ6dOnSUxMpFChQiQmJlKmTBn+/PNPihYtiqoSFxdHzZo1/enO4XA4Cjy5FYY3kFgMFLN7xgCISFMRSS0PwHfAQyJSxLarb8PKlgP22wm/HebLHNKOfw8m3v09nn4wmeWynHQmO6hJ+LPFE/dfDH5/cfuwFLhTRAqJSDXOC/xigMoico2IFBaRIlbZDyZB0kbgbmCK5/f1hzJlyvDLL78QFRVFkyZNqFOnDtHR0axbt47169czaNAgYmNjs/goDofDUbDI95O+3efoDvxTRP6yKWBHAntTaT4ZiAZW2Vj5kzBfytOBljZlbG9MLHxUNRZYJiLrReR1n77eAYLsPTOAPqp6hgvHPcB9Nl/BBowAMSNu8oqt79kO+A6TXjgOM9H/ZcvvxnzBf4vZ548CWotICCZT4NWY7Y7L8fKaSA2Per9GjRpJ6v1q1apRrly5ZEv5CQkJTJs2jRtvvNGf53c4HI4CT75X7zvSJr24+qo6RkR2k0Ox9W3fRTArL6+p6lyfukyr98eMGUPx4sUZONA3rEDukh9Vv3lNoNvo7MsegW4fBL6NuaXeLwh7+o60SeuNz1O+FpguIrOB2basI3CL175/cczWBcD3aU34lnEYr4G5vhWq+j4mcyAhISEaHh7OrFmzuOKKK5LS4u7evZvly5cTHh7OCy+8QOHChZk5cyZBQXm7YBUZGUl4eHiejpkZAt0+CHwbnX3ZI9Dtg8C3Mbfsc5N+gOAVfc+bf6vqulwcNr24+pCDsfWtGLIO4Pdnuaoyffp0VqxYgYgQHR1N165defDBB5k6dSqnTp1i1apVtGyZvluqw+FwOAz5fk//YkFVW6lqmM+RaxO+XdpfDuyxAYwQkdFAT+Anq7ivpapLgCEYMaN3bH2x91zhx1gtMPv6vawLoF/ccccdSelyz507R+HChRk1ahSTJ0+mUqVKlCpVirvvvpsXX3wxM4/ucDgcBRY36Tt6A8+LSBRwL7BQVf/C+Nl/YoWIq7Gx9TERDItgYutvsNdJiEhqq0cDMSsIS0QkSkQm+2vcCy+8wKZNm3jjjTdo3rw59evX5+zZs2zfvp2WLVvy6aefMmzYsKw8t8PhcBQ43PJ+AUe94up7RHy26iHMRH0W438/yrovvo0JwnMWE944wi7d34ZZCVgEtPX0LyJXAi2AupgXid8wCXnS5FTCuRRln3/+OT179szqYzocDocDp94vsOSlcl9EXsYI/koAO1V1ZCptktT7lSpVbjFr1sykuoSEBHr06MGUKVOSpcl99NFHeeihhwgJCcnuz5Ep8qPqN68JdBudfdkj0O2DwLfRxd53R44eGFHdep+yEcAT9nwB8AXQCyhty1YC6zE++FGYfAKhmJj9U9IZqyiwBvgVKJSRbZ7Y+x5mz56tHTp0UF/atm2rK1asSFGe2+THmN15TaDb6OzLHoFun2rg25hbsffdnn7BJS3l/kF7fjMmVn5zYIXdq/co9z1Cw9qqutG2T1O5D/wDs/RfhkyGLz5y5AgDBw5k3bp1hIaG8ssvvzBixAhq1KjBypUrufvuu5k/f35munQ4HI4Ci5v0CyiqGkdy5X5F4EZyWLlvmQQ8j4lsODozdvbv35/Dhw+zadMm1qxZQ2hoKBs3buTEiRMkJCRw9OhRxo0bl5kuHQ6Ho8DihHwFm97ARBF5w16/oKp/2ch5n4hIOczX/Xg1e/ovYUR4a+2LwRagS3oDiEhvIEFVPxWRQsDPInKDqi5O7z6Ao0eP8ssvv3D8+HE82QGLFi1KaGgoV155JYMHD86gB4fD4XB449ekLyL1MAKsMyISDjQFpqpx4XJcpKiXct+nPAGTGti3/BTwX+8yq85/HJOdrxRG3Henqq6390wVkZLWJRDMqsHzmHC8qeJR72/ZsoXKlSvTt29f1qxZQ4sWLZK+6idMmMDUqVNp2bIlY8eOpUIF350Kh8PhcPjil3rf/ge7JRAMzAcigEaq2jlXrXNcFPijzrft/Iq971Hvx8TE0L9/f95++20aNmzI22+/TalSpejWrRvlypVDRPjwww+JjY1lyJAhufuQXuRH1W9eE+g2OvuyR6DbB4Fv4wVV7wOr7N8nMfnmAVb7c6878v+Bn+p8TObBFzLqz6Pe37Nnj9apU0c9LF26VDt37qzebNmyRRs1aqR5SX5U/eY1gW6jsy97BLp9qoFvY26p9/3d008QkZ6YiG1dbZnfOdEd+R6POr8I5os/hZI/K7H3L7nkEqpVq0anTp3Ytm0bBw8epFOnTjz66KMsXryYoKAgTp06RYMGDXLoMRwOhyN/4696vy9wDfCKqm4RkUuBablnliOnEZG4jFslax8uIvP8bJ6uOj+rsfcBKlWqRExMDEWLFqV169a88sor7N27F1UlMTGRokWLUrZs2cx06XA4HAUWv770VTVaRIZgU6iq6hYy6XrlyJ/4qc73jr0PZgnq/oz6Pnr0KOvXr2fLli1J6n0wIXk9jBw5ku3bt+fQ0zgcDkf+xq8vfRHpionAtsBeh4nInNw0zJE72C/4SBH5QkQ2ich0L7/7G23ZKkwsfc89pUTkQxH5TURWi8ittnwcEKyqt4tIJ2AJcI33hG/bbFHV6hif/2NYwV5apKbev+KKK7j//vs5ccLsHDz77LPUqlWL6dOnuyx7DofD4Sf+qvd/B24AIlX1Clu2Xr3itjsCGxGJU9XS1uUyAmgE7AaWYQSaK4E/MP/OfwIzgJKq2iWdmPsKrMB8yb8HdFaToc973JIZtbHt/Fbv9+vXL+m+6dOnEx8fT9++fXPqp8qQ/Kj6zWsC3UZnX/YIdPsg8G280Or9hQSAdAAAIABJREFU5fbvaq+ytf7c647AODCJdADCge+9yt/FxNcPA5Z6ld+CScgDacTct3WtMRn3Hk5n7AzbeB+ZUe9v27bNqfd9CHT7VAPfRmdf9gh0+1QD38YLrd7fICJ3A4VE5HLgEeBnP+91BB5nvM7PkbG2wxNzPyaVuiaYOP7V07nfnzYpSEu9v2LFCp555hm2bt1KUFBQnmfZczgcjosVf9X7D2OWg88AnwJHgUdzyyjHBWETEGyjLwJ4J69PNea+iNQBnsAs9d8kIq18O/WnTXqkpt6/5557WLduHSVKlCAoKIiaNWtm9lkdDoejQJLhl75VZH+jqu2AZ3PfpKwhIpdg4sJfCRwB9gGPqurmbPZbHrhbVd+x19Uxseh7ZNNkT/+/AsUw6vYSwC5b1R94PKfG8eJGoIOIrMb41hcDjqjqabuv/o2InAR+xGTFA0gRc9+KOz8AVgENgfuAj0TkSlU9bZ9NbJvBqrpbRFK0SY+01PsiwurVq6lWrRp79uwhPDw8B34Wh8PhyP9kOOmr6jkRSRSRcqp6NC+Myix2cvka+FhV77JlzYCqwGavdoVV9Wwmuy+PmYDfAVDV3UCOTcSq2sra1gdoqarewWtybAtFVT2KkE3A/zzjiEg74DMRma6qC4AUkW40lZj7ln/6XDfxuU+926jq775tUiOj2Pv79u2jWrVqgNkC2LdvX0ZdOhwOhwP/1fsRmOXZ7/GKtqaqj+Seaf5j08OOUNU2qdSFY75UD2MmtFBgFEbQVgyYqKqTRKQ0RtVeARNZ7jlVjRCRz4FbgRjM80/ECNwai0hxjBCuJUao9riqLrET+C1ASaAe8LWqPpXBM/TBa9IXkWCvcfoA3YBSwOXAGEzo239jtlw6q+ohuzQ/EagMnAT+o6qb0hvHlr0IlFHVx1LrA9gDrAUuVdVEm1hnE1AX+J+18wubfGectfMM/H975x7v1ZT//+erji66kYrI6EIJkULMuBQKiTJjRv0YCt8ZMxOGCWOYGRkzDBky+BoTNQrlnhiDqc6XEJKTQhGVyiUiKd3U+/fHWp/69Olzzvmcy+ec3Tnv5+OxH2fvtdZe+7VXnbP2Wuu93m+Oi3Vs095Z3j9n6/3HHnuMp57a4jfolFNOYdKkSZlV5o2aaPVb1SRdo+urGEnXB8nXWN3W++dkO3K5tyoOgmHhLcXk9SR8qLSL1z8jdOgQOqEZQDvCrEfTmN6CsG1NhCBDc9Lq23xNWKu+N57vS7BqbwAMBj4kRJRrACwixKcv6R0GA7cX85zBUU8TQmf8NXBBzLuFsIwBMBnYJ573AKaU9pyYNgB4pqQ6CB9EveL5GcCoeD6GMPNRL77zoTG9aWzTrO1dUluUZr3fsWNH+/jjj83M7OOPP7ZU+aqiJlr9VjVJ1+j6KkbS9ZklX2O1Wu+b2b9yKZdgXrPgRRCgD3CgpNQUfTPC6HkJ8BdJRwObgD0IywMlcSTwdwAzmytpEdAx5k22uBwi6R2C3/nFFXiHqWb2DfCNpK+B1NB2dnyfxoStcQ+nrX/Xz7HulIFeSXVMIHT2U4GBxOWONDoBn5jZ6wBmtjLWWVx7L6AUdtttN9q0aUPnzp3p0KEDhxxyCI0aNWLFihV069aNDh06cMQRR9C/f/8cX9NxHKd2k1OnL2kBwRHLVphZ+0pXVD7epuR19vQAMCLsF382vUCc9m4JdDezDZIWEkbp5aWs2+LKUt+mtOtNse46BIO8rsVVEP3vZwt4czDwbmYdcWlkWCzzJOGjqDnQnRAiNxeytneu9OjRgzFjxrB06VIKCgqYPXs2TzzxBFdffTWzZ8/mww8/ZM6cOeWp2nEcp9aR65a9QwhW8YcCRwG3AePyJaocTAHqx3VhACQdKOmoLGWfBX4RY7sjqWNco24GLIsdfi/CyBzgG7ZYsWfyInBmqh5CbIJse9nzThxZL5D046hH0ZixRCQdQ5iC/2dmHZEmsf5VBM96Iwlr+BszqpoHtI7r+khqIqmA4tu7VJYsWUJRUREPP/wwRx99NE888QQFBQXUqVOHyZMnc8kllzB48GCaN2+eS3WO4zi1nlyn95dnJN0aXfP+ofIllR0zM0mnEXRdAawFFhJ8CeyRUXwUYb18ZrT6/5ywpn0/MEnSbMK689xY93JJL0maAzxDMHJLcSfwv/Ge74DBZrYufXtZFXNm1HM1wRhxPCHOfSZnxZkNI4zEXyK+LyFi3mhJ/wLWAF9A8L9PCKHbC5gvqb+ZTQQOAxpYMOS7BZgcZ4bWECz3RwGXAytju+zAFpe8WUlZ7//617/mxhtv5JtvvtmcN2rUKPr27UvDhg1p2rQp06dPL2MTOY7j1F5ytd7vlnZZhzDy/4WZlTqSdJJDdfnfT3t+d2A0cJRlbP/MtN6/9NJLmD59OpdccglFRUVMmDCB66+/nj/84Q8MHDiQ/fbbj/Hjx7N48WIuu+yySm2n0qiJVr9VTdI1ur6KkXR9kHyN1W29PzXteB64G+iUy71+JOegev3vtwDeAQ4oTWfHjh3tt7/9re2xxx6211572a677moNGza0vn37Wvv27S3FokWLrHPnzlbV1ESr36om6RpdX8VIuj6z5Gusbt/755nZh+kJktrleK8TSfO+l85PzWx2NcipMv/70avjeOBaM8vJ6u7666/n+uuvZ+PGjXTq1AkzY+LEiTRu3Jh27drRqFEjdtppJ/e77ziOUwZyNeR7JMc0pwTMrIeZdc04qqPDL468+N8nOOd5y8zGl1XQyJEj2WuvYFNZUFDAsGHDaNy4MXXr1uW9996jW7dupdTgOI7jpCix05e0r6QfAc0k/TDtGEzFtrM5CcSCP/yU//2ZwLK07D8RjPDekvQ28KdM3/oE//ujoqfCdIYBfSQVxePUXPQsWbKEp59+mquuuopDDz0UgOuuu47Zs2cza9YsrrzyStauLdWFv+M4jhMpbUq3E9CP4H/+lLT0bwjuWZ3tCIv+982sEChMSx+adl4m//uSXia448XM3pD0ZCw3Mq3YL4AL4nkz4BLCvv8SyWa9n2LDhg2MHTuWkSNHZrnTcRzHyUau1vtHmNkrVaDH2c6IMQIeM7NuMQLf+8Bhtu02T+Je/SnAjWY2KSMvJ+v9FCNGjKBBgwYMHZrN11B+qYlWv1VN0jW6voqRdH2QfI35st7P1ZDvTUm/Imzx2jx1a2bnlkmRU+Mws4WSlsc1/l2BN7N1+JGRBF/+20THMbO7CbtC6NSpk61cuZI33niDwYMHs3btWlauXMmoUaMYN24cw4cPp6CggIceeog6dXI1S6k8CgsLEx3ON+n6IPkaXV/FSLo+SL7GfOnL9S/mWGA34ATg/4A2hCl+x4HggGcwMAS4N1uBaAeyFzA8lwr/+Mc/svvuu9OsWTMaNGhAmzZtGDduHPvssw833ngjH374IW3atGHAgAGV9AqO4zg1n1w7/b3N7PfAagvBd04mRGBzHIDHgRMJbpq38bEfnfIMA84ys025VFi/fn2mTJnCrFmzGDVqFMuWLWP69OksWLCA3XbbjTp16vDtt99Wy0jfcRxneyXX6f0N8ecKSQcAnwKt8iPJ2d4ws/WSphKC9WT65Ifgra85MDXu+JthZueXVKekzetZ3//+92nfvj2S+O677wBYuXIle+21F2PGjKnEN3Ecx6nZ5DpMulvSzsDvCVbX7wA35k2Vs10g6VBJb0lqSPDK96P4UZhe5lpglpntbiF639OEcMClsnHjRrp27UqrVq3o3bs3PXpsmVx64oknOO6442jatGnlvZDjOE4NJyfrfccpDkl3AQOBBcBDZnZ9Rn5bcrDuz7Tef/jhhzbnrVq1it///vdcdNFFtGsXHEFeccUV9O3bl2OOOSZv71YcNdHqt6pJukbXVzGSrg+Sr7G6fe/vSnDC8ky83o/gmrfa/cn7Ub0HUI8Qye9VoG4xZZ4neOw7EXiktDo7duxomQwfPtxuuukmMzP7/PPPrXnz5rZmzZptylUFNdFnd1WTdI2ur2IkXZ9Z8jXmy/d+rtP7YwgGWinf6u8RwtY6zi5AY6AJxXtpLNW6P5PFixfTrVs3DjroIDp37syoUaPYd999GTx48GZf/IcffjhFRUWV8Q6O4zi1glw7/RZm9hCwCcDMviMEaXG2EyStKmP5npKeyqHoPwi2HvcDfy2mTInW/dn48ssv2bhxY/gyrVOHTZs20aJFCwB22203HnjgAYqKiujatWsu1TmO4zjkbr2/WtIuhNjpSDoc+LrkW5yajqSzgQ1m9kCMpPeypGPNbEp6OSvdun8bDjroIGbNmgXAt99+y5FHHkm0/Gf48OGceOKJlfoujuM4tYFcR/qXEqz2O0h6CbgPuDBvqpy8EUfwhZIekTRX0v1pkfNOjGkzgR+m3dNI0r2SXpP0pqT+Mas7YT0f4HhCuN7CtPvqSHpfUivgcOBeSfMltcxFa3HW+1dddRUHHnggl1xyCevWrSulFsdxHCdFidb7kr5nZh/F8wJCAB4B88xsQ7E3OolD0iozayypJzCR4FL5Y+Al4DJgBsGy/lhgPjAB2NHM+kn6C/COmY2TtBPwGsEwz4DXCfvw7wL6mtkHGc+9nWDd/y/C1P7PzexHWfTlZL3ftGlTmjdvzoYNG7j55pvZfffdOeeccyqrmXKiJlr9VjVJ1+j6KkbS9UHyNVaL9T4wM+380dKsAv1I7gGsij97As+npf8vcBbQFXghLf1U4Kl4PgOYAxTF4yOgc8z7PvAdcGExz90z9f8IGA/0K01radb7KaZOnWonn3zyNmXzTU20+q1qkq7R9VWMpOszS77G6rLeV9p5+1LKOtsP6XPiGyndtkPAj8ysazy+Z2bvxrwuwHK27OzYCjNbDHwm6VjgMOCZXAQWZ73/ySefAHDhhRfSu3dvDjjggFJqchzHcVKU1ulbMedOHpC0UVKRpLclzZL0m+jQpjx1HSLptjLcMhdoK6lDvB4U6ykkrNtfqEBbSfNj3l7AbwhT/SdJKi4ewyhgHPCw5WjIV5z1/plnnsnee+/N2LFjMTOuvvrqMryi4zhO7aa0DuUgSSslfQMcGM9XSvpG0sqqEFjLWBNH0vsDvYGTgD+WpyIzm2FmF5Wh/FrCmvrT0ZBvWVr2PcAOwFvAc0CraPx3DzDMzD4GzgNGScq2V/9Jwl7+0amEaO1fLCnr/bfeeovXX3+dVq1aIYnnn3+ePffck3fffZcGDRokek3OcRwnaZTY6ZtZXTNramZNzKwgnqeu3el5HjGzZYROeGgcYdeVdJOk16O/+58DSBov6eTUfZLGSDo9fZ+9pMbAw5JmA7cRO19JfQgW+BdJehiYZmb7mlk3M7vYzPrFateZ2c/NrAvQB/gorh9dDFwtqSjW+UMzWyvprGjpXyTpH4SZgFnADEk3S5oFHFFaG2Sz3r/99ts59dRTad26dYXb2HEcp7bhvvcTRMrCPiNtBWHXRH+glZldJ6k+wer+xwQDvAFmdo6kesAHQEdC6ONhFqzv/wrUN7Nfxzp3BuoCjwEnmdlqSVfEMtdmPL8w1jMjXrclGPgdIOnvwHQzuz8+uy7QlhCM6YdmtkHSK8A+wADgReAMC46eMt+9VOv9wYMHM2rUKG699Vbq1q3LSSedxDPP5GQiUKnURKvfqibpGl1fxUi6Pki+xnxZ7+fqnMepfvoQllhOj9fNCJ3pM8DI+CFwIsECf03KkU3keMK2OQDM7CtJ/QgxFF6KZesBr2R5bravwlTaK8BVktoQguq8L+k4wuzB67HehsDtZjZN0kbg0WwvZ2Z3A3cDdOrUyXr27LlV/syZM1mxYgWff/455513HgDr1q3j/PPPZ/78+dmqzBuFhYVk6ksSSdcHydfo+ipG0vVB8jXmS593+glGUnuCdf0yggX9hWa2jRvbOBo/ATiDsC0up+oJW/cGlVJuObBz2nVz4AsAC574XgVOBv4dlxwE/MvMrsxS19pcDfkWL15M//792bhxI+vXr2f16tXceeedHHfcccyYMYMddtiBunXr8u6775ZemeM4jgPk7pHPqWKi17q7CKNkIzi2+YWkHWJ+R0mNYvEJhGA2RwH/yVLd88Cv0ureGZgO/EDS3jGtkaSOWe4tBM7SlqmDc4Cp8Z72wIdmdhvB4c+BwGTg9OiFD0nNo5V/mSjJen/u3LnMnj0bgFGjRpW1asdxnFqLj/SrAEkDCEFnOpvZ3BKKNoxGcTsQHN40I0zZQ9j21haYGTvgzwnr5BAs6scCEy34uW9LMKx7O+ZfB9whaQ6wI2E0/jXQEpgd+/O1wIOSppvZfWma7gb2BWZJMoKjntQo/ifATyVtAD4F/mJmX0q6GngubjfcQPjgWJRjcwHF+97v27fv5jI33ngjS5YsKUu1juM4tRrv9KuGQcC0+HObLXiSCszsOzOrm36dXsbMNgG/iwcZeRsI0+7prE5Z35vZKsIIPfO5haQZ6WXDzNYT3Oxmy7sBuCFL+gTC7ENmepmsUjZu3Ej37t2ZP38+v/rVrzb73gfYsGEDY8eOZeTIkWWp0nEcp1bj1vt5Jm6Xmwf0AiaZWaeY3hP4E/AVYST9s/RrM+uY5i9/PDDWzJ6O944BniKMuscCqWn+oWb2crqFfSnaCtnaMv8agrveETHvTcKSQSPgbMIIvwswwcyujvecBVxEMAR8Ffhl+rq9pGYEX/2nmtk8SQ8CU8zsnxlacvK9365dOwBGjBhBgwYNGDo06/dIXqmJVr9VTdI1ur6KkXR9kHyN1eJ7349K8Xl/JnBPPH8Z6G5bfOCvBtplu45pKX/5pxGM4yB0rosJVvE7Ag1i+j5Ev8uEZYA5OWgrBA5Ju76G8BGQyvtrPL+YEJynNVAfWALsAnQGJgE7xHJ3AmdneU5vgqX/QOA/pekqzff+NddcY/3797eNGzduU64qqIk+u6uapGt0fRUj6frMkq8xX773fXo//wwCUnPQ4+P1G/H6NTNbkFY28zpFcdvymgG3S+pKsPLPZohXEZ6MP2cDb5vZJwCSPiQE0jmSbbfnLcusxMyel/Rj4A7goFweXJz1/g033MCIESNo2bIlgwYNYuzYsdSrV6/CL+o4jlMb8E4/j0hqTghV2yUawdUFTNJlscjqjFsyr4HgIreYbXmXAJ8ROtI6BGO8XLW1BQ7NSO4JfAOMiNepwDyb2DpIzybC/52StuelP6sOYVbgW8L2v1Kt74qz3v/d735Hy5Ytadq0KZMnT+ZHP/oRkyZNKq06x3EcB9+yl29OJ6zF72Vmbc1sT2ABYZ28rGTbltcM+MSCkd9PCR8VVUm27XnZojFeArwL/D9gdGrbYUkU53u/efPmLF26lKKiIiZNmsT69esr9YUcx3FqMt7p55dBhK166Twa08vKc8AxwH8tWNRDWEM/J/qy35diZgrKyR7AQ5LeAv4Am/fy3wt0I0TN2we4muBT/2tgIfBIeiWS7gAuA35jZi8SliFyGppn+t7v0KEDO+20EwUFYYKqTZs2LF26tBJe1XEcp3bg1vu1lGwW/hnW+x8TjArXSdrJzFZI+gvwjpmNk7QTwSr/YEIMgOuAA83syyzPeczMusVp/veBw8xseUa5Uq33zz33XG644Qbuv/9+AJYtW8YVV1zB6NGjqUpqotVvVZN0ja6vYiRdHyRfo/vedyqb4r72UulvAfdLegJ4Iqb1AU6VNCxeNwC+F8+fz+zwAcxsoaTlkg4GdgXezOzwY7lSfe+vXbuWdevWceSRR1JQUMArr7xCx44dq9x/dm312V2ZJF2j66sYSdcHydfovvedMiPpBOCvGckLzOw0tvWpD8HBT2r3wMnA0cAphKA6XQiGez8ys3kZz+lByUsLo4DBwG7AvbloLyoqYujQoXz55ZeYGevXr2fkyJH06tWL888/n+nTp7Ns2TK6d++eS3WO4zgOvqZfozGzZ82sa8ZxWsxbBXwi6VjYvNPgRGBanIbf08ymAlcQDAYbE/z/X5jywx9H77nweKz70FhHqSxfvpwvvviCgoICJLFixQrat2/PgAEDePzxx9mwYQPHHXcc996b0zeE4ziOg4/0aztnE3zy/y1eDzezD6J1/bjoB0DAbXFN/0/ArcBb8cNgAdCvtIdYiAcwFVhhOUbZO+6445g7d0uYgv79+7N06VImTpzIo48+yvHHH1+mF3Ucx3F8pF+rMbN3zKxX2izA/TF9g5kdaWZdzOwACz72MbM1ZvbzmL6/bfHtP8bMsvrDlXRo3AFwBPCApLcllegeOJOFCxfy5ptv0qNHD9577z1efPFFevTowTHHHMPrr79esUZwHMepRbj1vpNXJO0HvATMJ1j7LzGz67OU22y937Jly+4PPRSs99esWcPFF1/MWWedxdFHH82QIUM4+OCDufDCC5k7dy7XXnstDzzwAFsi/+afmmj1W9UkXaPrqxhJ1wfJ1+i+9/3Ybg9CvIBZhIA8dUsrn/K9v379euvTp4/dfPPNluKEE06wKVOmbL5u3769LVu2zKqSmuizu6pJukbXVzGSrs8s+Rrd976zPbMLwRBwB8I2v1KdCH300UccdthhrF+/niVLllC3bl0uvvhi6tSpw8CBA2ndujWNGjVizZo1tGjRIt/6HcdxagS+pu9UBf8Afg/cz7ZbCLMyc+ZMPvvsM9q0aUPdunW5/PLLueuuu7jvvvvo3bs33333HR999BHdu3ev0ql9x3Gc7Rkf6Tt5RdLZwAYze0BSXeBlScea2ZSS7hswYEBqaQAI1vsdOnSgRYsWjBs3DoDrr7+ejz76KJ/yHcdxahTe6Tt5xczuk7SjpKKY1Iww6i+x008n3Xof4KqrruK+++6jWbNmTJ06tfJFO47j1FDcet+pMuL+/ynAjWY2KSMvJ+v9dO6//37Wr1/PkCFDqkR/ippo9VvVJF2j66sYSdcHydfo1vt+bPcHISrg8NLKlWS9n86iRYts//33z5qXT2qi1W9Vk3SNrq9iJF2fWfI1uvW+s10jaTCwF5DViU8mxVnvn3/++UybNo169eoB0KFDh/yJdhzHqWF4p+/kHUndgWHAUWa2KZd7Utb7Xbp0AeDyyy+nfv36zJ07l4KC8N/222+/ZY899siXbMdxnBqHd/pOVTCUEMFvatxeN8PMzi/phuKs96dNm7Y57fHHH+eRRx7Jj2LHcZwaiO/Td/KKpGuBWWa2u5l1BZ4GZpeljkzr/RT33nsvJ510UuWJdRzHqeG49b6TVyS1BR4zs24xMt/7wGFmtjyjXJms98eNG8e8efO49tprq9w5T020+q1qkq7R9VWMpOuD5Gt0630/ttsDeB44GDgReKS08qVZ748ePdoOP/xwW716tVUHNdHqt6pJukbXVzGSrs8s+RrzZb3v0/tOVTAKGAwMAe7N5YYhQ4bQpEkTZsyYwaWXXgpAUVER++67LxdccAHffvstc+bMyZtgx3Gcmoh3+k5V8DhhlH8o8GwuN3Tv3p1169axevVqunbtSteuXRkyZAgrV66kZcuWfPPNNxx77LFccMEFeRXuOI5Tk/BOv4qQNECSSdq3jPe9XM7ntZWUdSgs6QRJRfFYJWlePL9P0gXRX36lYWbrganAQ2a2MZd7hg4dyoIFC9h7770pKiqiqKiIVq1accstt7B48WL+/Oc/c+qpp3LXXXdVplTHcZwajW/ZqzoGAdPizz9mZkoqMLPvMq/N7PuVLcTMniWOuCUVAsPMbEZlPydFNOA7HPhxReq59dZbOeGEExg2bBibNm3i5ZfL9T3kOI5Ta3Hr/SpAUmNgHtALmGRmnWJ6T+BPwFfAvgTr9c3XZtZR0iozayxpPDDWzJ6O944BngJmAGOBRvFxQ83s5Wg1/5SZHVCKtkLSOn1J1wCrzGxEzHsTOCrWfzZwJdAFmGBmV8d7zgIuAuoBrwK/TI3oJe0HPBfzPgbqAgeY2VYm99ms9z/99FOuvPJKRo8eDcBtt93GQQcdxDHHHMPUqVN56qmnuPnmm0tr/kqnJlr9VjVJ1+j6KkbS9UHyNbr1/nZ8AGcC98Tzl4Hu8bwnsBpol+06pq2KP08D/hXP6wGLgYbAjkCDmL4P0XoTaAvMyUFbIXBI2vU1hI+AVN5f4/nFhE67NVAfWALsAnQGJgE7xHJ3AmeX8LybgJtK0pSy3l+wYMFWvvWbNm1qmzZtMjOzTZs2WZMmTaw6qIlWv1VN0jW6voqRdH1mydfovve3bwYBI+P5+Hj9Rrx+zcwWpJXNvE7xDDBSUn2CUdwLZrZGUjPgdkldgY1Ax0rW/mT8ORt428w+AZD0IbAncCTQHXg97pdvCCzLVpGkM4BuQJ/SHnruuecyceJEVq9evTmtSZMmtGvXjo8++ojbb7+dffbZp/xv5TiOUwtxQ748I6k5cCwwStJC4DLgJ9riUWZ1xi2Z1wCY2VrCyPsE4AxgQsy6BPgMOAg4BGgYY9c/B3SQ9Ju4pl4emgAXxvNNwLq0vE0EmxARZiC6xqOTmV2TWZGkAwizCAMtB2O+RYsWIYl169bRpk0b7rnnHm644QYaNWrEjjvuyJ133sndd99dztdyHMepnfhIP/+cTliL/3kqQdL/EdbJy8oE4HxC5z44pjUDlpjZJknnAJhZ17im/wxwEtCULMaDOfANMKKUMpOBiZJuMbNl8SOniZktShWQtBPwIGHa//NcHjx58mQWLlxIv379ttqPf9ZZZ9GzZ09GjBhB9+7dy/xCjuM4tRkf6eefQYR96uk8GtPLynPAMcB/LWyDg7CGfo6kWQRjwHQ2EozjhipQV9JNkl6X9Jak1IfInyWdnHbfGZJOB3YCbolpDYGDJM2W9BbQIqa3AQxYKGkF4SOgdaoiSQXATKA98E9Jn0n6tBzv7jiO41QQH+nnGTPrlSXttrTLwrT0wvTrmNY47XwDIVpdev77wIGpa0m/iukLgQNiWl2gFdAf+NrMDo22AS8RttH9ooXEAAAYt0lEQVR1BX5CCIbzF+C8eP4FISQuBCPDR83s17HOnQmW+H8DDjaz1ZKuAOqb2fQ0fd9JOgV4JNZ1E7B15BzHcRynSvBOv3bRBzgwjuIhLA3sQ/FGgun3Hg8MTF2Y2VeS+gH7AS/FsvWAVzIfamZvSxpL2GJ4RNosxWYytuxRWFjIp59+yurVqyksLNyq7IoVK3jjjTdYtWpVedqgwqxatWobTUki6fog+RpdX8VIuj5Ivsa86SvNvN+P7esgbvFLuz4H+A4oAlYAHwCPZ7nvPuBU4AHg1JjWk7DXH8Jug30y7jkFeDBHXQ8CnwInlFY2tWXv6quvtvr169v+++9vAwcOtDVr1piZ2THHHGOvv/66VRc1catPVZN0ja6vYiRdn1nyNXrAHSdXGkWXum9HN7y/A/5sIZb95YStdz8BkNRRUsqpzwRCQJyjgP/EtE7A/vH8eeBXqYfE6f3pwA8k7R3TGknaZsugpB8SliWOBv4eDftKZMCAAVx//fVs3LiRFStWMH/+fC677DLatGnDK6+8wsknn8wJJ5xQtpZxHMep5XinX7NRPFKMAt4BZsYPgn+wZYknm5HgPODteH4dsLOkOdFosJcFS/zBwIPRuO8VMowJJbUAbgDON7P3gNvZ4rOgWO644w5at27NZ599xsKFC9l111055ZRTWLJkCevWreOzzz7j2Wdzit3jOI7jRHxNv+axOo7qAZDUnuA45xrCB8AOwFqCV73xZvZ1movf5vGeMYT19y8y6t5EsNRP/5goICwf1CUsHUzJuKce8C0wKa77dyFY8pfIHnvswbBhw/je975Hw4YN6dOnD336lOrTx3EcxykBH+nXcMzsQ0KH3Ipglf+1mR1KCHP7P5LaEab2U1P+9YDjCNb76fw+3tvFzA4EpsRR/NXA8WbWjRAH4NKM539s0XEP8E/CDoBFlMJXX33FxIkTWbBgAR9//DGrV69m3Lhx5W4Hx3Ecx0f6tY1qsd4HkPQD4H8Ibnuz5W9lvT9y5EgaNGjA22+H1YXOnTvz8MMP06ZNm3K9eGVSa61+K5Gka3R9FSPp+iD5Gt1634+cDra13m8PLCdMxz9KMdbz5NF6n+CsZy7QMZd36Nixo02YMMHq169vXbp0sQMPPNAKCgrstNNOsyRQE61+q5qka3R9FSPp+sySr9Gt950yI6klcBdwe/wP8SzwC0k7xPzSrPfTKZf1fnzWw8AVFgz5cuInP/kJV1xxBevXr2fjxo0UFBRwww035Hq74ziOkwXv9KsASQMkmaRMN7ml3fdyOR7XUNI7ktYC/yVY5Q+PeaOANcBKSWsIlvxFMUDPQML0/wLL4jyHclrvA98nxAoYHrcSFknaPZcXGT58OHPnzuVvf/sb3bp1o2PHyg4g6DiOU7vwNf2qYRAwLf7cJvCNpAIz+y7z2sy+X9YHmVndGGznKTM7KCNvE2k+/yUVAsPMbEZM+mlG+UKiW2AzW0Vw9JP5vCkEo8Di9Pwf0KCs75HO+PHjGTSoPKEKHMdxnHR8pJ9nJDUmGK+dR5ohnKSekl6U9CTwTuZ1LLMq/hyfHhBH0hhJp0tqG++ZGY8yfyRkaL1G0rB4XijpFkkzJL0r6VBJj0l6X9J1afecJem1OIL/R/Tzn17nsZKeSLvuLSkzAFGxrF+/nieffJIf//jHFXk1x3EcBx/pVwX9gf+Y2XuSlkvqbmZvxLxuwAFmtkBSz/TrjDpSW+qeTttS9wuCcV5vM1sraR+Cq9tDKlH7ejM7RNLFwESgO/Al8IGkWwjbAM8AfmBmGyTdCZxJMApMMRW4U1LLuBwwBLg380HZfO8DTJs2jXbt2vHuu+/y7rvvVuKrlZ9aa/VbiSRdo+urGEnXB8nXmC993unnn0Fs8UA3Pl6nOv3XMjr4zOsUxW2pawbcLqkrIYxuZS96Pxl/zgbeNrNPACR9COxJmMHoTnD+AyH87rL0CszMYrCdsySNBo4Azs58kJndDdwN0KlTJ+vZsydt27ZlxYoVNG3alGHDhjFjxozM26qFwsJCevbsWd0yiiXp+iD5Gl1fxUi6Pki+xnzp804/j0hqDhwLdJFkBCc5JumyWGR1xi2Z1wDEkXwhcAJhZD0+Zl0CfAYcRFiqWStpI8F9bvtocHcfcEtczy8r7SWdATwGrEtL30T4vyPgX2Z2ZSn1jAYmETwBPpxuv1ASmzZtok6dOsyePZtmzZqVXb3jOI6zFb6mn19OJ7i33cvM2prZnsACwra4spJtS10z4JPYof+U8FGxBugLzAd6AyeRxXgwR941s4tKyJ8MnC6pFYSPHEl7ZRYys4+Bjwne+0bn+vA6derw3nvveYfvOI5TSXinn18GAZlGa4+SZkFfBrIFxLkTOCeO6PclY6bAzJYR1smHKlBX0k2SXo9b7FpDdkNBoCXQXdJTMbmupNGSZhPsBo41s3eAh4AFcQvgB2T41ZfUQdJM4H5gMfBdvC4VSfTp04fu3btz991359ZKjuM4TrH49H4eMbNeWdJuS7ssTEsvTL+OaY3TzjcQwtOm578PHJiWdIWkVWa2EDgglvkwWtS3IhgVfm1mh0b7gJcI3vomAD8xs3OioeAHBPuAHsApZlYo6SSgvpl1geCcJ/rePwpoZWarJV0Rr6emafxA0tfAAILv/SHkONqfNm0ae+yxB8uWLaN3797su+++HH300bnc6jiO42RBwVGbU1OInX7jjLQVQCfgDsJHwrcxqxnwc+AF4D2CH/4TCR8AZ8YdBcPMrJ+kN4CB8UMjVW8/YAywJCbVA14xs/Mynr8g5u0NzAEOM7PlGWXSrfe7P/TQQ1u915gxY2jYsCFnnHFGmdukslm1ahWNGzcuvWA1kXR9kHyNrq9iJF0fJF9jefT16tXrDTMreQdXaX56/di+DtJ87xMM/94hhL4tAlYALxVzXz597zcgfFT0Bx4qrXzHjh1t1apVtnLlSjMzW7VqlR1xxBH2zDPPWBKoiT67q5qka3R9FSPp+sySr9F97zvlYSZhFP5nC6FtLwc+r0rf+xB2HxD8/v8vOU7tv/nmm7Ru3ZoGDRqwyy67sPPOO3PiiSfm9NKO4zhOdrzTLyeSdosGcB9IekPSv7N1eOWodydJv0y73l3SI2WoomH0jvc22X3vH0rwvb+eMNX+ioLv/ZXAyWxtKJhOeX3vp7ifsNXvuVxeokOHDrzwwgusXbuWzz//nPfff5933nknl1sdx3GcYnBDvnKg4InmccIe9YEx7SBgV8I0dqrcVj71c2Qn4JcEy3wsbHc7PdebzaxuCXmbgD2itsHAIWY2NK1IvYzyhVTQ934aRwKjzWxjDmVp3bo1rVu3BqBJkyZ07tyZpUuXst9+++Vyu+M4jpMFH+mXj17ABjO7K5VgZrPM7MUsPvW32iYn6ecQfPJLmhx95s+W1D9WdQPQIY7Wb1Lwrz8n3tMgtW1O0puSesX0wdEv/n+ib/wby/pCGc8ZLOkJSc9LWihpqKRL4zOnR6dDqe14/4kzHS8qI4qgpDpRz9MEL3y3SZqvEPI3ZxYuXMibb75Jjx49yvpajuM4Tho+0i8fB7DFlW420n3q/4yMbXKSniPsWT/NzFbGrW/T44fCb+O9XSF0xmn1/org2bZL7GCfS1tS6AocTPCcN0/S381scQXf8WCCEd584AozO1jB5/7ZwK0Et7kXmNn7knoQZieOTVVgZpskjYvvf6ukPsCsuCywFSrG9/6aNWu4+OKLOf/885k5M6ft/XmntvrsrkySrtH1VYyk64Pka3Tf+9sX6T70+wAHSkpN0TcjbI1bAvxF0tGEte49CMsDJXEk8HcAM5sraRFb/O1PNrOvASS9A+xF+LAoL1PN7Bvgm7jPflJMnx3fpzHwfeDhsNoBQP0s9dxLCNZzK3AuxRjyWRbf+xs2bKBfv35ccMEFXHrppRV4lcqltvrsrkySrtH1VYyk64Pka3Tf+8nibUpeZ0/3jCfgQjN7Nr1AXFNvCXS3EKFuIRWLO5/uG38jFf+3zfS1vy7tvICwNLQiNSNRHGa2WNJnko4FDiNE4SuVIUOG8OCDD9KoUSOeffbZ0m9wHMdxSsXX9MvHFKB+nJIGQNKBkrL51H8W+EWWbXLNgGWxw+9FGJkDfAM0Kea5LxI7zTit/z1CcJ0yoRCU51pgkKRZkn5D+DjJGTNbSXC/e4Wk2xQ4qJjio4BxhGA7ORnyde/enXXr1rF69Wq6du1K165d+fe//10WiY7jOE4G3umXg+gE4TTg+Lhl723geuDTLMVHERzkzIyGcv8gjJTvBw5R8GV/NjA31r2csO4/R9JNGXXdCdSJ90wABpvZOsrOGuAPwINsCcpzcTnqOZNg1HgMYfajfzHlngQaU4ZgO0OHDmXBggXsvffeFBUVUVRURN++fcsh0XEcx0nh0/vlJG6l+0mWrPfZ2qf+JuB38cjkiGLq/n8ZSSk/+msJDnQyy48huMNNXfcrSXvmPXHG4nWghYKf/v2BHnHv/R1m1jb6JBibuk8hKM9ThN0GKVe9jSWNJgTkMWC4mT1KNNADHpD0ATAkbgF0HMdxqhD3vV8LUcn++fsTAuhcpy1BeX5M2B0wwLIH5Ul1+n8lBOX5daxzZ+ASgifAfmb2X4WgPPXN7NqM52/je//TTz/lyiuvZPTonCcIqoSa6LO7qkm6RtdXMZKuD5KvMV++932kX4OR9CrbWtT/tJTbittt8AwwMn4InAi8YGZr0iz3AY4HBqYuzOwrSa8Bq4ARsWw9gue+rchmvb9w4UIaNWqUOAvb2mr1W5kkXaPrqxhJ1wfJ1+jW+06ZMbOs3mwyOmoktSdY/C+jmN0GsVwhIYjPGcD4HGUIeN7MBuUsHDj33HOZOHEiq1evLr2w4ziOkxNuyFfLid7x7gJujwaKxe02gDwG5clk0aJFSGLdunW0adOGe+65p/wv6TiO4wDe6ZcbJTfgTmn1vwo0krRe0nfARwTDvefic4rbbQAhWM4x5C8oz2YmT57MjBkz2H///VmyZAnnnXdeRV7bcRzHwaf3y4WU3IA7pZGa8lf2gDsvxZ9ZdxuY2QageUZaIZUXlMdxHMfJI269Xw6id7lrzOzoLHk9gT8BXxFGtJ0J29p6Eozq7jCzf0Q3thOBnYEdgKvNbKKk8QQL+nmE6fI7gKfM7ABJDQgx6Q8BvgMuNbOpsQM/FdgR6AA8bmaXl/IOg0nr9KOP/9RzBgMDgEYEI74RBAO8nxI88/U1sy8ldYj6WgLfAv9jZnMznjMSWG5m10o6AbgK6Bm3MqaXc+v9SiLp+iD5Gl1fxUi6Pki+xnxZ72NmfpTxAC4CbikmryfBDW+7eP0zQocOodOfAbQjzLI0jektCEFtBLQF5qTVt/ka+A1wbzzflzA134Awff4hwdK+AbAI2LOUdxhMWMfP9pzBUU8TQof+NSGwDsAtwK/j+WRgn3jeA5iS5Tk7Ehz39CJ8yHQorX07duxoZmYLFiyw/fff35LG1KlTq1tCiSRdn1nyNbq+ipF0fWbJ11gefcAMK+Xvq0/v5wcPuBMxs28l/Q/wAnCJmX1QAU2O4zhOBXBDvvLxNtC9hPxsAXe6xqOdmT1HcGGbCrjTFfiM7TTgTtrRuZi6ugDLgd1zffigQYM44ogjmDdvnlvvO47jVBLe6ZeP7TrgTmVgWwLu/DjqyRpwR9JehGWJg4GTJGX1HZDJgw8+yCeffMKGDRvcet9xHKeS8E6/HMS1k+054E5lcSZwXtyet03AnbjL4R6Cm96PgfOAUdEg0XEcx6li3HrfSRySvqGaZjBypAXwRXWLKIGk64Pka3R9FSPp+iD5Gsujby8za1lSATfkc5LIPCtt20k1ImmG66sYSdfo+ipG0vVB8jXmS593+jWY4gLumNns6tDjOI7jVC/e6ddgrJiAO47jOE7txA35nCRyd3ULKAXXV3GSrtH1VYyk64Pka8yLPjfkcxzHcZxago/0HcdxHKeW4J2+4ziO49QSvNN3EoOkEyXNkzRf0m+r8Ll7Spoq6R1Jb0u6OKY3l/S8pPfjz51juiTdFnW+JalbWl3nxPLvS9omzHAFddaV9Kakp+J1O0mvRh0TJNWL6fXj9fyY3zatjitj+rwY9bAy9e0k6RFJcyW9K+mIJLWhpEviv+8cSQ9KalDdbSjpXknLouOuVFqltZmk7pJmx3tuiw6zKqrvpvhv/JakxyXtlJaXtW2K+90urv0roi8t7zeSTFKLeJ2I9ovpF8Y2fFvSjWnp+W+/0iLy+OFHVRxAXeADoD0hjO8sYL8qenZroFs8bwK8B+wH3Aj8Nqb/FvhrPO8LPEOIq3A48GpMb06IdticEDL5Q2DnStR5KfAAIQQywEPAwHh+F/CLeP5L4K54PhCYEM/3i+1anxDp8QOgbiXq+xdwfjyvB+yUlDYkBLRaADRMa7vB1d2GwNFAN7aOrFlpbQa8Fssq3ntSJejrAxTE87+m6cvaNpTwu11c+1dEX0zfk+ACfRHQImHt1wv4L1A/XreqyvbL+x9UP/zI5QCOAJ5Nu74SuLKatEwEehO8AraOaa0JToMguFIelFZ+XswfBPwjLX2rchXU1IYQyvhY4Kn4R+iLtD++m9sv/rE7Ip4XxHLKbNP0cpWgrxmhU1VGeiLakNDpL45/2AtiG56QhDZk23DaldJmMW9uWvpW5cqrLyPvNOD+eJ61bSjmd7uk/8MV1Qc8AhwELGRLp5+I9iN01MdnKVcl7efT+05SSP1RTrEkplUpcRr3YOBVYFcz+yRmfcqW0MfFac3nO9wKXE6IcgiwCyHK4XdZnrVZR8z/OpbPp752wOfAaIUliFEKgaUS0YZmthQYAXwEfEJokzdIVhumqKw22yOe51PruYQRcHn0lfR/uNxI6g8sNbNZGVlJab+OwFFxWv7/JB1aTn3laj/v9B0nIqkx8CjwawtRBDdj4VO6Wva3SupHiMj4RnU8P0cKCNOY/2tmBxPCS29ll1HNbbgzISBUO0KI50bAidWhpSxUZ5uVhqSrgO8IwcMSgaQdgd8Bf6huLSVQQJhxOhy4DHiorLYCFcE7fScpLCWsw6VoE9OqBIXQx48Spiofi8mfSWod81sDy0rRmq93+AFwqqSFwHjCFP9IYCdJKa+a6c/arCPmNwOW51EfhFHGEjN7NV4/QvgISEobHg8sMLPPzWwD8BihXZPUhikqq82WxvNK1yppMNAPODN+mJRH33KKb//y0oHwYTcr/r60IUQ43a0c+vLVfkuAxyzwGmH2rkU59JWv/cq6PuGHH/k4CF+/HxJ+YVPGKvtX0bMF3AfcmpF+E1sbVN0Yz09ma4Og12J6c8K69s7xWAA0r2StPdliyPcwWxvx/DKe/4qtjdAeiuf7s7Wh0IdUriHfi0CneH5NbL9EtCHQgxD+ecf4zH8BFyahDdl2zbfS2oxtDdH6VoK+EwnhwltmlMvaNpTwu11c+1dEX0beQras6Sel/S4Aro3nHQlT96qq9qu0P0Z++FHRg2Bd+x7BUvWqKnzukYQp1LeAonj0JayZTQbeJ1jbpv4QCLgj6pwNHJJW17nA/HgMyYPWnmzp9NvHP0rz4y9/yhq4QbyeH/Pbp91/VdQ9jzJaIuegrSswI7bjE/EPaGLaEBgOzAXmAGPjH9dqbUPgQYKNwQbCCPC8ymwz4JD4vh8At5NhaFlOffMJHVXqd+Wu0tqGYn63i2v/iujLyF/Ilk4/Ke1XDxgX650JHFuV7edueB3HcRynluBr+o7jOI5TS/BO33Ecx3FqCd7pO47jOE4twTt9x3Ecx6kleKfvOI7jOLWEgtKLOI7j1AwkbSRs10oxwMwWVpMcx6lyfMue4zi1BkmrzKxxFT6vwLb4Rnecasen9x3HcSKSWkt6QVKRpDmSjorpJ0qaKWmWpMkxrbmkJ2Js9umSDozp10gaK+klYKyklpIelfR6PH5Qja/o1HJ8et9xnNpEQ0lF8XyBmZ2Wkf//COFJ/yypLrCjpJbAP4GjzWyBpOax7HDgTTMbIOlYgivnrjFvP+BIM1sj6QHgFjObJul7hJCpnfP4jo5TLN7pO45Tm1hjZl1LyH8duDcGYHrCzIok9QReMLMFAGb2ZSx7JPCjmDZF0i6Smsa8J81sTTw/HtgvLZBaU0mNzWxV5b2W4+SGd/qO4zgRM3tB0tGE4CxjJP0N+KocVa1OO68DHG5maytDo+NUBF/TdxzHiUjaC/jMzP4JjCKEB54OHC2pXSyTmt5/ETgzpvUEvjCzlVmqfY4Q0S/1jJJmGhwnr/hI33EcZws9gcskbQBWAWeb2eeSfgY8JqkOIb59b0L44HslvQV8C5xTTJ0XAXfEcgXAC4Twqo5T5fiWPcdxHMepJfj0vuM4juPUErzTdxzHcZxagnf6juM4jlNL8E7fcRzHcWoJ3uk7juM4Ti3BO33HcRzHqSV4p+84juM4tYT/D0I7r8keWRqOAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "plot_importance(xgb_model, max_num_features=40)\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "yA0ylVDO0hZE",
        "outputId": "e88df7b6-2f1f-475b-e613-a73e47bdb6a7"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "plot_importance(xgb_model, max_num_features=45)\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "LJTvlhh70jhI",
        "outputId": "36fe6d25-0726-4d91-ff0a-9e16ec66bbbe"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "plot_importance(xgb_model, max_num_features=50)\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "RThsO5t-0kpl",
        "outputId": "bf0fc2ff-f410-4c9c-f83a-025314224288"
      },
      "outputs": [
        {
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAf0AAAEWCAYAAABsT07JAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOydeZyPVfvH39fY9+WxZJ8oY6wTSqkYeVCiKE+lPEI9PYXSoqQFrSgqotKjFKnQYpCkHyalFDG2YbTYd2Md2wxz/f445zu+853tO6uvmfN+ve7X3Pc55z7nur+ep3Pf53yu6xJVxeFwOBwOR/4n6EIb4HA4HA6HI29wk77D4XA4HAUEN+k7HA6Hw1FAcJO+w+FwOBwFBDfpOxwOh8NRQHCTvsPhcDgcBQQ36TscDgcgIs+IyOQLbYfDkZuI89N3OBzZRUS2AlWBc17F9VV1dzb7vF9V/y971l18iMgI4DJV7XWhbXHkL9yXvsPhyCm6qmppryPLE35OICKFL+T4WeVitdtxceAmfYfDkWuISDkR+UBE9ojILhF5WUQK2bp6IrJYRGJF5KCITBeR8rZuGlAbmCsicSLylIiEi8hOn/63isg/7fkIEflCRD4RkWNAn/TGT8XWESLyiT0PFhEVkb4iskNEDovIgyJypYisFZEjIjLB694+IrJMRCaIyFER2SQi7b3qq4vIHBE5JCJ/ish/fMb1tvtB4BngTvvsa2y7viKyUUSOi8jfIvJfrz7CRWSniDwhIvvt8/b1qi8hImNFZJu17ycRKWHrrhaRn+0zrRGR8Cz9YzsuCtyk73A4cpOPgLPAZcAVQEfgflsnwEigOhAK1AJGAKjqv4HtnF89eM3P8W4FvgDKA9MzGN8fWgGXA3cCbwHPAv8EGgF3iEhbn7Z/AZWA4cBXIlLR1n0O7LTP2gN4VURuSMPuD4BXgRn22ZvZNvuBLkBZoC/wpog09+rjEqAcUAO4D5goIhVs3RigBdAaqAg8BSSKSA3gG+BlWz4Y+FJEKmfiN3JcRLhJ3+Fw5BSz7dfiERGZLSJVgc7Ao6p6QlX3A28CdwGo6p+q+r2qnlHVA8AbQNu0u/eLX1R1tqomYibHNMf3k5dU9bSqLgROAJ+p6n5V3QX8iHmR8LAfeEtVE1R1BhAD3CwitYBrgSG2ryhgMtA7NbtV9VRqhqjqN6r6lxp+ABYC13s1SQBetOPPB+KAEBEJAvoBg1R1l6qeU9WfVfUM0AuYr6rz7djfAyvt7+bIh7i9I4fDkVN08xbdichVQBFgj4h4ioOAHba+KjAOM3GVsXWHs2nDDq/zOumN7yf7vM5PpXJd2ut6lyZXRm/DfNlXBw6p6nGfupZp2J0qInITZgWhPuY5SgLrvJrEqupZr+uT1r5KQHHMKoQvdYB/iUhXr7IiwJKM7HFcnLhJ3+Fw5BY7gDNAJZ/JyMOrgAJNVPWQiHQDJnjV+7oWncBMdADYvXnfZWjvezIaP6epISLiNfHXBuYAu4GKIlLGa+KvDezyutf3WZNdi0gx4EvM6kCEqiaIyGzMFklGHAROA/WANT51O4BpqvqfFHc58iVued/hcOQKqroHswQ9VkTKikiQFe95lvDLYJagj9q95Sd9utgH1PW63gwUF5GbRaQI8BxQLBvj5zRVgEdEpIiI/AujU5ivqjuAn4GRIlJcRJpi9tw/SaevfUCwXZoHKIp51gPAWfvV39Efo+xWx4fAG1ZQWEhErrEvEp8AXUWkky0vbkWBNTP/+I6LATfpOxyO3KQ3ZsKKxizdfwFUs3UvAM2Boxgx2Vc+944EnrMagcGqehToj9kP34X58t9J+qQ3fk7zK0b0dxB4BeihqrG2ricQjPnq/xoYnkH8gVn2b6yIrLIrBI8AMzHPcTdmFcFfBmO2AlYAh4DRQJB9IbkV4y1wAPPl/yRubsi3uOA8DofDkU1EpA8mkNB1F9oWhyM93Nucw+FwOBwFBDfpOxwOh8NRQHDL+w6Hw+FwFBDcl77D4XA4HAUE56fvCDjKly+vl1122YU2I01OnDhBqVKlLrQZaRLo9kHg2+jsyx6Bbh8Evo1Zse/3338/qKrphlB2k74j4KhatSorV6680GakSWRkJOHh4RfajDQJdPsg8G109mWPQLcPAt/GrNgnItsyauOW9x0Oh8PhyEH69etHlSpVaNy4cVLZiBEjqFGjBmFhYYSFhTF//nwAYmNjadeuHaVLl2bgwIFJ7U+fPs3NN99MgwYNaNSoEU8//XRS3bZt22jfvj1NmzYlPDycnTszCldxHjfpOxwOh8ORg/Tp04cFCxakKH/ssceIiooiKiqKzp1NTqPixYvz0ksvMWbMmBTtBw8ezKZNm1i9ejXLli3j22+/TSrv3bs3a9euZdiwYQwdOtRv29ykHyCISDebv7tBJu75OYtjBYvI+jTqOolIlD3iRCTGnk+1+cR7p3afw+FwOAxt2rShYsWKGTcESpUqxXXXXUfx4sWTlRcvXpx27doBULRoUZo3b570RR8dHc0NN5jMzO3atSMiIsJv29ykHzj0BH6yf5MhIoVTu1bV1jlthKp+p6phqhqGSbF5j73urarvqerUnB7T4XA4CgITJkygadOm9OvXj8OH/U8oeeTIEebOnUv79u0BaNasGV99ZaJWf/311xw/fpzY2Nj0ukjC+ekHACJSGpN7ux0wV1VDRCQceAkTZ7sB8ID3tarWF5E4VS0tIp9jMmV9Y/v7CJiHmbSnAR4J6EBV/VlEgoF5qnp+wyl1uyKBwaq60l6PAOJUdYytW41Ji1oKE+N8KNAEmKGqz9l7emFihhfFxCbvr6rnUhnrAfuMVKpUucWwt/7n78+X51QtAftSzXgeGAS6fRD4Njr7skeg2we5Z2OTGuUA2Lt3L0OHDmXKlCkAHDp0iHLlyiEifPjhh8TGxjJkyJCk+xYsWEBMTAyDBg0CIC4ujtKlS3Pu3DmeeeYZrrzySnr06AHAwYMHGT9+PHv27KFp06YsXbqUKVOm0LVr199VtSXpoaruuMAHcA/wgT3/GWgBhGMSilxqy5Nd27I4+7c78LE9L4pJmlECk4a0uC2/HFhpz4OB9X7YFQm09LoegXkJ8NSNtueDMIlEqmEyge0E/oHJMjYXKGLbvQP0zmjc+vXrayCzZMmSC21CugS6faqBb6OzL3sEun2quW/jli1btFGjRn7XTZkyRQcMGJB07bGvb9+++vDDD6c5zvHjx7VGjRqqqur5b3x6h1veDwx6Ap/b8885v8T/m6pu8Wrne+3hW6CdTZV5E7BUVU8BRYD/icg6TNauhjlstyfL1zpgg6ruUdUzwN9ALaA95gVmhYhE2eu6qfZUQNmxYwft2rWjYcOGNGrUiHHjxgEQFRXF1VdfTVhYGC1btuS3334D4PDhw3Tv3p2mTZty1VVXsX79eWmGRzHct2/fC/IsDocjbfbs2ZN0/vXXXydT9qfFc889x9GjR3nrrbeSlR88eJDExEQARo4cSb9+/fy2w/npX2BEpCJwA9BERBQoBCgm1egJn+a+1wCo6mm73N4JuJPzLxCPYfJyN8PoN07nsPln7N9Er3PPdWFAMCsQ/ktLCxiFCxdm7NixNG/enOPHj9OiRQs6dOjAU089xfDhw7npppuYP38+Tz31FJGRkbz66quEhYXx9ddfs2nTJgYMGMCiRYsAoxgeOHAgt99++wV+KoejYNOzZ08iIyM5ePAgNWvW5IUXXiAyMpKoqChEhODgYCZNmpTUPjg4mGPHjhEfH8/s2bNZuHAhBw4c4JVXXqFBgwY0b94cgIEDB3L//fcTGRnJ0KFDERHatGnDxIkT/bbNTfoZICLdMPmvQ1V1Uybu+1n9E9r1wOzH/9feFwxswOyV+1LZfjEDXAaUtNdrgWPAMKAq0Me2KQfsVNVEEbkX80KRWeaISFNVPWiv64nIPD/vXQREiMibqrrfvuCUUdUMA0gUFKpVq0a1aia9e5kyZQgNDWXXrl2ICMeOHQPg6NGjVK9eHTCqXY+/boMGDdi6dSv79u2jatWqtGnThq1bt16Q53A4HOf57LPPUpTdd999abZP7f+3+/fv92yrpqBHjx5J+/uZxU36GeOtqh/uWykihVX1rO+1nxO+p//RPmVHbflfPuUHVLWLHScSuEqNyh4RKYL5qo9Q1Xjb/h3gS+tmt4A0VgpyC1WNFpHngIUiEgQkAANEZDtGRJqYl/YEOlu3bmX16tW0atWKt956i06dOjF48GASExP5+WfjnelR7V5//fX89ttvbNu2jZ07d1K1atULbL3D4bgYcOr9dEhNVW/LwykgynoR2YoR8x30evbBqtpFRNoC42xTBdqo6nEReRK4AyPq+1pVh9vn+s6O0wLo7P3FXxDV+x6VL8CpU6cYNGgQvXr1ok2bNowfP55mzZrRtm1blixZwrx58xg7diwnTpxgwoQJ/PHHH9StW5ft27czePBgPLkK9u7dy5AhQ/j444+zb2Au4lEmByrOvuwR6PZB4NuYFfvatWvn1PvZOUhFVa+pKOl9rzUfKeuBrUAlr+twzEsJ9v5r7XlpzMpRR+B9zH5+EOYFp419rkTg6oyeraCp9+Pj47Vjx446duzYpLKyZctqYmKiqqomJiZqmTJlUtyXmJioderU0aNHjyaVbdmyRYODg3PUvtwg0NXdzr7sEej2qQa+jVmxD6fezzZpqeohfyjrWwPHROQU0A94TUS+9ukntaUgT9ky4A0ReQQor2abo6M9VgOrMCshl9v221R1eY48XQCQlvL+0KFDdOjQgcsvv5wOHTokC8IRGRlJWFgYjRo1om3btqgqd911Fxs2bGDy5MlJ/VSvXp0ffvgBgMWLF3P55eYnPHLkCPHxZvdm8uTJtGnThrJly+bxkzscjosVN+mngZeqfrJd4n4SuENExDbxW1mP+fL2KOtn2CpvZX1LzCpATuKPsv59VS1hj+JAaVXt7tNPLFDB67oi4BH1Lcd4BJQAltkQwgKMVBvVT1UvU9UPbPs81RTkNh7lfXR0NMuXL2fixIlER0czatQo2rdvzx9//EH79u0ZNWoUYCbs/v37M2fOHDZs2MCsWbNYtmwZX331FSVKlKBo0aIUKlSI1157jWeeeYYnnniCZs2a8cwzz/D+++8DsHHjRho3bkxISAjffvtt0osGGMXwNddcw44dO6hZsyYffPBBqnY7HI6CixPypU0yVT2AiPxA6qr6jJgB3I+Z3PvYspxQ1meH1JT1kkq7SODfwDARKQT0AmbbuurAcVUdLSJXYr7qvwNeEpHpqhonIjUwAr58R1rK+4iICCIjIwG49957CQ8PZ/To0Xz66afcdttt1K5dG4AqVapQpUqVFArdW2+9lUsuuYTff/89xZjXXHMNmzdvTtUej2I40FOGOhyOC4f70k+bnhhXPW++JJXY+H6wEGgL/J8mV9bfKyJrMJNlnivrAY+yfi3wPXbSF5FwEYkUkS+AW4De1s7VwDngGRFZBQwB2tj7E4FuwMtADWC93bpYiRERApQWkaVWyZ+v8Fbe79u3L+ll4JJLLmHfvn0AbN68mcOHDxMeHk6LFi2YOjVlGgPvfhwOhyOncep9RxJeHgfhQATQCCMCXIbZ3lgJ/IHZ9vgTs4JRUo2S/1UgWlU/EZHywG/AFZj9/xXAQOA9jGrf1xXxolPvV6mYtvK+S5cuzJt3PpRB165dmTt3LuPGjSMmJoaxY8cSHx/PgAEDGDlyJLVq1Uq1n6wS6KpkCHwbnX3ZI9Dtg8C30an33ZHrB+c9DsKB773K38Us64dhhIie8ls4r+RfCawHouyxHRPQCIxg8CzwsD92XEzq/dSU9/Xr19fdu3erquru3bvV8zwjR47UYcOGJbXr16+fzpw5M81+csK+QCXQbXT2ZY9At0818G106v0Chk9ee8/hu92Qm3iL/86Rsf5DgNv1vICvtqputHVNMILA6rlgZ57giWvvHS/7jjvuoGrVqqxatYrx48cTFhYGwGWXXUbz5s0JCwujefPmbN68maioKG699Va+/PJLmjRpQmhoKHPmzCE0NBRV5b777iM0NJTHH3/8Qj2iw+EoALhJP0BRr7z2Xoevsj6v2QQEi0g9e+2tb/gOeNjj3SAiV9i/dYAnMEv9N4nIRblZ3adPHxYsWJCs7JFHHuHw4cNUq1aNkydPsm/fPubPn8/UqVNp2LAhJ06coE6dOtSpU4ewsDCqVKnCzp07iY+Pp1ChQgQHB7Nv3z6WLVvGtGnTWLx4MWFhYYSFhTF//vwL9KQOhyM/UyDV+yJyCfAWcCVwBOM696iqpi6L9r/f8sDdqvqOva4OjFfVrAVJTtn/r5gAOxUxbnK7bFV/4PGcGsdrvBFAnOdaTWKfYsACETkO/AiUsdUvYX7TtVaot0VEugIfYIIG7RaR+4CPRORKNa6MFw2pxbW/7rrrkpbMateuncyf3pME55lnnsHj5fn333/TokWLpLpp06bx5Zdf8s4773i2SxwOhyNXKXCTvv0S/RoTJe8uW9YMk6hms1e7ZDH1/aQ8ZgJ+B0BVd2Nc/3IEVW1lbeuDicY30Kv65xzov7T9GwlE2kkfn3FOAdfo+QQ8nntPAf8lJf/0avM7Jptgvvrf3Y8//kjVqlWTJnxvZsyYQUREBGCW/WNiYti6dSs1a9Zk9uzZSYF2HA6HIy/IV//x9ZN2QIKqvucpUNU1kDKmvoiEAqMwwrZiwERVnWRj8kdggtYUAZ5T1Qjbtp7NfPc9MBEbR19EimMEcS0xorbHVXWJncBvwYTlrYeJVf9UZh7IO16/7a8bJub+5cAYTOCff2P26Tur6iG7RD8RqAycBP6jmcsiWAqYCdTExBh4SVVniEgL4A1MWN6DQB9V3WPzAUQB1wGfAWPT6vtUwjmCn/7G7+fPbbaOujnd+s8++4yePVN6cv7666+ULFkySQdQoUIF3n33Xe68806CgoJo3bo1f/2VwpHB4XA4co2COOk3BlJGPTlPc6Cxqm6xbmRHVfVKu6y9TEQWYuLnd1fVYyJSCVguInOAp+29nsx3wV79DgBUVZvYyHULRaS+rQvD7HmfAWJE5G1V3ZHNZ7wCKI5xrRuiqleIyJuY5DtvYeLjP6iqf9h99ncwrnj+ciOwW1VvBhCRcjbT39vArap6QETuBF7BhPgFKKppuJN4u+xVrlyZmTeWSq3ZBcETaGfv3r2cOHGCuLi4pLJz584xY8YMJk2alFTmYeLEibRq1SpZeZkyZRg92iRVnDt3LsWLF09xX3bxti9QCXQbnX3ZI9Dtg8C3MbfsK4iTfkZ4x9DvCDQVEc8SfTnM1/NO4FURaYMJSlMDsz2QHtdhJkRUdZOIbAM8k/4iVT0KICLRQB3Mi0VWWaKqx4HjInIUkxgHTBz+pnalojUw63xUYYql0k9aG81q+xorIqMxqww/ikhjzAvH97bfQsAer/tmpOjJ06Hq+5gXEUJCQjSQIsr169ePefPmUb58eUqVKkXp0qV59913iYmJ4dixY5w8eZJXXnmFqKiopHu2bt3Kl19+yZAhQ5Ki4y1YsIABAwYgItxzzz0sWrSImTNnUr9+/TRGzhoXQ0S+QLfR2Zc9At0+CHwbc8u+gjjpbyD9fXbvyHiC8S3/zruBXUKvjMm6l2Bj8xfPhk2ZdY/LTH/esfc9cfeDgCOeFYl0iMVk5/OmjL03VkSaA52Bl0VkEUYrsUFVr0mjv4sy9n6fPn3YvXt3kgDvX//6F6NGjWLGjBn06dOHnTt3pgimc++991KhQgX+8Y9/AGZFYMCAATRs2JA///yT0aNH8/LLL+f4hO9wOBzpURBd9soA14vIME+BiDQVkdRi6n8HPCQiRUTkZxGpb/eyywH77YTfDvNlDnCc82p2D0VEZD1G6X6PHa8+UNseLwJ3ikiciMRgYvsPFZEHRaR3jj31ebpjIuSVFZFdIjJBRMpbMaMvS4FbRKSMtfs2YI2qnrOeCScxwsXXMdsiMUBlEbnGti8iIo1y4RnylDZt2vDee+8REhJCQkICs2bN4r777gNgypQpxMTEJNvTnz17NldddRUPP/xwUtlvv/3GZZddxty5c9m4cSPDhw8nISFfpiRwOBwBTEGc9HtiQsTeKSJ/icgGYCSw19Z7J52ZDGzEpIgtC0zCfClPB1ra2PK9Mf7rqGosZt9/vYi87jPuO0CQvWcGRuA2Dxhmr1diXgp+xGSpe09VUwZnzxnuwSzDrwfuwLj+3erbSFXXAhOAn6w48UFM4iAwAXd+wwgQhwMvq8kr0AMYbWP1R2G2EfItvsr9uLg4Ro8ezfDhw5O127VrV1K4XYCaNWuya9cuHA6HIy8pUMv7di/7OoyCf66q1rPl4cCHGNX+5b4qflWtb+PSNxaRzzHZ9zxfsx8Bz9vzHzGq+XiMCn+riHTC7HmfBvr62qSqH2F81yPtdRfb1whMWNwxtm41ZhWgFOZF4xYR+QOYoarPAY1FpBfmy7uoiEwC+qtqsPdYdmsCq1voJCZz3p8YbwRsH49gFP+/2v4EqKeqT9o2fTAug0098fpt+RBMuN5E4FtVfVpE6onIAswKyJsikqGXQCCp9zOr3B8xYgSPPfZYQMf0djgcBZcCNeljvmYXqOpmEYkVkRbWdxySq/bDva99+piB+Tr+RkSKAu2BhzATYwcbwOZyjFta+okPMke8qrYUkUGYCboFcAj4y6ryqwB3AtfabYd3MF/06a4W2KX6NRgXxfg0+vgS+AWTdAfb5hXvfkTkJszv20pVT4pJ1Qt+egkEqnrfV7kfGRmZpKpNTbm/cOFCPvnkEx555BHi4uIICgpix44d1K9fnzVr1iS1W7p0abL+c5JAVyVD4Nvo7MsegW4fBL6NuWZfRsH589MBzMNMzGC+Zsfo+QQzS7zaJbvW5MloimOSyRTDTHLTbXk5YBpG1R4FnLTlwcB6P2yLxHw9e653YpbdozBR8f7ALKnfQPJkOEsxLn8DMRnxPAlvYoARGY1jyyIwE3mafWDSA18N/APYwvkMjZ7fZSzG19+739KYYD5RXsfGjH6L7Cbc2bRpkzZr1izpKFOmjL755puqqjp+/HgNCQnRhg0b6pNPPpl0z5o1a/Tqq6/Whg0bauPGjfXUqVPJ+tyyZYs2atRIVc8nwvj222+1TZs2adoxfPhwff3111VVNSEhQS+99FL9+++/9cyZM9q0aVNdv359tp4zLQI9kYhq4Nvo7MsegW6fauDbmFsJdwrMl7798rwBExFOMe5kKiKer1dfZXmqSnM1X/KRQCfMRPm5rXoME863GUYrkd0ws5NJvrw/WFXX2VUIX3V+YcxKw8eqOjRFT+lgl/ebYLQLVdLp43PMCscmzNaFP3Fj/fUSyFFCQkKS3OfOnTtHjRo16N69O0uWLCEiIoI1a9ZQrFgx9u/fD8DZs2fp1asX06ZNo1mzZsTGxlKkSJGk/nr27ElkZCQHDx6kZs2a9OzZk/DwcD7//PNUg/KkRuHChZkwYQKdOnXi3Llz9OvXj0aNLnqNo8PhuMgoMJM+RmA2TVWTQsWKyA+YfXLPdTeM29m9GfQ1AyNoawn0EZGfMfvfO1U1UUTuxbxUpIt3JL1UqusBN9g99ibAFyLiyRNQTUR6a3Kh3yIgQkTeVNX99iWnjKpuS2f8Iphl+h2qulZEJgOXpNHH18CzmKA/Q7y68azDfw8ME5Hpapf31UT+2yIi/1LVWWKc95uqjYCYFyxatIh69epRp04dnnzySZ5++mmKFTMhCapUqQKYJfmmTZvSrJlxYPC42Xn47LPPkl17ltw++uijdMceMWJEsuvOnTvTuXPnLD6Jw+FwZJ+CpN7viZm4vPmS5JniegI/kUZkOjkfM34h0BYTVCdeVVtj9qrv9eyPk32f9L+AN+1X8nHgeXs+EtjuM+GjqtHAc5hIf2sxk7Cvj72H6bbNesyk7VHun8SECk7Rh6oexqwG1FHV33w7VNUFwBxgpVX6D7ZV9wD32d9lA6l4CeQm3l/jmzdv5scff6RVq1a0bduWFStWJJWLCJ06daJ58+a89tpreWmiw+Fw5BkF5ktfVdulUjbecy4iKzF72B5lf4gtD8co+RcD0VZw9hLm5aC1bROnqqVtNL1pqvoNMMQq++cBXbyU/QADVTVZghxVDfe5HuF1GYWZcFHVSBEJF5HBqjrG1t8jIu9xXtk/FLM60AVYbm30VuVHYpT951L5qb5X1ZEiEgeMAyaJyClMaN0uInKpiPyC2a+PIPnLzTl7XQzwOKGHYfIThAGXAD+IyPuqupc0yI5631ttHx8fz5w5cxg5ciRglvEPHTrE8uXLWbFiBXfccQd///03Z8+e5aeffmLFihWULFmS9u3b06JFC9q3b58lGxwOhyNQKTCTvh84ZX9ySgHLVfVZEXkN+A/wMuZF4F1VnSoiAzyNRaQjJkTxVfZ554hIG1X9WkRux+QeuBEYntqEn1PqfW+1608//cSll17Kxo0b2bhxIyVLlqRu3br88MMPgHkpiIiI4NixY9SvX5/169cDEBoayqxZsyhUKPUdmgKr+s1BAt1GZ1/2CHT7IPBtdOr9i1/Z/38YJfspjPguChPxLyvK/hEYYZ+n7lp7nmPKfoxY0KPQvxOYbM9jgSL2vKzXs48BtnqN8Sdwn62rgPFE+NKff4ucUu+XL19ea9WqlaTef/fdd/X555/XMWPGKKDVq1fXxMREHTFihJYoUUKbNGmiDRs2VEA/++yzNPvPj6rfvCbQbXT2ZY9At0818G106v1cJI+U/VGYBD4eZX8TzFJ7XbvfPRWzh5+YhUeoKyaj3VdkT9nfAPM7rMSkH/Yo9H3zAXgr90uKyTQomEiCk1Lpt6a1paqIBGXxGf0mJCSEZcuWUbt2baKiomjYsCHdu3enWrVq3HXXXSxcuJAiRYowYcIERIThw4dTr149Ro4cSVxcHLVq1eKuu+7KTRMdDofjglCQhHzp4VH211HVYFWthfFFTy0ef0bMwETeux5YYMvKAXvsZPdvzEvFKUyymj+BDsBNmHC2WWGjqj6STv0ioIeIVAHzkiMiddJpnx7LAM+MeI9X+XdAPxv1EBGpISJVrPjxQ4xIcunocZQAACAASURBVCPweBbHzRSlSpUiNjaWlStXJqn3ixYtSlBQEMuWLaN69epcf/35f95evXqxYcMGrr32Wp5//vm8MNHhcDjyHDfpG/xR9vuLR9n/f2pi0UMGyn5V3Y/Zzx4ohkIi8rqIrLAq+moAIvK5iCQp1axQsDLQQkTm2eJCIjJFTIz/lsANapT9M4EtVpT3F1A3g+coKSIviMgqTOCdcrZ8ODBBRE5bmz3PsBDYDOy3Y6yy90zABChaifEueEVEumYwdo7hrd6PiIigRo0aSa55vpw8eZIFCxZw++2355V5DofDkae45X0yVvZj9ro95ZHe17astNd5AlDRp/4PoKlX0RCr+N+KSXyDqv5tA+VUwegBjqrqlSJSDPN1HYsVCqrqvVYo+BdQH2gFdFWj7L8JKKaqTQBEpIJdfr8eqKKqJ8TEyL8eWOLz2AuAv+35duCgqjYXkf4Y8SKYVYw3VPVF+wIyz44TCpQHyul5seA1qvqgiBzEiABLYLQEc31/b29yQ71/8uRJXn31VRYuXJjmvXPnzuXaa6+lYsWKabZxOByOixk36QcmHYGmItLDXpfDKOO/BcbZF4EbgaWqesrEvEnin5xffkdVD4tIF6AhJgMgGC3BL37Y8ZX9+ztwmz1v4zlX1W9E5LAtb4/xGlhhxygB7Ld1L2LS+Z7GiCRTkNvq/b///pvNmzcTEhICwIEDB2jUqBHvvvtu0iQ/YcIE2rZtm6FitsCqfnOQQLfR2Zc9At0+CHwbnXo/nx1Y1bs97wREA2cxgr8jwLI07psK3AJ8Ctxiy8Ixkf3ATNCX+9zTFfjMD5s+AnrY861AJXveEoi051FAXa97DgGVgIcxQr7U+q2GWZWIBkplZEd21fse7rzzTv3www9TratTp44eOHAg6frIkSNaoUIFjYuLy7Df/Kj6zWsC3UZnX/YIdPtUA9/G3FLvuz39wGAVJsHOK2qi7j0FHLBhchGR+iLi+fRNTSjozfcYn3jsvRUwAXquFZHLbFkpEamfRVuXAnfbfm7CuONB+mLBSZj0w9OB0Vkc129iYmJo0qQJX3zxBW+88QZly5blrbfeSqofO3Ys27ZtIzY2FoDp06fTuHFjEhMT6dChA2vW5FmUYIfD4chT3PL+haOEDVdbBPOFPw14w9ZNxojfVtl49Q0w2e3ACAWnARF6XijozcuYOP3HMB4IL6jqVyLSB/jMbg2AEdVtzsDGp0RkP8k1DN0wQX96Aj9j9v5R1WgR8YQBDsJE5BsgIm0x7n+fWs3CzyJyg6ouzmDsLBMSEsK6deuA5Al3AHbs2MHChQupXbt2Uoz9Sy+9lLVr11KhQgW+/fZbHnjgAX799dfcMs/hcDguGG7Sv0CoapoJedS49j1jD0+Y36O2LjWhYCR2YlbVOBEZhQne08WrzWLgygxs6uN1HiwiI+z5SswWAhh/+3+p6sFU7p+BWYnwZjk28p+asL+tvHIY5DreCXcAHnvsMV577TVuvfV8CoDWrVsnnV999dXs3Lkzr8xzOByOPMVN+hcRNgTwCOAgRvX/O9BLVVVEbgTewiTN+cnrnlLA27Z9EYx6PkJExgGxalT4nTAZ9MLVz8A5tt+ZmMA7hYCXVHWGiLTArFiUtnb2UdU9NmhRFHAdJgzx2LT6zin1PmTOZQ/ggw8+4KabbsrS2A6HwxHoeMKsOgIYr4Q+4ZjY+o0wYXWXAU9ifOD/wETT+xPztV1STYKcV4FoVf1ERMoDv2HS4ypGUT8QeA/orKp/+Yw7AiM4HONVthUj7GsL3Kiq/7Hl5TAvHD9gkvMcsFECO6lqPzvpR6tq/zSe0Vu932LmzJnZ+MUMCQkJ9OjRgylTplCyZEkee+wxXn/9dUqXLs1dd93FpEmTKFeuXFL71atX89ZbbzF+/Phk5b7ExcVRunTpNOsvNIFuHwS+jc6+7BHo9kHg25gV+9q1a/e7qqaf1yUjpZ87LvzB+fj24SSPrf8u0AsTX3+pV/ktnFfzr8Sk0PXExN8OhNq61hg9wcNpjDscG+Pfq2wrRl9Q356PBq63dY2BY15jrQMW2rpIoK0/z5tT6v3Zs2drhw4dVFV17dq1WrlyZa1Tp47WqVNHCxUqpLVq1dI9e/aoquqaNWu0bt26GhMTk2G/+VH1m9cEuo3OvuwR6PapBr6NLva+w4N3bH3fmPipIcDtqhqTSl0TTNCf6mncG4uNBuhFGeCIqsaKSHNMKOGXRWQRJqrhBlW9Jo3+Us1Z4A87duygd+/e7Nu3DxHhgQceYNCgQTz//PNEREQQFBRElSpV+Oijj6he3TzOuHHj+PPPP2nUqBGVKlVi//79Sf0FBwezcuVKKlWqxPbt27ntttuYNm0a9etn1anB4XA4Ah/nspfLiMglNnzuXyLyu4jMz4a7nKfP8jZKnoejQEsRqWevvcMHfwc8bL0AEJEr7N86wBOYpf6bRKSVV/+/Ws+CZzHRA9eKSJSIDAXOqeo5EakOnFTVT4DXMRH7YoDKInKN7aeIiDTKzrN6KFy4MGPHjiU6Oprly5czceJEoqOjefLJJ1m7di1RUVF06dKFF198EYBdu3axdOlS5s+fz4YNG5g1a1aafb/44ovExsbSv39/wsLCaNkyJ7MeOxwOR+DgvvRzETvRfo3JcHeXLWsGVMW6y4lIYVU9m8muywP9Oa/Y3yIitwHfiMhJ4EfMFznASxiB31rrSrfFxr7/ALN0v1tE7gM+EpErVfW0qraytvUB7uW8t0A74Gp73gR4XUQSMe55D6lqvI0iON7u8Re2Y2/I5POloFq1alSrZhYdypQpQ2hoKLt27aJhw4ZJbU6cOIEnOmFERARPP/00jRs3BqBKlSrJ+tu6dWvS+eTJk5k8eXJ2TXQ4HI6Ax036uUs7jI/6e54CVV0jIuEi8iNwGGhg49aPwuzZFwMmquokMRnrIoDNNoHOc2rEeZ8D9TD78adEJBgYo6oNRKQ4Zq+/joisBh5X1f/aCfwWoCTmheNrVZ1jbfodM4mnxgZVHQhgx5mH2buvhonTXwoTIvhqEWmDySJ4BminqodEpJ6ILMC8hLwpIv9R1U3p/Wipqfe9Vflbt25l9erVtGplFieeffZZpk6dSrly5ViyxKQT2Lx5MwkJCYSHh3P8+HEGDRpE79690xvW4XA48j1OvZ+LiMgjwKWq+phPeTjwDdDYfqU/gEmG87JXgp1/ATswKvxjNmnOcswEWwcj1Gts+wv2XIvIE0AjNYr5BphgPvUx8fiHYZbzz2CW4q9T1R3p2N8HaOk76dtx+mAC/FwBFMd4DQxR1fdE5E1gm6q+Zff6H1TVP+wWwkhVvSGVsfxS7586dYpBgwbRq1cv2rRpk6xu+vTpxMfH07dvX8aNG0dMTAxjx44lPj6eAQMGMHLkSGrVqpXW4/pNflT95jWBbqOzL3sEun0Q+DbmlnrffelfOH5T1S32PK0EOzuBV+0XdCJQA7M1kB7XYfzyUdVNIrINM+kDLFIb5EdEojEvD2lO+n6wRFWPA8dF5CjgyZ63zj5PaYyHwCw5nxSoWMpuQFXfB94HCAkJ0fDw8BRtEhIS6NKlCw8++CCPP/54ivq6devSuXNnPv74Y5YvX07Tpk2TfO7nzJlD8eLFSa3fzBIZGZkj/eQWgW4fBL6Nzr7sEej2QeDbmFv2uUk/d9kA9EijzlvJLhi3ue+8G9iv6cpACzXpardivqqzSmaV/5npL9HrOtH2HYRR+odlcxy2b9/OVVddRXx8PDt37qRQoUIMGjSI8ePHM2nSJDZu3MgTTzxBgwYNABNa9/7772f27NkkJiZy9OhRHnvssQxGcTgcjvyNU+/nLouBYnbpGgARaYpJluPNd8BDqSTYKQfstxN+O8yXOcBxzgv1fPkRuMfTD1Abs5Sf56jqMYxw8F/WHrFCxkyzatUq9u3bR82aNSlUqBBPPfUU7733HvPmzSMhIYGSJUuyfPlyxo0bB0DXrl157rnnSEhIICEhgUOHDiW9EDgcDkdBJVcnfRHpJiJq95Yzc9/PWRwvWETWp1HXybqdRYlInIjE2POpIvKgiOSGyqswJurdWBE5Y5X1k4G9Pu0mY9LOrrL2T7L3Tse44q0DegObAFQ1FjgoIhtE5HXbRx0bce8dIMjeMwMTBvcMmUBEJlqXvVeABz2/G3CT1zj+cg9wn4iswax83JpB+1Tp1q0bqsratWtZu3YtN954I/Xq1WPhwoVs3ryZli1b8uabb1KjRg0ASpYsyZAhQ4iOjubbb7+lVKlSGYzgcDgc+Z/cXt7viYkD3xMT3S0Zvu5qnmtVbe3bNrvYpfPv7DiRGHe1lTk9jg8vYb7IK6nqGRGpiolKNxP4n1c7UdWkBDs+pBXo5h8YkZ0n8Y33l39f38aq+hHwkdd1F982XnUDIJlwz3t5/l3717e/4NTGsrqFG9MaKzV81fu+8fR91ftp8euvv9KvXz+2bdvGtGnTKFzY7WY5HI6CTa6p962IKwbjtjZXVUNseThmMjyMSRn7gPe1qtaX87HmPwemqeo39t6PMC5jKzHpZT2fbwNV9WdvdXkGtkXiNemLV4x5W7caswRfCvOFPRTj0jZDVZ+z9/QCHgGKAr8C/dVkkfOMURIjkrvULnP72hCH+aL/JzAAk0o3RX8i8i4mO14J4AtVHW69AsbY3/egqraz+/0tVfWgiDwO9LNDTbYq+mDgW8xLWGtgFyZG/ql0fqcUv6ecj71fGliA8ShojVnRmAK8AFQB7lHV3ySNhD+pjJUt9f6jjz7KQw89REhISIp7tm3bxqhRoxg3bhxFixZN63H9Jj+qfvOaQLfR2Zc9At0+CHwbL7rY+5hl3Q/s+c8YMRoYX/QTmMkwxbUmjzXfHRPYBsxkuAMz+ZUEitvyy7HxhjET53o/bIvETJCe6xHYGPO2brQ9H4RJbFMNozrfifnCDsUo1YvYdu8AvX3GaAqsTscGBe6w52n2B1S0fwtZ25rq+Rj4lbz62wpUAlpg1POlMBPzBoxbXTDGrz/Mtp+Jidv/K+dj5XuOJmn9nl7jePprgtkm+h34ECNKvBWYbdu/iskECCao0GagVHr/PmnF3o+Pj9eOHTvq2LFjU9S1bdtWV6xYkep9qqrt2rVLtz4z5MeY3XlNoNvo7MsegW6fauDbeDHG3u8JjLPnn9vr3+21t7taatcevgXGWd/1GzFJZU7ZaG8TRCQMo0LP6YDpc+zfdZjgNHsARORvoBbGLa4FsMK6opUA9qfST3qcA7605+3T6e8O+xVcGPPy0RBYm06/12EC75ywNn+FWbWYA2xR1Sjb7ncgWG30vSyyRVXX2XE2YFwC1eoJgm2bjsAtIjLYXhfHiAs3+jtIv379mDt3LgkJCfTp04fHH3+cESNG8L///Y/KlSsDkJhoMgJv3bqV0NBQgoODKVq0KNdccw1Dhw5l06ZNBAcHpzOKw+Fw5H9yZdIXkYqYNK9NREQxX6kqIk/aJr6JV1JNxKKqp+1yeyfgTszLA8BjwD6gGeYr83SOPkBy1zNft7TCmK/Zj1V1aDp9/AnUFpGymsryPnBaz28HpNqfiFwKDAauVNXDdnsjJ132SvhxTxERWa/Jt0yewoTezchlD9JP+OMXffr04dprr+X+++9n8eLFhIWFsXfvXm6++Wa6dOnCww8/zIEDB7j55psJCQmhXr16DBkyhFGjRrF8+XK6d+/OO++8Q6VKlbJqgsPhcOQLcku93wOzF19HVYNVtRawhZSuav4wAyNMux6zhwzGlW2PqiZiwr4WygGbM8MioIeIVAHzkiMmgU0SqnoSE99+nIgUte0qe9zX/OyvLOaF6KgVAd7kdU9abns/At1EpKTdT+9uyy4UKRL+iEimXjbbtGlD+/btadSoUVJynQcffJDQ0FC6d+/Ozp07OXPmDPv27WPq1KkA/Pvf/2bDhg1ERUWxatUqunXrlvNP5nA4HBcZuTXp98QkmvHmS5Jnf/OXhUBb4P9UNd6WvQPca93AGpCNlK1ZQVWjMSFoF4rIWuB7UqagxbY5AERbV7x5mHzzfvWnqmswosJNwKeY8Lwe3gcWiMgSn75WYZTzv2H26yer6uqsP226VBSRaGtzWwD7ovEaUFdM7P/VGAHfdhE5Zp9tUXqdetT7vvH3fZkwYQJNmzalX79+HD58OKl8y5YtXHHFFbRt25Yff7yQ7zsOh8MRWLjY+450SUPBPwKIAx7HCDDPiEh5VT0iIq8C0ar6ifXn/w0jJPwX8DJGiHgolXHSVe/v3buXoUOHMmXKFAAOHTpEuXLlEBE+/PBDYmNjGTJkCPHx8Zw6dYpy5coRExPD888/z5QpU3LUTz8/qn7zmkC30dmXPQLdPgh8Gy869b478seBiQLoq+AfATyB2W75AuMFUNrWrQTWc94TYDvGO6EPMMWfMVNT72/ZskUbNWqUojyjuoxU/VkhP6p+85pAt9HZlz0C3T7VwLfxYlTvXzBEpBMw2qd4i6p2vxD2BDIiMhG41qd4nKpOseexQAWf+ooYjcbNQBugK/CsiDQhDeGemAx7md6GefPNN5k8eTIJCQns37+f06dPM2DAAH755RcKFy5M/fr1admyJY0bm4WIAwcOULFiRQoVKsTff//NH3/8Qd26dTM7rMPhcORL8uWkr17R9xzpozb6Xjr1cSKyR0RuUNXF1jPjRow7Zi1VXSIiP2FS95bmvHDvYVVVEblCs6gp2LVrF+PHj+fKK6/kxx9/5NixY1SvXp0XX3yR+Ph4oqKi+OGHH4iOjmbRIiMTWLp0KcOGDaNIkSIEBQXx3nvvUbFixawM73A4HPmOfDnp+yIilwBvYSLbHcG4+z2qqpuz2W954G5VfcdeVwfGq2pamfUy2/+vmKBAFTHudbtsVX/g8Zwax2u8PpigRQN9qnoDE0XkDXv9AmbZfomINMJ4EoxXs6f/Eua3XisiQZgVgTRD/mbE2bNnee+99yhbtizdunXjkUceoWPHjgwcOBBVpX///gQHB1OtmtFR3n777dx+++1ZHc7hcDjyNfl+0reuYl9j/ODvsmXNMHnpN3u1S5YHwE/KYybgdwBUdTdpp9LNNGoD56QxGWcpKVEW7YjGhFP25TpPaGRV/cK2PQX8N5U+PsIrVn96eNT7W0fdzODBg6lduzYlSpSgY8eOdOzYEYC+ffsyf/58GjZsyNixY7P0XA6Hw1HQyPfqfRG5ARPvvU0qdeEkzwMQCozChAYuBkxU1Uk2j0AEZm+7CPCcqkbY3AC3YmLgfw9MxCrdRaQ4JjlNS0y42sftUngf4BZMKOF6mOh5T2XwDH3wmvS9FfW2rhsm7O7lmJj8RTHxC84AnVX1kIjUs/ZVBk4C/1HVTWmNYyfzY9b+S4CnVPUL+xL1NtABExY5HvjQ1rUA3sAs8x/EiPdOYhT8t6hqjIh8BixW1f/5jJ1CvX/8+HGGDx/OsGHDKF26NCNGjKBt27Z06NABgHPnzjF+/HgaNGjATTd5hzDIXfKj6jevCXQbnX3ZI9Dtg8C30an3s64+fwR4M426cJLnAXgAM6GDmfRXApdiVkTK2vJKmGh7gk9seu9rjLr9Q3veALMcXhwzEf6NCTBUHNiG2RtP7xn6ABPSGKePtacMZkI/Cjxo697EbGOA8Y2/3J63wky8aY6D+SqfhYnl0BD405bfhnnBKQRUx2yX9MC8DP0MVLbt7vR6/g7AL5h9/wUZ/Zt51PszZ87Ufv36qYePP/5YH3roIfXmhx9+0Jtvvlnzkvyo+s1rAt1GZ1/2CHT7VAPfRqfezz284/53BJqKiGeJvhzm63kn8KqItMGEmK2B2R5Ij+swX8So6iYR2cb5HAGLVPUogIhEY9zidmTjGZao6nHguIgcxSTvAZM7oKldqWgNzLKB8cC81GTEbDVRD6NtREAwav3P1IQQ3i0ii215CCaT3vd2jELAHgBV/d5GIpyICZ3sF7Vr1+abb74hNDSUoKAgTp48ycMPP8wtt9xCTEwMhQsXRkSSlvwdDofDkT4FYdLfQPr77N5uZAI8rEb9f77QLHtXxmQKTLDpZXMyBn52/x0yioEfBBxR1TDvm9JIRXyVV3Ic736F9BFMcqJrUlQYQV8oZqm/AuYlKkNq1qzJ6dOnSUxMpFChQiQmJlKmTBn+/PNPihYtiqoSFxdHzZo1/enO4XA4Cjy5FYY3kFgMFLN7xgCISFMRSS0PwHfAQyJSxLarb8PKlgP22wm/HebLHNKOfw8m3v09nn4wmeWynHQmO6hJ+LPFE/dfDH5/cfuwFLhTRAqJSDXOC/xigMoico2IFBaRIlbZDyZB0kbgbmCK5/f1hzJlyvDLL78QFRVFkyZNqFOnDtHR0axbt47169czaNAgYmNjs/goDofDUbDI95O+3efoDvxTRP6yKWBHAntTaT4ZiAZW2Vj5kzBfytOBljZlbG9MLHxUNRZYJiLrReR1n77eAYLsPTOAPqp6hgvHPcB9Nl/BBowAMSNu8oqt79kO+A6TXjgOM9H/ZcvvxnzBf4vZ548CWotICCZT4NWY7Y7L8fKaSA2Per9GjRpJ6v1q1apRrly5ZEv5CQkJTJs2jRtvvNGf53c4HI4CT75X7zvSJr24+qo6RkR2k0Ox9W3fRTArL6+p6lyfukyr98eMGUPx4sUZONA3rEDukh9Vv3lNoNvo7MsegW4fBL6NuaXeLwh7+o60SeuNz1O+FpguIrOB2basI3CL175/cczWBcD3aU34lnEYr4G5vhWq+j4mcyAhISEaHh7OrFmzuOKKK5LS4u7evZvly5cTHh7OCy+8QOHChZk5cyZBQXm7YBUZGUl4eHiejpkZAt0+CHwbnX3ZI9Dtg8C3Mbfsc5N+gOAVfc+bf6vqulwcNr24+pCDsfWtGLIO4Pdnuaoyffp0VqxYgYgQHR1N165defDBB5k6dSqnTp1i1apVtGyZvluqw+FwOAz5fk//YkFVW6lqmM+RaxO+XdpfDuyxAYwQkdFAT+Anq7ivpapLgCEYMaN3bH2x91zhx1gtMPv6vawLoF/ccccdSelyz507R+HChRk1ahSTJ0+mUqVKlCpVirvvvpsXX3wxM4/ucDgcBRY36Tt6A8+LSBRwL7BQVf/C+Nl/YoWIq7Gx9TERDItgYutvsNdJiEhqq0cDMSsIS0QkSkQm+2vcCy+8wKZNm3jjjTdo3rw59evX5+zZs2zfvp2WLVvy6aefMmzYsKw8t8PhcBQ43PJ+AUe94up7RHy26iHMRH0W438/yrovvo0JwnMWE944wi7d34ZZCVgEtPX0LyJXAi2AupgXid8wCXnS5FTCuRRln3/+OT179szqYzocDocDp94vsOSlcl9EXsYI/koAO1V1ZCptktT7lSpVbjFr1sykuoSEBHr06MGUKVOSpcl99NFHeeihhwgJCcnuz5Ep8qPqN68JdBudfdkj0O2DwLfRxd53R44eGFHdep+yEcAT9nwB8AXQCyhty1YC6zE++FGYfAKhmJj9U9IZqyiwBvgVKJSRbZ7Y+x5mz56tHTp0UF/atm2rK1asSFGe2+THmN15TaDb6OzLHoFun2rg25hbsffdnn7BJS3l/kF7fjMmVn5zYIXdq/co9z1Cw9qqutG2T1O5D/wDs/RfhkyGLz5y5AgDBw5k3bp1hIaG8ssvvzBixAhq1KjBypUrufvuu5k/f35munQ4HI4Ci5v0CyiqGkdy5X5F4EZyWLlvmQQ8j4lsODozdvbv35/Dhw+zadMm1qxZQ2hoKBs3buTEiRMkJCRw9OhRxo0bl5kuHQ6Ho8DihHwFm97ARBF5w16/oKp/2ch5n4hIOczX/Xg1e/ovYUR4a+2LwRagS3oDiEhvIEFVPxWRQsDPInKDqi5O7z6Ao0eP8ssvv3D8+HE82QGLFi1KaGgoV155JYMHD86gB4fD4XB449ekLyL1MAKsMyISDjQFpqpx4XJcpKiXct+nPAGTGti3/BTwX+8yq85/HJOdrxRG3Henqq6390wVkZLWJRDMqsHzmHC8qeJR72/ZsoXKlSvTt29f1qxZQ4sWLZK+6idMmMDUqVNp2bIlY8eOpUIF350Kh8PhcPjil3rf/ge7JRAMzAcigEaq2jlXrXNcFPijzrft/Iq971Hvx8TE0L9/f95++20aNmzI22+/TalSpejWrRvlypVDRPjwww+JjY1lyJAhufuQXuRH1W9eE+g2OvuyR6DbB4Fv4wVV7wOr7N8nMfnmAVb7c6878v+Bn+p8TObBFzLqz6Pe37Nnj9apU0c9LF26VDt37qzebNmyRRs1aqR5SX5U/eY1gW6jsy97BLp9qoFvY26p9/3d008QkZ6YiG1dbZnfOdEd+R6POr8I5os/hZI/K7H3L7nkEqpVq0anTp3Ytm0bBw8epFOnTjz66KMsXryYoKAgTp06RYMGDXLoMRwOhyN/4696vy9wDfCKqm4RkUuBablnliOnEZG4jFslax8uIvP8bJ6uOj+rsfcBKlWqRExMDEWLFqV169a88sor7N27F1UlMTGRokWLUrZs2cx06XA4HAUWv770VTVaRIZgU6iq6hYy6XrlyJ/4qc73jr0PZgnq/oz6Pnr0KOvXr2fLli1J6n0wIXk9jBw5ku3bt+fQ0zgcDkf+xq8vfRHpionAtsBeh4nInNw0zJE72C/4SBH5QkQ2ich0L7/7G23ZKkwsfc89pUTkQxH5TURWi8ittnwcEKyqt4tIJ2AJcI33hG/bbFHV6hif/2NYwV5apKbev+KKK7j//vs5ccLsHDz77LPUqlWL6dOnuyx7DofD4Sf+qvd/B24AIlX1Clu2Xr3itjsCGxGJU9XS1uUyAmgE7AaWYQSaK4E/MP/OfwIzgJKq2iWdmPsKrMB8yb8HdFaToc973JIZtbHt/Fbv9+vXL+m+6dOnEx8fT9++fXPqp8qQ/Kj6zWsC3UZnX/YIdPsg8G280Or9hQSAdAAAIABJREFU5fbvaq+ytf7c647AODCJdADCge+9yt/FxNcPA5Z6ld+CScgDacTct3WtMRn3Hk5n7AzbeB+ZUe9v27bNqfd9CHT7VAPfRmdf9gh0+1QD38YLrd7fICJ3A4VE5HLgEeBnP+91BB5nvM7PkbG2wxNzPyaVuiaYOP7V07nfnzYpSEu9v2LFCp555hm2bt1KUFBQnmfZczgcjosVf9X7D2OWg88AnwJHgUdzyyjHBWETEGyjLwJ4J69PNea+iNQBnsAs9d8kIq18O/WnTXqkpt6/5557WLduHSVKlCAoKIiaNWtm9lkdDoejQJLhl75VZH+jqu2AZ3PfpKwhIpdg4sJfCRwB9gGPqurmbPZbHrhbVd+x19Uxseh7ZNNkT/+/AsUw6vYSwC5b1R94PKfG8eJGoIOIrMb41hcDjqjqabuv/o2InAR+xGTFA0gRc9+KOz8AVgENgfuAj0TkSlU9bZ9NbJvBqrpbRFK0SY+01PsiwurVq6lWrRp79uwhPDw8B34Wh8PhyP9kOOmr6jkRSRSRcqp6NC+Myix2cvka+FhV77JlzYCqwGavdoVV9Wwmuy+PmYDfAVDV3UCOTcSq2sra1gdoqarewWtybAtFVT2KkE3A/zzjiEg74DMRma6qC4AUkW40lZj7ln/6XDfxuU+926jq775tUiOj2Pv79u2jWrVqgNkC2LdvX0ZdOhwOhwP/1fsRmOXZ7/GKtqaqj+Seaf5j08OOUNU2qdSFY75UD2MmtFBgFEbQVgyYqKqTRKQ0RtVeARNZ7jlVjRCRz4FbgRjM80/ECNwai0hxjBCuJUao9riqLrET+C1ASaAe8LWqPpXBM/TBa9IXkWCvcfoA3YBSwOXAGEzo239jtlw6q+ohuzQ/EagMnAT+o6qb0hvHlr0IlFHVx1LrA9gDrAUuVdVEm1hnE1AX+J+18wubfGectfMM/H975x7v1ZT//+erji66kYrI6EIJkULMuBQKiTJjRv0YCt8ZMxOGCWOYGRkzDBky+BoTNQrlnhiDqc6XEJKTQhGVyiUiKd3U+/fHWp/69Olzzvmcy+ec3Tnv5+OxH2fvtdZe+7VXnbP2Wuu93m+Oi3Vs095Z3j9n6/3HHnuMp57a4jfolFNOYdKkSZlV5o2aaPVb1SRdo+urGEnXB8nXWN3W++dkO3K5tyoOgmHhLcXk9SR8qLSL1z8jdOgQOqEZQDvCrEfTmN6CsG1NhCBDc9Lq23xNWKu+N57vS7BqbwAMBj4kRJRrACwixKcv6R0GA7cX85zBUU8TQmf8NXBBzLuFsIwBMBnYJ573AKaU9pyYNgB4pqQ6CB9EveL5GcCoeD6GMPNRL77zoTG9aWzTrO1dUluUZr3fsWNH+/jjj83M7OOPP7ZU+aqiJlr9VjVJ1+j6KkbS9ZklX2O1Wu+b2b9yKZdgXrPgRRCgD3CgpNQUfTPC6HkJ8BdJRwObgD0IywMlcSTwdwAzmytpEdAx5k22uBwi6R2C3/nFFXiHqWb2DfCNpK+B1NB2dnyfxoStcQ+nrX/Xz7HulIFeSXVMIHT2U4GBxOWONDoBn5jZ6wBmtjLWWVx7L6AUdtttN9q0aUPnzp3p0KEDhxxyCI0aNWLFihV069aNDh06cMQRR9C/f/8cX9NxHKd2k1OnL2kBwRHLVphZ+0pXVD7epuR19vQAMCLsF382vUCc9m4JdDezDZIWEkbp5aWs2+LKUt+mtOtNse46BIO8rsVVEP3vZwt4czDwbmYdcWlkWCzzJOGjqDnQnRAiNxeytneu9OjRgzFjxrB06VIKCgqYPXs2TzzxBFdffTWzZ8/mww8/ZM6cOeWp2nEcp9aR65a9QwhW8YcCRwG3AePyJaocTAHqx3VhACQdKOmoLGWfBX4RY7sjqWNco24GLIsdfi/CyBzgG7ZYsWfyInBmqh5CbIJse9nzThxZL5D046hH0ZixRCQdQ5iC/2dmHZEmsf5VBM96Iwlr+BszqpoHtI7r+khqIqmA4tu7VJYsWUJRUREPP/wwRx99NE888QQFBQXUqVOHyZMnc8kllzB48GCaN2+eS3WO4zi1nlyn95dnJN0aXfP+ofIllR0zM0mnEXRdAawFFhJ8CeyRUXwUYb18ZrT6/5ywpn0/MEnSbMK689xY93JJL0maAzxDMHJLcSfwv/Ge74DBZrYufXtZFXNm1HM1wRhxPCHOfSZnxZkNI4zEXyK+LyFi3mhJ/wLWAF9A8L9PCKHbC5gvqb+ZTQQOAxpYMOS7BZgcZ4bWECz3RwGXAytju+zAFpe8WUlZ7//617/mxhtv5JtvvtmcN2rUKPr27UvDhg1p2rQp06dPL2MTOY7j1F5ytd7vlnZZhzDy/4WZlTqSdJJDdfnfT3t+d2A0cJRlbP/MtN6/9NJLmD59OpdccglFRUVMmDCB66+/nj/84Q8MHDiQ/fbbj/Hjx7N48WIuu+yySm2n0qiJVr9VTdI1ur6KkXR9kHyN1W29PzXteB64G+iUy71+JOegev3vtwDeAQ4oTWfHjh3tt7/9re2xxx6211572a677moNGza0vn37Wvv27S3FokWLrHPnzlbV1ESr36om6RpdX8VIuj6z5Gusbt/755nZh+kJktrleK8TSfO+l85PzWx2NcipMv/70avjeOBaM8vJ6u7666/n+uuvZ+PGjXTq1AkzY+LEiTRu3Jh27drRqFEjdtppJ/e77ziOUwZyNeR7JMc0pwTMrIeZdc04qqPDL468+N8nOOd5y8zGl1XQyJEj2WuvYFNZUFDAsGHDaNy4MXXr1uW9996jW7dupdTgOI7jpCix05e0r6QfAc0k/TDtGEzFtrM5CcSCP/yU//2ZwLK07D8RjPDekvQ28KdM3/oE//ujoqfCdIYBfSQVxePUXPQsWbKEp59+mquuuopDDz0UgOuuu47Zs2cza9YsrrzyStauLdWFv+M4jhMpbUq3E9CP4H/+lLT0bwjuWZ3tCIv+982sEChMSx+adl4m//uSXia448XM3pD0ZCw3Mq3YL4AL4nkz4BLCvv8SyWa9n2LDhg2MHTuWkSNHZrnTcRzHyUau1vtHmNkrVaDH2c6IMQIeM7NuMQLf+8Bhtu02T+Je/SnAjWY2KSMvJ+v9FCNGjKBBgwYMHZrN11B+qYlWv1VN0jW6voqRdH2QfI35st7P1ZDvTUm/Imzx2jx1a2bnlkmRU+Mws4WSlsc1/l2BN7N1+JGRBF/+20THMbO7CbtC6NSpk61cuZI33niDwYMHs3btWlauXMmoUaMYN24cw4cPp6CggIceeog6dXI1S6k8CgsLEx3ON+n6IPkaXV/FSLo+SL7GfOnL9S/mWGA34ATg/4A2hCl+x4HggGcwMAS4N1uBaAeyFzA8lwr/+Mc/svvuu9OsWTMaNGhAmzZtGDduHPvssw833ngjH374IW3atGHAgAGV9AqO4zg1n1w7/b3N7PfAagvBd04mRGBzHIDHgRMJbpq38bEfnfIMA84ys025VFi/fn2mTJnCrFmzGDVqFMuWLWP69OksWLCA3XbbjTp16vDtt99Wy0jfcRxneyXX6f0N8ecKSQcAnwKt8iPJ2d4ws/WSphKC9WT65Ifgra85MDXu+JthZueXVKekzetZ3//+92nfvj2S+O677wBYuXIle+21F2PGjKnEN3Ecx6nZ5DpMulvSzsDvCVbX7wA35k2Vs10g6VBJb0lqSPDK96P4UZhe5lpglpntbiF639OEcMClsnHjRrp27UqrVq3o3bs3PXpsmVx64oknOO6442jatGnlvZDjOE4NJyfrfccpDkl3AQOBBcBDZnZ9Rn5bcrDuz7Tef/jhhzbnrVq1it///vdcdNFFtGsXHEFeccUV9O3bl2OOOSZv71YcNdHqt6pJukbXVzGSrg+Sr7G6fe/vSnDC8ky83o/gmrfa/cn7Ub0HUI8Qye9VoG4xZZ4neOw7EXiktDo7duxomQwfPtxuuukmMzP7/PPPrXnz5rZmzZptylUFNdFnd1WTdI2ur2IkXZ9Z8jXmy/d+rtP7YwgGWinf6u8RwtY6zi5AY6AJxXtpLNW6P5PFixfTrVs3DjroIDp37syoUaPYd999GTx48GZf/IcffjhFRUWV8Q6O4zi1glw7/RZm9hCwCcDMviMEaXG2EyStKmP5npKeyqHoPwi2HvcDfy2mTInW/dn48ssv2bhxY/gyrVOHTZs20aJFCwB22203HnjgAYqKiujatWsu1TmO4zjkbr2/WtIuhNjpSDoc+LrkW5yajqSzgQ1m9kCMpPeypGPNbEp6OSvdun8bDjroIGbNmgXAt99+y5FHHkm0/Gf48OGceOKJlfoujuM4tYFcR/qXEqz2O0h6CbgPuDBvqpy8EUfwhZIekTRX0v1pkfNOjGkzgR+m3dNI0r2SXpP0pqT+Mas7YT0f4HhCuN7CtPvqSHpfUivgcOBeSfMltcxFa3HW+1dddRUHHnggl1xyCevWrSulFsdxHCdFidb7kr5nZh/F8wJCAB4B88xsQ7E3OolD0iozayypJzCR4FL5Y+Al4DJgBsGy/lhgPjAB2NHM+kn6C/COmY2TtBPwGsEwz4DXCfvw7wL6mtkHGc+9nWDd/y/C1P7PzexHWfTlZL3ftGlTmjdvzoYNG7j55pvZfffdOeeccyqrmXKiJlr9VjVJ1+j6KkbS9UHyNVaL9T4wM+380dKsAv1I7gGsij97As+npf8vcBbQFXghLf1U4Kl4PgOYAxTF4yOgc8z7PvAdcGExz90z9f8IGA/0K01radb7KaZOnWonn3zyNmXzTU20+q1qkq7R9VWMpOszS77G6rLeV9p5+1LKOtsP6XPiGyndtkPAj8ysazy+Z2bvxrwuwHK27OzYCjNbDHwm6VjgMOCZXAQWZ73/ySefAHDhhRfSu3dvDjjggFJqchzHcVKU1ulbMedOHpC0UVKRpLclzZL0m+jQpjx1HSLptjLcMhdoK6lDvB4U6ykkrNtfqEBbSfNj3l7AbwhT/SdJKi4ewyhgHPCw5WjIV5z1/plnnsnee+/N2LFjMTOuvvrqMryi4zhO7aa0DuUgSSslfQMcGM9XSvpG0sqqEFjLWBNH0vsDvYGTgD+WpyIzm2FmF5Wh/FrCmvrT0ZBvWVr2PcAOwFvAc0CraPx3DzDMzD4GzgNGScq2V/9Jwl7+0amEaO1fLCnr/bfeeovXX3+dVq1aIYnnn3+ePffck3fffZcGDRokek3OcRwnaZTY6ZtZXTNramZNzKwgnqeu3el5HjGzZYROeGgcYdeVdJOk16O/+58DSBov6eTUfZLGSDo9fZ+9pMbAw5JmA7cRO19JfQgW+BdJehiYZmb7mlk3M7vYzPrFateZ2c/NrAvQB/gorh9dDFwtqSjW+UMzWyvprGjpXyTpH4SZgFnADEk3S5oFHFFaG2Sz3r/99ts59dRTad26dYXb2HEcp7bhvvcTRMrCPiNtBWHXRH+glZldJ6k+wer+xwQDvAFmdo6kesAHQEdC6ONhFqzv/wrUN7Nfxzp3BuoCjwEnmdlqSVfEMtdmPL8w1jMjXrclGPgdIOnvwHQzuz8+uy7QlhCM6YdmtkHSK8A+wADgReAMC46eMt+9VOv9wYMHM2rUKG699Vbq1q3LSSedxDPP5GQiUKnURKvfqibpGl1fxUi6Pki+xnxZ7+fqnMepfvoQllhOj9fNCJ3pM8DI+CFwIsECf03KkU3keMK2OQDM7CtJ/QgxFF6KZesBr2R5bravwlTaK8BVktoQguq8L+k4wuzB67HehsDtZjZN0kbg0WwvZ2Z3A3cDdOrUyXr27LlV/syZM1mxYgWff/455513HgDr1q3j/PPPZ/78+dmqzBuFhYVk6ksSSdcHydfo+ipG0vVB8jXmS593+glGUnuCdf0yggX9hWa2jRvbOBo/ATiDsC0up+oJW/cGlVJuObBz2nVz4AsAC574XgVOBv4dlxwE/MvMrsxS19pcDfkWL15M//792bhxI+vXr2f16tXceeedHHfcccyYMYMddtiBunXr8u6775ZemeM4jgPk7pHPqWKi17q7CKNkIzi2+YWkHWJ+R0mNYvEJhGA2RwH/yVLd88Cv0ureGZgO/EDS3jGtkaSOWe4tBM7SlqmDc4Cp8Z72wIdmdhvB4c+BwGTg9OiFD0nNo5V/mSjJen/u3LnMnj0bgFGjRpW1asdxnFqLj/SrAEkDCEFnOpvZ3BKKNoxGcTsQHN40I0zZQ9j21haYGTvgzwnr5BAs6scCEy34uW9LMKx7O+ZfB9whaQ6wI2E0/jXQEpgd+/O1wIOSppvZfWma7gb2BWZJMoKjntQo/ifATyVtAD4F/mJmX0q6GngubjfcQPjgWJRjcwHF+97v27fv5jI33ngjS5YsKUu1juM4tRrv9KuGQcC0+HObLXiSCszsOzOrm36dXsbMNgG/iwcZeRsI0+7prE5Z35vZKsIIPfO5haQZ6WXDzNYT3Oxmy7sBuCFL+gTC7ENmepmsUjZu3Ej37t2ZP38+v/rVrzb73gfYsGEDY8eOZeTIkWWp0nEcp1bj1vt5Jm6Xmwf0AiaZWaeY3hP4E/AVYST9s/RrM+uY5i9/PDDWzJ6O944BniKMuscCqWn+oWb2crqFfSnaCtnaMv8agrveETHvTcKSQSPgbMIIvwswwcyujvecBVxEMAR8Ffhl+rq9pGYEX/2nmtk8SQ8CU8zsnxlacvK9365dOwBGjBhBgwYNGDo06/dIXqmJVr9VTdI1ur6KkXR9kHyN1eJ7349K8Xl/JnBPPH8Z6G5bfOCvBtplu45pKX/5pxGM4yB0rosJVvE7Ag1i+j5Ev8uEZYA5OWgrBA5Ju76G8BGQyvtrPL+YEJynNVAfWALsAnQGJgE7xHJ3AmdneU5vgqX/QOA/pekqzff+NddcY/3797eNGzduU64qqIk+u6uapGt0fRUj6frMkq8xX773fXo//wwCUnPQ4+P1G/H6NTNbkFY28zpFcdvymgG3S+pKsPLPZohXEZ6MP2cDb5vZJwCSPiQE0jmSbbfnLcusxMyel/Rj4A7goFweXJz1/g033MCIESNo2bIlgwYNYuzYsdSrV6/CL+o4jlMb8E4/j0hqTghV2yUawdUFTNJlscjqjFsyr4HgIreYbXmXAJ8ROtI6BGO8XLW1BQ7NSO4JfAOMiNepwDyb2DpIzybC/52StuelP6sOYVbgW8L2v1Kt74qz3v/d735Hy5Ytadq0KZMnT+ZHP/oRkyZNKq06x3EcB9+yl29OJ6zF72Vmbc1sT2ABYZ28rGTbltcM+MSCkd9PCR8VVUm27XnZojFeArwL/D9gdGrbYUkU53u/efPmLF26lKKiIiZNmsT69esr9YUcx3FqMt7p55dBhK166Twa08vKc8AxwH8tWNRDWEM/J/qy35diZgrKyR7AQ5LeAv4Am/fy3wt0I0TN2we4muBT/2tgIfBIeiWS7gAuA35jZi8SliFyGppn+t7v0KEDO+20EwUFYYKqTZs2LF26tBJe1XEcp3bg1vu1lGwW/hnW+x8TjArXSdrJzFZI+gvwjpmNk7QTwSr/YEIMgOuAA83syyzPeczMusVp/veBw8xseUa5Uq33zz33XG644Qbuv/9+AJYtW8YVV1zB6NGjqUpqotVvVZN0ja6vYiRdHyRfo/vedyqb4r72UulvAfdLegJ4Iqb1AU6VNCxeNwC+F8+fz+zwAcxsoaTlkg4GdgXezOzwY7lSfe+vXbuWdevWceSRR1JQUMArr7xCx44dq9x/dm312V2ZJF2j66sYSdcHydfovvedMiPpBOCvGckLzOw0tvWpD8HBT2r3wMnA0cAphKA6XQiGez8ys3kZz+lByUsLo4DBwG7AvbloLyoqYujQoXz55ZeYGevXr2fkyJH06tWL888/n+nTp7Ns2TK6d++eS3WO4zgOvqZfozGzZ82sa8ZxWsxbBXwi6VjYvNPgRGBanIbf08ymAlcQDAYbE/z/X5jywx9H77nweKz70FhHqSxfvpwvvviCgoICJLFixQrat2/PgAEDePzxx9mwYQPHHXcc996b0zeE4ziOg4/0aztnE3zy/y1eDzezD6J1/bjoB0DAbXFN/0/ArcBb8cNgAdCvtIdYiAcwFVhhOUbZO+6445g7d0uYgv79+7N06VImTpzIo48+yvHHH1+mF3Ucx3F8pF+rMbN3zKxX2izA/TF9g5kdaWZdzOwACz72MbM1ZvbzmL6/bfHtP8bMsvrDlXRo3AFwBPCApLcllegeOJOFCxfy5ptv0qNHD9577z1efPFFevTowTHHHMPrr79esUZwHMepRbj1vpNXJO0HvATMJ1j7LzGz67OU22y937Jly+4PPRSs99esWcPFF1/MWWedxdFHH82QIUM4+OCDufDCC5k7dy7XXnstDzzwAFsi/+afmmj1W9UkXaPrqxhJ1wfJ1+i+9/3Ybg9CvIBZhIA8dUsrn/K9v379euvTp4/dfPPNluKEE06wKVOmbL5u3769LVu2zKqSmuizu6pJukbXVzGSrs8s+Rrd976zPbMLwRBwB8I2v1KdCH300UccdthhrF+/niVLllC3bl0uvvhi6tSpw8CBA2ndujWNGjVizZo1tGjRIt/6HcdxagS+pu9UBf8Afg/cz7ZbCLMyc+ZMPvvsM9q0aUPdunW5/PLLueuuu7jvvvvo3bs33333HR999BHdu3ev0ql9x3Gc7Rkf6Tt5RdLZwAYze0BSXeBlScea2ZSS7hswYEBqaQAI1vsdOnSgRYsWjBs3DoDrr7+ejz76KJ/yHcdxahTe6Tt5xczuk7SjpKKY1Iww6i+x008n3Xof4KqrruK+++6jWbNmTJ06tfJFO47j1FDcet+pMuL+/ynAjWY2KSMvJ+v9dO6//37Wr1/PkCFDqkR/ippo9VvVJF2j66sYSdcHydfo1vt+bPcHISrg8NLKlWS9n86iRYts//33z5qXT2qi1W9Vk3SNrq9iJF2fWfI1uvW+s10jaTCwF5DViU8mxVnvn3/++UybNo169eoB0KFDh/yJdhzHqWF4p+/kHUndgWHAUWa2KZd7Utb7Xbp0AeDyyy+nfv36zJ07l4KC8N/222+/ZY899siXbMdxnBqHd/pOVTCUEMFvatxeN8PMzi/phuKs96dNm7Y57fHHH+eRRx7Jj2LHcZwaiO/Td/KKpGuBWWa2u5l1BZ4GZpeljkzr/RT33nsvJ510UuWJdRzHqeG49b6TVyS1BR4zs24xMt/7wGFmtjyjXJms98eNG8e8efO49tprq9w5T020+q1qkq7R9VWMpOuD5Gt0630/ttsDeB44GDgReKS08qVZ748ePdoOP/xwW716tVUHNdHqt6pJukbXVzGSrs8s+RrzZb3v0/tOVTAKGAwMAe7N5YYhQ4bQpEkTZsyYwaWXXgpAUVER++67LxdccAHffvstc+bMyZtgx3Gcmoh3+k5V8DhhlH8o8GwuN3Tv3p1169axevVqunbtSteuXRkyZAgrV66kZcuWfPPNNxx77LFccMEFeRXuOI5Tk/BOv4qQNECSSdq3jPe9XM7ntZWUdSgs6QRJRfFYJWlePL9P0gXRX36lYWbrganAQ2a2MZd7hg4dyoIFC9h7770pKiqiqKiIVq1accstt7B48WL+/Oc/c+qpp3LXXXdVplTHcZwajW/ZqzoGAdPizz9mZkoqMLPvMq/N7PuVLcTMniWOuCUVAsPMbEZlPydFNOA7HPhxReq59dZbOeGEExg2bBibNm3i5ZfL9T3kOI5Ta3Hr/SpAUmNgHtALmGRmnWJ6T+BPwFfAvgTr9c3XZtZR0iozayxpPDDWzJ6O944BngJmAGOBRvFxQ83s5Wg1/5SZHVCKtkLSOn1J1wCrzGxEzHsTOCrWfzZwJdAFmGBmV8d7zgIuAuoBrwK/TI3oJe0HPBfzPgbqAgeY2VYm99ms9z/99FOuvPJKRo8eDcBtt93GQQcdxDHHHMPUqVN56qmnuPnmm0tr/kqnJlr9VjVJ1+j6KkbS9UHyNbr1/nZ8AGcC98Tzl4Hu8bwnsBpol+06pq2KP08D/hXP6wGLgYbAjkCDmL4P0XoTaAvMyUFbIXBI2vU1hI+AVN5f4/nFhE67NVAfWALsAnQGJgE7xHJ3AmeX8LybgJtK0pSy3l+wYMFWvvWbNm1qmzZtMjOzTZs2WZMmTaw6qIlWv1VN0jW6voqRdH1mydfovve3bwYBI+P5+Hj9Rrx+zcwWpJXNvE7xDDBSUn2CUdwLZrZGUjPgdkldgY1Ax0rW/mT8ORt428w+AZD0IbAncCTQHXg97pdvCCzLVpGkM4BuQJ/SHnruuecyceJEVq9evTmtSZMmtGvXjo8++ojbb7+dffbZp/xv5TiOUwtxQ748I6k5cCwwStJC4DLgJ9riUWZ1xi2Z1wCY2VrCyPsE4AxgQsy6BPgMOAg4BGgYY9c/B3SQ9Ju4pl4emgAXxvNNwLq0vE0EmxARZiC6xqOTmV2TWZGkAwizCAMtB2O+RYsWIYl169bRpk0b7rnnHm644QYaNWrEjjvuyJ133sndd99dztdyHMepnfhIP/+cTliL/3kqQdL/EdbJy8oE4HxC5z44pjUDlpjZJknnAJhZ17im/wxwEtCULMaDOfANMKKUMpOBiZJuMbNl8SOniZktShWQtBPwIGHa//NcHjx58mQWLlxIv379ttqPf9ZZZ9GzZ09GjBhB9+7dy/xCjuM4tRkf6eefQYR96uk8GtPLynPAMcB/LWyDg7CGfo6kWQRjwHQ2EozjhipQV9JNkl6X9Jak1IfInyWdnHbfGZJOB3YCbolpDYGDJM2W9BbQIqa3AQxYKGkF4SOgdaoiSQXATKA98E9Jn0n6tBzv7jiO41QQH+nnGTPrlSXttrTLwrT0wvTrmNY47XwDIVpdev77wIGpa0m/iukLgQNiWl2gFdAf+NrMDo22AS8RttH9ooXEAAAYt0lEQVR1BX5CCIbzF+C8eP4FISQuBCPDR83s17HOnQmW+H8DDjaz1ZKuAOqb2fQ0fd9JOgV4JNZ1E7B15BzHcRynSvBOv3bRBzgwjuIhLA3sQ/FGgun3Hg8MTF2Y2VeS+gH7AS/FsvWAVzIfamZvSxpL2GJ4RNosxWYytuxRWFjIp59+yurVqyksLNyq7IoVK3jjjTdYtWpVedqgwqxatWobTUki6fog+RpdX8VIuj5Ivsa86SvNvN+P7esgbvFLuz4H+A4oAlYAHwCPZ7nvPuBU4AHg1JjWk7DXH8Jug30y7jkFeDBHXQ8CnwInlFY2tWXv6quvtvr169v+++9vAwcOtDVr1piZ2THHHGOvv/66VRc1catPVZN0ja6vYiRdn1nyNXrAHSdXGkWXum9HN7y/A/5sIZb95YStdz8BkNRRUsqpzwRCQJyjgP/EtE7A/vH8eeBXqYfE6f3pwA8k7R3TGknaZsugpB8SliWOBv4eDftKZMCAAVx//fVs3LiRFStWMH/+fC677DLatGnDK6+8wsknn8wJJ5xQtpZxHMep5XinX7NRPFKMAt4BZsYPgn+wZYknm5HgPODteH4dsLOkOdFosJcFS/zBwIPRuO8VMowJJbUAbgDON7P3gNvZ4rOgWO644w5at27NZ599xsKFC9l111055ZRTWLJkCevWreOzzz7j2Wdzit3jOI7jRHxNv+axOo7qAZDUnuA45xrCB8AOwFqCV73xZvZ1movf5vGeMYT19y8y6t5EsNRP/5goICwf1CUsHUzJuKce8C0wKa77dyFY8pfIHnvswbBhw/je975Hw4YN6dOnD336lOrTx3EcxykBH+nXcMzsQ0KH3Ipglf+1mR1KCHP7P5LaEab2U1P+9YDjCNb76fw+3tvFzA4EpsRR/NXA8WbWjRAH4NKM539s0XEP8E/CDoBFlMJXX33FxIkTWbBgAR9//DGrV69m3Lhx5W4Hx3Ecx0f6tY1qsd4HkPQD4H8Ibnuz5W9lvT9y5EgaNGjA22+H1YXOnTvz8MMP06ZNm3K9eGVSa61+K5Gka3R9FSPp+iD5Gt1634+cDra13m8PLCdMxz9KMdbz5NF6n+CsZy7QMZd36Nixo02YMMHq169vXbp0sQMPPNAKCgrstNNOsyRQE61+q5qka3R9FSPp+sySr9Gt950yI6klcBdwe/wP8SzwC0k7xPzSrPfTKZf1fnzWw8AVFgz5cuInP/kJV1xxBevXr2fjxo0UFBRwww035Hq74ziOkwXv9KsASQMkmaRMN7ml3fdyOR7XUNI7ktYC/yVY5Q+PeaOANcBKSWsIlvxFMUDPQML0/wLL4jyHclrvA98nxAoYHrcSFknaPZcXGT58OHPnzuVvf/sb3bp1o2PHyg4g6DiOU7vwNf2qYRAwLf7cJvCNpAIz+y7z2sy+X9YHmVndGGznKTM7KCNvE2k+/yUVAsPMbEZM+mlG+UKiW2AzW0Vw9JP5vCkEo8Di9Pwf0KCs75HO+PHjGTSoPKEKHMdxnHR8pJ9nJDUmGK+dR5ohnKSekl6U9CTwTuZ1LLMq/hyfHhBH0hhJp0tqG++ZGY8yfyRkaL1G0rB4XijpFkkzJL0r6VBJj0l6X9J1afecJem1OIL/R/Tzn17nsZKeSLvuLSkzAFGxrF+/nieffJIf//jHFXk1x3EcBx/pVwX9gf+Y2XuSlkvqbmZvxLxuwAFmtkBSz/TrjDpSW+qeTttS9wuCcV5vM1sraR+Cq9tDKlH7ejM7RNLFwESgO/Al8IGkWwjbAM8AfmBmGyTdCZxJMApMMRW4U1LLuBwwBLg380HZfO8DTJs2jXbt2vHuu+/y7rvvVuKrlZ9aa/VbiSRdo+urGEnXB8nXmC993unnn0Fs8UA3Pl6nOv3XMjr4zOsUxW2pawbcLqkrIYxuZS96Pxl/zgbeNrNPACR9COxJmMHoTnD+AyH87rL0CszMYrCdsySNBo4Azs58kJndDdwN0KlTJ+vZsydt27ZlxYoVNG3alGHDhjFjxozM26qFwsJCevbsWd0yiiXp+iD5Gl1fxUi6Pki+xnzp804/j0hqDhwLdJFkBCc5JumyWGR1xi2Z1wDEkXwhcAJhZD0+Zl0CfAYcRFiqWStpI8F9bvtocHcfcEtczy8r7SWdATwGrEtL30T4vyPgX2Z2ZSn1jAYmETwBPpxuv1ASmzZtok6dOsyePZtmzZqVXb3jOI6zFb6mn19OJ7i33cvM2prZnsACwra4spJtS10z4JPYof+U8FGxBugLzAd6AyeRxXgwR941s4tKyJ8MnC6pFYSPHEl7ZRYys4+Bjwne+0bn+vA6derw3nvveYfvOI5TSXinn18GAZlGa4+SZkFfBrIFxLkTOCeO6PclY6bAzJYR1smHKlBX0k2SXo9b7FpDdkNBoCXQXdJTMbmupNGSZhPsBo41s3eAh4AFcQvgB2T41ZfUQdJM4H5gMfBdvC4VSfTp04fu3btz991359ZKjuM4TrH49H4eMbNeWdJuS7ssTEsvTL+OaY3TzjcQwtOm578PHJiWdIWkVWa2EDgglvkwWtS3IhgVfm1mh0b7gJcI3vomAD8xs3OioeAHBPuAHsApZlYo6SSgvpl1geCcJ/rePwpoZWarJV0Rr6emafxA0tfAAILv/SHkONqfNm0ae+yxB8uWLaN3797su+++HH300bnc6jiO42RBwVGbU1OInX7jjLQVQCfgDsJHwrcxqxnwc+AF4D2CH/4TCR8AZ8YdBcPMrJ+kN4CB8UMjVW8/YAywJCbVA14xs/Mynr8g5u0NzAEOM7PlGWXSrfe7P/TQQ1u915gxY2jYsCFnnHFGmdukslm1ahWNGzcuvWA1kXR9kHyNrq9iJF0fJF9jefT16tXrDTMreQdXaX56/di+DtJ87xMM/94hhL4tAlYALxVzXz597zcgfFT0Bx4qrXzHjh1t1apVtnLlSjMzW7VqlR1xxBH2zDPPWBKoiT67q5qka3R9FSPp+sySr9F97zvlYSZhFP5nC6FtLwc+r0rf+xB2HxD8/v8vOU7tv/nmm7Ru3ZoGDRqwyy67sPPOO3PiiSfm9NKO4zhOdrzTLyeSdosGcB9IekPSv7N1eOWodydJv0y73l3SI2WoomH0jvc22X3vH0rwvb+eMNX+ioLv/ZXAyWxtKJhOeX3vp7ifsNXvuVxeokOHDrzwwgusXbuWzz//nPfff5933nknl1sdx3GcYnBDvnKg4InmccIe9YEx7SBgV8I0dqrcVj71c2Qn4JcEy3wsbHc7PdebzaxuCXmbgD2itsHAIWY2NK1IvYzyhVTQ934aRwKjzWxjDmVp3bo1rVu3BqBJkyZ07tyZpUuXst9+++Vyu+M4jpMFH+mXj17ABjO7K5VgZrPM7MUsPvW32iYn6ecQfPJLmhx95s+W1D9WdQPQIY7Wb1Lwrz8n3tMgtW1O0puSesX0wdEv/n+ib/wby/pCGc8ZLOkJSc9LWihpqKRL4zOnR6dDqe14/4kzHS8qI4qgpDpRz9MEL3y3SZqvEPI3ZxYuXMibb75Jjx49yvpajuM4Tho+0i8fB7DFlW420n3q/4yMbXKSniPsWT/NzFbGrW/T44fCb+O9XSF0xmn1/org2bZL7GCfS1tS6AocTPCcN0/S381scQXf8WCCEd584AozO1jB5/7ZwK0Et7kXmNn7knoQZieOTVVgZpskjYvvf6ukPsCsuCywFSrG9/6aNWu4+OKLOf/885k5M6ft/XmntvrsrkySrtH1VYyk64Pka3Tf+9sX6T70+wAHSkpN0TcjbI1bAvxF0tGEte49CMsDJXEk8HcAM5sraRFb/O1PNrOvASS9A+xF+LAoL1PN7Bvgm7jPflJMnx3fpzHwfeDhsNoBQP0s9dxLCNZzK3AuxRjyWRbf+xs2bKBfv35ccMEFXHrppRV4lcqltvrsrkySrtH1VYyk64Pka3Tf+8nibUpeZ0/3jCfgQjN7Nr1AXFNvCXS3EKFuIRWLO5/uG38jFf+3zfS1vy7tvICwNLQiNSNRHGa2WNJnko4FDiNE4SuVIUOG8OCDD9KoUSOeffbZ0m9wHMdxSsXX9MvHFKB+nJIGQNKBkrL51H8W+EWWbXLNgGWxw+9FGJkDfAM0Kea5LxI7zTit/z1CcJ0yoRCU51pgkKRZkn5D+DjJGTNbSXC/e4Wk2xQ4qJjio4BxhGA7ORnyde/enXXr1rF69Wq6du1K165d+fe//10WiY7jOE4G3umXg+gE4TTg+Lhl723geuDTLMVHERzkzIyGcv8gjJTvBw5R8GV/NjA31r2csO4/R9JNGXXdCdSJ90wABpvZOsrOGuAPwINsCcpzcTnqOZNg1HgMYfajfzHlngQaU4ZgO0OHDmXBggXsvffeFBUVUVRURN++fcsh0XEcx0nh0/vlJG6l+0mWrPfZ2qf+JuB38cjkiGLq/n8ZSSk/+msJDnQyy48huMNNXfcrSXvmPXHG4nWghYKf/v2BHnHv/R1m1jb6JBibuk8hKM9ThN0GKVe9jSWNJgTkMWC4mT1KNNADHpD0ATAkbgF0HMdxqhD3vV8LUcn++fsTAuhcpy1BeX5M2B0wwLIH5Ul1+n8lBOX5daxzZ+ASgifAfmb2X4WgPPXN7NqM52/je//TTz/lyiuvZPTonCcIqoSa6LO7qkm6RtdXMZKuD5KvMV++932kX4OR9CrbWtT/tJTbittt8AwwMn4InAi8YGZr0iz3AY4HBqYuzOwrSa8Bq4ARsWw9gue+rchmvb9w4UIaNWqUOAvb2mr1W5kkXaPrqxhJ1wfJ1+jW+06ZMbOs3mwyOmoktSdY/C+jmN0GsVwhIYjPGcD4HGUIeN7MBuUsHDj33HOZOHEiq1evLr2w4ziOkxNuyFfLid7x7gJujwaKxe02gDwG5clk0aJFSGLdunW0adOGe+65p/wv6TiO4wDe6ZcbJTfgTmn1vwo0krRe0nfARwTDvefic4rbbQAhWM4x5C8oz2YmT57MjBkz2H///VmyZAnnnXdeRV7bcRzHwaf3y4WU3IA7pZGa8lf2gDsvxZ9ZdxuY2QageUZaIZUXlMdxHMfJI269Xw6id7lrzOzoLHk9gT8BXxFGtJ0J29p6Eozq7jCzf0Q3thOBnYEdgKvNbKKk8QQL+nmE6fI7gKfM7ABJDQgx6Q8BvgMuNbOpsQM/FdgR6AA8bmaXl/IOg0nr9KOP/9RzBgMDgEYEI74RBAO8nxI88/U1sy8ldYj6WgLfAv9jZnMznjMSWG5m10o6AbgK6Bm3MqaXc+v9SiLp+iD5Gl1fxUi6Pki+xnxZ72NmfpTxAC4CbikmryfBDW+7eP0zQocOodOfAbQjzLI0jektCEFtBLQF5qTVt/ka+A1wbzzflzA134Awff4hwdK+AbAI2LOUdxhMWMfP9pzBUU8TQof+NSGwDsAtwK/j+WRgn3jeA5iS5Tk7Ehz39CJ8yHQorX07duxoZmYLFiyw/fff35LG1KlTq1tCiSRdn1nyNbq+ipF0fWbJ11gefcAMK+Xvq0/v5wcPuBMxs28l/Q/wAnCJmX1QAU2O4zhOBXBDvvLxNtC9hPxsAXe6xqOdmT1HcGGbCrjTFfiM7TTgTtrRuZi6ugDLgd1zffigQYM44ogjmDdvnlvvO47jVBLe6ZeP7TrgTmVgWwLu/DjqyRpwR9JehGWJg4GTJGX1HZDJgw8+yCeffMKGDRvcet9xHKeS8E6/HMS1k+054E5lcSZwXtyet03AnbjL4R6Cm96PgfOAUdEg0XEcx6li3HrfSRySvqGaZjBypAXwRXWLKIGk64Pka3R9FSPp+iD5Gsujby8za1lSATfkc5LIPCtt20k1ImmG66sYSdfo+ipG0vVB8jXmS593+jWY4gLumNns6tDjOI7jVC/e6ddgrJiAO47jOE7txA35nCRyd3ULKAXXV3GSrtH1VYyk64Pka8yLPjfkcxzHcZxago/0HcdxHKeW4J2+4ziO49QSvNN3EoOkEyXNkzRf0m+r8Ll7Spoq6R1Jb0u6OKY3l/S8pPfjz51juiTdFnW+JalbWl3nxPLvS9omzHAFddaV9Kakp+J1O0mvRh0TJNWL6fXj9fyY3zatjitj+rwY9bAy9e0k6RFJcyW9K+mIJLWhpEviv+8cSQ9KalDdbSjpXknLouOuVFqltZmk7pJmx3tuiw6zKqrvpvhv/JakxyXtlJaXtW2K+90urv0roi8t7zeSTFKLeJ2I9ovpF8Y2fFvSjWnp+W+/0iLy+OFHVRxAXeADoD0hjO8sYL8qenZroFs8bwK8B+wH3Aj8Nqb/FvhrPO8LPEOIq3A48GpMb06IdticEDL5Q2DnStR5KfAAIQQywEPAwHh+F/CLeP5L4K54PhCYEM/3i+1anxDp8QOgbiXq+xdwfjyvB+yUlDYkBLRaADRMa7vB1d2GwNFAN7aOrFlpbQa8Fssq3ntSJejrAxTE87+m6cvaNpTwu11c+1dEX0zfk+ACfRHQImHt1wv4L1A/XreqyvbL+x9UP/zI5QCOAJ5Nu74SuLKatEwEehO8AraOaa0JToMguFIelFZ+XswfBPwjLX2rchXU1IYQyvhY4Kn4R+iLtD++m9sv/rE7Ip4XxHLKbNP0cpWgrxmhU1VGeiLakNDpL45/2AtiG56QhDZk23DaldJmMW9uWvpW5cqrLyPvNOD+eJ61bSjmd7uk/8MV1Qc8AhwELGRLp5+I9iN01MdnKVcl7efT+05SSP1RTrEkplUpcRr3YOBVYFcz+yRmfcqW0MfFac3nO9wKXE6IcgiwCyHK4XdZnrVZR8z/OpbPp752wOfAaIUliFEKgaUS0YZmthQYAXwEfEJokzdIVhumqKw22yOe51PruYQRcHn0lfR/uNxI6g8sNbNZGVlJab+OwFFxWv7/JB1aTn3laj/v9B0nIqkx8CjwawtRBDdj4VO6Wva3SupHiMj4RnU8P0cKCNOY/2tmBxPCS29ll1HNbbgzISBUO0KI50bAidWhpSxUZ5uVhqSrgO8IwcMSgaQdgd8Bf6huLSVQQJhxOhy4DHiorLYCFcE7fScpLCWsw6VoE9OqBIXQx48Spiofi8mfSWod81sDy0rRmq93+AFwqqSFwHjCFP9IYCdJKa+a6c/arCPmNwOW51EfhFHGEjN7NV4/QvgISEobHg8sMLPPzWwD8BihXZPUhikqq82WxvNK1yppMNAPODN+mJRH33KKb//y0oHwYTcr/r60IUQ43a0c+vLVfkuAxyzwGmH2rkU59JWv/cq6PuGHH/k4CF+/HxJ+YVPGKvtX0bMF3AfcmpF+E1sbVN0Yz09ma4Og12J6c8K69s7xWAA0r2StPdliyPcwWxvx/DKe/4qtjdAeiuf7s7Wh0IdUriHfi0CneH5NbL9EtCHQgxD+ecf4zH8BFyahDdl2zbfS2oxtDdH6VoK+EwnhwltmlMvaNpTwu11c+1dEX0beQras6Sel/S4Aro3nHQlT96qq9qu0P0Z++FHRg2Bd+x7BUvWqKnzukYQp1LeAonj0JayZTQbeJ1jbpv4QCLgj6pwNHJJW17nA/HgMyYPWnmzp9NvHP0rz4y9/yhq4QbyeH/Pbp91/VdQ9jzJaIuegrSswI7bjE/EPaGLaEBgOzAXmAGPjH9dqbUPgQYKNwQbCCPC8ymwz4JD4vh8At5NhaFlOffMJHVXqd+Wu0tqGYn63i2v/iujLyF/Ilk4/Ke1XDxgX650JHFuV7edueB3HcRynluBr+o7jOI5TS/BO33Ecx3FqCd7pO47jOE4twTt9x3Ecx6kleKfvOI7jOLWEgtKLOI7j1AwkbSRs10oxwMwWVpMcx6lyfMue4zi1BkmrzKxxFT6vwLb4Rnecasen9x3HcSKSWkt6QVKRpDmSjorpJ0qaKWmWpMkxrbmkJ2Js9umSDozp10gaK+klYKyklpIelfR6PH5Qja/o1HJ8et9xnNpEQ0lF8XyBmZ2Wkf//COFJ/yypLrCjpJbAP4GjzWyBpOax7HDgTTMbIOlYgivnrjFvP+BIM1sj6QHgFjObJul7hJCpnfP4jo5TLN7pO45Tm1hjZl1LyH8duDcGYHrCzIok9QReMLMFAGb2ZSx7JPCjmDZF0i6Smsa8J81sTTw/HtgvLZBaU0mNzWxV5b2W4+SGd/qO4zgRM3tB0tGE4CxjJP0N+KocVa1OO68DHG5maytDo+NUBF/TdxzHiUjaC/jMzP4JjCKEB54OHC2pXSyTmt5/ETgzpvUEvjCzlVmqfY4Q0S/1jJJmGhwnr/hI33EcZws9gcskbQBWAWeb2eeSfgY8JqkOIb59b0L44HslvQV8C5xTTJ0XAXfEcgXAC4Twqo5T5fiWPcdxHMepJfj0vuM4juPUErzTdxzHcZxagnf6juM4jlNL8E7fcRzHcWoJ3uk7juM4Ti3BO33HcRzHqSV4p+84juM4tYT/D0I7r8keWRqOAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "plot_importance(xgb_model, max_num_features=55)\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "bC-aXahDBnVy",
        "outputId": "91bf599f-648e-4d83-ff17-ad7cef22b99c"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "plot_importance(xgb_model, max_num_features=20, importance_type=\"weight\")\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "HkDgEcSZCG9L",
        "outputId": "912693a0-82ce-4e81-de34-9be12578a96c"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "plot_importance(xgb_model, max_num_features=20, importance_type=\"gain\")\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "nHppZGpiCKaH",
        "outputId": "d81dd945-4bdd-46f7-c072-8e464c7fa888"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "plot_importance(xgb_model, max_num_features=20, importance_type=\"cover\")\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "E0jIdc3HCcCw"
      },
      "outputs": [],
      "source": [
        "y_pred = xgb_model.predict(X_validation)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "7EJwsH06C96L",
        "outputId": "edec4ba8-18de-4c52-8d51-149f825d6058"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "0.9063146650389134"
            ]
          },
          "execution_count": 49,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "sum(y_pred==y_validation) / len(y_pred)"
      ]
    }
  ],
  "metadata": {
    "colab": {
      "collapsed_sections": [
        "K-tlPjlCKvhW",
        "suCyhBPxO5X9",
        "MmcEiCpni0l0",
        "fXmgdiKO_FkA"
      ],
      "name": "PROJECT2_PART2_322691718_206684011.ipynb",
      "provenance": []
    },
    "gpuClass": "standard",
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}