4. Modelling.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Modelling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:21.930707Z",
     "start_time": "2021-01-21T08:04:19.823686Z"
    }
   },
   "outputs": [],
   "source": [
    "# Import libraries\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import pickle\n",
    "\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.model_selection import train_test_split, GridSearchCV, cross_val_score\n",
    "from sklearn.pipeline import Pipeline\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.metrics import (confusion_matrix, classification_report, plot_roc_curve, roc_auc_score, \n",
    "accuracy_score, precision_score, recall_score, f1_score, auc, precision_recall_curve, average_precision_score)\n",
    "from sklearn.ensemble import (RandomForestClassifier, ExtraTreesClassifier, GradientBoostingClassifier, \n",
    "                              AdaBoostClassifier)\n",
    "from sklearn.svm import SVC\n",
    "\n",
    "import xgboost as xgb\n",
    "from xgboost import XGBClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:21.946700Z",
     "start_time": "2021-01-21T08:04:21.932696Z"
    }
   },
   "outputs": [],
   "source": [
    "pd.set_option('display.max_rows', 400)\n",
    "pd.set_option('max_colwidth', 120)\n",
    "pd.set_option('display.max_columns', None)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:22.138749Z",
     "start_time": "2021-01-21T08:04:21.950702Z"
    }
   },
   "outputs": [],
   "source": [
    "train = pd.read_csv('./datasets/train_processed.csv')\n",
    "train = train._get_numeric_data().drop(columns='n_comments', errors='ignore')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:22.217766Z",
     "start_time": "2021-01-21T08:04:22.138749Z"
    }
   },
   "outputs": [],
   "source": [
    "test = pd.read_csv('./datasets/test_processed.csv')\n",
    "test = test._get_numeric_data()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "full = pd.read_csv('./datasets/full_processed.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In this notebook, we'll carry out of the following steps\n",
    "\n",
    "- Run a Train-Test-Split on our data\n",
    "- Run data through a pipeline and fit various models to our training data\n",
    "- Select the best model and tune hyper-parameters\n",
    "- Evaluate model based on accuracy and ROC-AUC\n",
    "\n",
    "We'll look to test a range of classification techniques including Logistic Regression, Random Forest, Boosting, Multinomial Naive Bayes classification and Support Vector Machine (SVM) classification.\n",
    "\n",
    "Accuracy and ROC-AUC will be our main metrics -- we want to minimize both false negatives and false positives, but we also want to know how good our models are at separating our positive and negative class.\n",
    "\n",
    "It's worth noting that in previous iterations of my modelling I tried modelling with purely text-based features (headline + abstract). This returned an unsatisfactory accuracy and ROC-AUC score."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Train Test Split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:22.233770Z",
     "start_time": "2021-01-21T08:04:22.219754Z"
    }
   },
   "outputs": [],
   "source": [
    "X = train.drop(columns=['is_popular'])\n",
    "y = train['is_popular']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:22.249762Z",
     "start_time": "2021-01-21T08:04:22.234757Z"
    }
   },
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=42)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Instantiate Models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:22.265764Z",
     "start_time": "2021-01-21T08:04:22.252762Z"
    }
   },
   "outputs": [],
   "source": [
    "feature_names = X_train.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:22.281769Z",
     "start_time": "2021-01-21T08:04:22.267764Z"
    }
   },
   "outputs": [],
   "source": [
    "models = {'lr': LogisticRegression(max_iter=5_000, random_state=42, solver='saga'),\n",
    "          'rf': RandomForestClassifier(random_state=42),\n",
    "          'gb': GradientBoostingClassifier(random_state=42),\n",
    "          'et': ExtraTreesClassifier(random_state=42),\n",
    "          'ada': AdaBoostClassifier(random_state=42),\n",
    "          'svc': SVC(random_state=42, probability=True),\n",
    "          'xgb': xgb.XGBClassifier(seed=42, use_label_encoder=False, eval_metric='auc'),\n",
    "        }"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Model Function Setup"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To get various feature importance scores from XGBoost, I created a custom transformer that scales all features with StandardScaler and returns it as a dataframe (instead of an array)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:22.297772Z",
     "start_time": "2021-01-21T08:04:22.283770Z"
    }
   },
   "outputs": [],
   "source": [
    "class CustomTransformer():\n",
    "    def __init__(self, func):\n",
    "        self.func = func\n",
    "\n",
    "    def transform(self, input_df, **transform_params):\n",
    "        return self.func(input_df)\n",
    "\n",
    "    def fit(self, X, y=None, **fit_params):\n",
    "        return self\n",
    "\n",
    "# This function takes a dataframe as input and returns a scaled version\n",
    "def scale_df(input_df):\n",
    "    ss = StandardScaler()\n",
    "    feature_names = input_df.columns\n",
    "    input_df = ss.fit_transform(input_df)\n",
    "    input_df = pd.DataFrame(input_df, columns=feature_names)\n",
    "    return input_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:14:55.097192Z",
     "start_time": "2021-01-21T08:14:55.087199Z"
    }
   },
   "outputs": [],
   "source": [
    "# Instantiate lists to store results\n",
    "init_list = []\n",
    "gs_list = []\n",
    "\n",
    "# Function to run model -- input scaler and model\n",
    "def run_model(mod, mod_params={}, grid_search=False):\n",
    "    \n",
    "    # Initial dictionary to hold model results\n",
    "    results = {}\n",
    "    \n",
    "    pipe = Pipeline([\n",
    "            ('ss', CustomTransformer(scale_df)),\n",
    "            (mod, models[mod])\n",
    "            ])\n",
    "    \n",
    "    if grid_search:\n",
    "        # Instantiate list to store gridsearch results\n",
    "        gs = GridSearchCV(pipe, param_grid=mod_params, cv=3, verbose=1, scoring='roc_auc', n_jobs=-1)\n",
    "        gs.fit(X_train, y_train)\n",
    "        pipe = gs.best_estimator_\n",
    "        \n",
    "    else:\n",
    "        pipe.fit(X_train, y_train)\n",
    "    \n",
    "    # Retrieve metrics\n",
    "    predictions = pipe.predict(X_test)\n",
    "    tn, fp, fn, tp = confusion_matrix(y_test, predictions).ravel()\n",
    "    y_test_pred_prob = pipe.predict_proba(X_test)[:,1]\n",
    "    y_train_pred_prob = pipe.predict_proba(X_train)[:,1]\n",
    "    \n",
    "    results['model'] = mod\n",
    "    results['train_auc'] = roc_auc_score(y_train, y_train_pred_prob)\n",
    "    results['test_auc'] = roc_auc_score(y_test, y_test_pred_prob)\n",
    "    results['precision'] = precision_score(y_test, predictions)\n",
    "    results['specificity'] = tn / (tn + fp)\n",
    "    results['recall'] = recall_score(y_test, predictions)\n",
    "    results['f_score'] = f1_score(y_test, predictions)\n",
    "    \n",
    "    if grid_search:\n",
    "        gs_list.append(results)\n",
    "        print('### BEST PARAMS ###')\n",
    "        display(pipe[1])\n",
    "        \n",
    "    else:\n",
    "        init_list.append(results)\n",
    "    \n",
    "    print('### METRICS ###')\n",
    "    display(results)\n",
    "    \n",
    "    print(f\"True Negatives: {tn}\")\n",
    "    print(f\"False Positives: {fp}\")\n",
    "    print(f\"False Negatives: {fn}\")\n",
    "    print(f\"True Positives: {tp}\")\n",
    "    \n",
    "    return pipe"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Initial Run"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:22.808901Z",
     "start_time": "2021-01-21T08:04:22.315776Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "### METRICS ###\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'model': 'lr',\n",
       " 'train_auc': 0.8441526452636806,\n",
       " 'test_auc': 0.848308634620144,\n",
       " 'precision': 0.7834254143646409,\n",
       " 'specificity': 0.7998979070954568,\n",
       " 'recall': 0.7546567323044172,\n",
       " 'f_score': 0.7687720249390079}"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "True Negatives: 1567\n",
      "False Positives: 392\n",
      "False Negatives: 461\n",
      "True Positives: 1418\n"
     ]
    }
   ],
   "source": [
    "lr = run_model('lr')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:24.342235Z",
     "start_time": "2021-01-21T08:04:22.809895Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "### METRICS ###\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'model': 'rf',\n",
       " 'train_auc': 1.0,\n",
       " 'test_auc': 0.8665311857419843,\n",
       " 'precision': 0.7942318358291736,\n",
       " 'specificity': 0.810617662072486,\n",
       " 'recall': 0.7621075039914849,\n",
       " 'f_score': 0.7778381314502987}"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "True Negatives: 1588\n",
      "False Positives: 371\n",
      "False Negatives: 447\n",
      "True Positives: 1432\n"
     ]
    }
   ],
   "source": [
    "rf = run_model('rf')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:25.649551Z",
     "start_time": "2021-01-21T08:04:24.343236Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "### METRICS ###\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'model': 'et',\n",
       " 'train_auc': 1.0,\n",
       " 'test_auc': 0.8685890722558594,\n",
       " 'precision': 0.7875275938189845,\n",
       " 'specificity': 0.8034711587544666,\n",
       " 'recall': 0.7594465141032464,\n",
       " 'f_score': 0.7732321863993498}"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "True Negatives: 1574\n",
      "False Positives: 385\n",
      "False Negatives: 452\n",
      "True Positives: 1427\n"
     ]
    }
   ],
   "source": [
    "et = run_model('et')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:27.673993Z",
     "start_time": "2021-01-21T08:04:25.651540Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "### METRICS ###\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'model': 'gb',\n",
       " 'train_auc': 0.877682750961473,\n",
       " 'test_auc': 0.8650679808886864,\n",
       " 'precision': 0.7859858772406301,\n",
       " 'specificity': 0.7988769780500256,\n",
       " 'recall': 0.7700904736562001,\n",
       " 'f_score': 0.7779569892473118}"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "True Negatives: 1565\n",
      "False Positives: 394\n",
      "False Negatives: 432\n",
      "True Positives: 1447\n"
     ]
    }
   ],
   "source": [
    "gb = run_model('gb')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:48.539740Z",
     "start_time": "2021-01-21T08:04:27.674992Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "### METRICS ###\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'model': 'svc',\n",
       " 'train_auc': 0.8885165314480307,\n",
       " 'test_auc': 0.8527667095630733,\n",
       " 'precision': 0.7967064168086314,\n",
       " 'specificity': 0.8172537008677897,\n",
       " 'recall': 0.746673762639702,\n",
       " 'f_score': 0.7708791208791208}"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "True Negatives: 1601\n",
      "False Positives: 358\n",
      "False Negatives: 476\n",
      "True Positives: 1403\n"
     ]
    }
   ],
   "source": [
    "svc = run_model('svc')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:49.306893Z",
     "start_time": "2021-01-21T08:04:48.541734Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "### METRICS ###\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'model': 'ada',\n",
       " 'train_auc': 0.8517299029883305,\n",
       " 'test_auc': 0.8485251541649041,\n",
       " 'precision': 0.7730769230769231,\n",
       " 'specificity': 0.7891781521184278,\n",
       " 'recall': 0.7488025545502927,\n",
       " 'f_score': 0.7607461476074615}"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "True Negatives: 1546\n",
      "False Positives: 413\n",
      "False Negatives: 472\n",
      "True Positives: 1407\n"
     ]
    }
   ],
   "source": [
    "ada = run_model('ada')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:50.806030Z",
     "start_time": "2021-01-21T08:04:49.307893Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "### METRICS ###\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'model': 'xgb',\n",
       " 'train_auc': 0.9934760781132403,\n",
       " 'test_auc': 0.8606760571492064,\n",
       " 'precision': 0.7783902976846747,\n",
       " 'specificity': 0.7947932618683001,\n",
       " 'recall': 0.7514635444385311,\n",
       " 'f_score': 0.7646899539669645}"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "True Negatives: 1557\n",
      "False Positives: 402\n",
      "False Negatives: 467\n",
      "True Positives: 1412\n"
     ]
    }
   ],
   "source": [
    "xgb = run_model('xgb')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:50.822034Z",
     "start_time": "2021-01-21T08:04:50.807030Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<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>model</th>\n",
       "      <th>train_auc</th>\n",
       "      <th>test_auc</th>\n",
       "      <th>precision</th>\n",
       "      <th>specificity</th>\n",
       "      <th>recall</th>\n",
       "      <th>f_score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>et</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.868589</td>\n",
       "      <td>0.787528</td>\n",
       "      <td>0.803471</td>\n",
       "      <td>0.759447</td>\n",
       "      <td>0.773232</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>rf</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.866531</td>\n",
       "      <td>0.794232</td>\n",
       "      <td>0.810618</td>\n",
       "      <td>0.762108</td>\n",
       "      <td>0.777838</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>gb</td>\n",
       "      <td>0.877683</td>\n",
       "      <td>0.865068</td>\n",
       "      <td>0.785986</td>\n",
       "      <td>0.798877</td>\n",
       "      <td>0.770090</td>\n",
       "      <td>0.777957</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>xgb</td>\n",
       "      <td>0.993476</td>\n",
       "      <td>0.860676</td>\n",
       "      <td>0.778390</td>\n",
       "      <td>0.794793</td>\n",
       "      <td>0.751464</td>\n",
       "      <td>0.764690</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>svc</td>\n",
       "      <td>0.888517</td>\n",
       "      <td>0.852767</td>\n",
       "      <td>0.796706</td>\n",
       "      <td>0.817254</td>\n",
       "      <td>0.746674</td>\n",
       "      <td>0.770879</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>ada</td>\n",
       "      <td>0.851730</td>\n",
       "      <td>0.848525</td>\n",
       "      <td>0.773077</td>\n",
       "      <td>0.789178</td>\n",
       "      <td>0.748803</td>\n",
       "      <td>0.760746</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>lr</td>\n",
       "      <td>0.844153</td>\n",
       "      <td>0.848309</td>\n",
       "      <td>0.783425</td>\n",
       "      <td>0.799898</td>\n",
       "      <td>0.754657</td>\n",
       "      <td>0.768772</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  model  train_auc  test_auc  precision  specificity    recall   f_score\n",
       "0    et   1.000000  0.868589   0.787528     0.803471  0.759447  0.773232\n",
       "1    rf   1.000000  0.866531   0.794232     0.810618  0.762108  0.777838\n",
       "2    gb   0.877683  0.865068   0.785986     0.798877  0.770090  0.777957\n",
       "3   xgb   0.993476  0.860676   0.778390     0.794793  0.751464  0.764690\n",
       "4   svc   0.888517  0.852767   0.796706     0.817254  0.746674  0.770879\n",
       "5   ada   0.851730  0.848525   0.773077     0.789178  0.748803  0.760746\n",
       "6    lr   0.844153  0.848309   0.783425     0.799898  0.754657  0.768772"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Results of our initial modelling\n",
    "pd.DataFrame(init_list).sort_values(by='test_auc', ascending=False).reset_index(drop=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Hyperparameter Tuning"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "cell_id": "00034-acfe72b2-8985-42e7-98bd-133e3a36df9d",
    "deepnote_cell_type": "markdown",
    "output_cleared": false
   },
   "source": [
    "### Logistic Regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:14:47.842145Z",
     "start_time": "2021-01-21T08:14:47.830144Z"
    },
    "cell_id": "00035-cac21cb8-4fdf-4b01-a757-4f5b80a882b9",
    "deepnote_cell_type": "code",
    "execution_millis": 1,
    "execution_start": 1608128623190,
    "output_cleared": false,
    "source_hash": "3e948428"
   },
   "outputs": [],
   "source": [
    "lr_params = {\n",
    "    # Trying different types of regularization\n",
    "    'lr__penalty':['l2','l1', 'elasticnet'],\n",
    "    \n",
    "    'lr__l1_ratio':[0, 0.5, 1],\n",
    "\n",
    "     # Trying different alphas of: 10, 1, 0.1  (C = 1/alpha)\n",
    "    'lr__C':[0.1, 1, 10],\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:14:48.874850Z",
     "start_time": "2021-01-21T08:14:48.857840Z"
    },
    "cell_id": "00036-00347ffc-97e1-49e6-b921-a614257a110e",
    "deepnote_cell_type": "code",
    "execution_millis": 17669,
    "execution_start": 1608128623234,
    "output_cleared": false,
    "source_hash": "fb9d1e6b"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 3 folds for each of 27 candidates, totalling 81 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 12 concurrent workers.\n",
      "[Parallel(n_jobs=-1)]: Done  26 tasks      | elapsed:    2.4s\n",
      "[Parallel(n_jobs=-1)]: Done  81 out of  81 | elapsed:    6.8s finished\n",
      "C:\\Users\\benja\\anaconda3\\lib\\site-packages\\sklearn\\linear_model\\_logistic.py:1315: UserWarning: l1_ratio parameter is only used when penalty is 'elasticnet'. Got (penalty=l1)\n",
      "  warnings.warn(\"l1_ratio parameter is only used when penalty is \"\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "### BEST PARAMS ###\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "LogisticRegression(C=1, l1_ratio=0, max_iter=5000, penalty='l1',\n",
       "                   random_state=42, solver='saga')"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "### METRICS ###\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'model': 'lr',\n",
       " 'train_auc': 0.8441560877822272,\n",
       " 'test_auc': 0.8483385181206755,\n",
       " 'precision': 0.7838584853510226,\n",
       " 'specificity': 0.8004083716181726,\n",
       " 'recall': 0.7546567323044172,\n",
       " 'f_score': 0.7689804772234274}"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "True Negatives: 1568\n",
      "False Positives: 391\n",
      "False Negatives: 461\n",
      "True Positives: 1418\n"
     ]
    }
   ],
   "source": [
    "lr_gs = run_model('lr', mod_params=lr_params, grid_search=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### SVC"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.369174Z",
     "start_time": "2021-01-21T08:04:19.869Z"
    }
   },
   "outputs": [],
   "source": [
    "svc_params = {\n",
    "    'svc__C':[0.05, 1],\n",
    "    'svc__gamma':[0.05, 0.1], \n",
    "    'svc__kernel':['rbf', 'sigmoid'],\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.369174Z",
     "start_time": "2021-01-21T08:04:19.871Z"
    },
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 3 folds for each of 8 candidates, totalling 24 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 12 concurrent workers.\n",
      "[Parallel(n_jobs=-1)]: Done  24 out of  24 | elapsed:   15.9s finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "### BEST PARAMS ###\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "SVC(C=1, gamma=0.05, probability=True, random_state=42)"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "### METRICS ###\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'model': 'svc',\n",
       " 'train_auc': 0.9051034339621418,\n",
       " 'test_auc': 0.8490527337833789,\n",
       " 'precision': 0.788826815642458,\n",
       " 'specificity': 0.8070444104134763,\n",
       " 'recall': 0.7514635444385311,\n",
       " 'f_score': 0.7696920141727991}"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "True Negatives: 1581\n",
      "False Positives: 378\n",
      "False Negatives: 467\n",
      "True Positives: 1412\n"
     ]
    }
   ],
   "source": [
    "svc_gs =  run_model('svc', mod_params=svc_params, grid_search=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "cell_id": "00047-e387c823-24f3-4b10-bd76-788e4ea8612a",
    "deepnote_cell_type": "markdown",
    "output_cleared": false
   },
   "source": [
    "### Random Forest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.370159Z",
     "start_time": "2021-01-21T08:04:19.873Z"
    },
    "cell_id": "00048-153c3751-b937-475e-9efe-3272e717f71e",
    "deepnote_cell_type": "code",
    "execution_millis": 2,
    "output_cleared": false,
    "source_hash": "818ae80f"
   },
   "outputs": [],
   "source": [
    "rf_params = {'rf__n_estimators': [100, 200],\n",
    "             'rf__max_depth': [20, 25, 30, 40],\n",
    "             'rf__min_samples_leaf': [2, 3, 4],\n",
    "            }"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.371177Z",
     "start_time": "2021-01-21T08:04:19.875Z"
    },
    "cell_id": "00049-f8899325-0bf6-48ff-8de3-14a8e474a66d",
    "deepnote_cell_type": "code",
    "execution_millis": 48356,
    "output_cleared": false,
    "source_hash": "b2289f2a"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 3 folds for each of 24 candidates, totalling 72 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 12 concurrent workers.\n",
      "[Parallel(n_jobs=-1)]: Done  26 tasks      | elapsed:    4.3s\n",
      "[Parallel(n_jobs=-1)]: Done  72 out of  72 | elapsed:   10.3s finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "### BEST PARAMS ###\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "RandomForestClassifier(max_depth=20, min_samples_leaf=3, n_estimators=200,\n",
       "                       random_state=42)"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "### METRICS ###\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'model': 'rf',\n",
       " 'train_auc': 0.9912806493288461,\n",
       " 'test_auc': 0.8696699041364471,\n",
       " 'precision': 0.7928767123287671,\n",
       " 'specificity': 0.8070444104134763,\n",
       " 'recall': 0.7700904736562001,\n",
       " 'f_score': 0.781317494600432}"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "True Negatives: 1581\n",
      "False Positives: 378\n",
      "False Negatives: 432\n",
      "True Positives: 1447\n"
     ]
    }
   ],
   "source": [
    "rf_gs =  run_model('rf', mod_params=rf_params, grid_search=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Extra Trees"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.372170Z",
     "start_time": "2021-01-21T08:04:19.877Z"
    }
   },
   "outputs": [],
   "source": [
    "et_params = {'et__n_estimators': [100, 200],\n",
    "             'et__max_depth': [20, 25, 30, 35, 40, 50],\n",
    "             'et__min_samples_leaf': [2, 3, 4],\n",
    "            }"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.372170Z",
     "start_time": "2021-01-21T08:04:19.879Z"
    },
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 3 folds for each of 36 candidates, totalling 108 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 12 concurrent workers.\n",
      "[Parallel(n_jobs=-1)]: Done  26 tasks      | elapsed:    2.6s\n",
      "[Parallel(n_jobs=-1)]: Done 108 out of 108 | elapsed:    9.7s finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "### BEST PARAMS ###\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "ExtraTreesClassifier(max_depth=40, min_samples_leaf=2, n_estimators=200,\n",
       "                     random_state=42)"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "### METRICS ###\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'model': 'et',\n",
       " 'train_auc': 0.9988750448088691,\n",
       " 'test_auc': 0.8728125617196161,\n",
       " 'precision': 0.7961538461538461,\n",
       " 'specificity': 0.810617662072486,\n",
       " 'recall': 0.7711548696114955,\n",
       " 'f_score': 0.7834549878345498}"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "True Negatives: 1588\n",
      "False Positives: 371\n",
      "False Negatives: 430\n",
      "True Positives: 1449\n"
     ]
    }
   ],
   "source": [
    "et_gs =  run_model('et', mod_params=et_params, grid_search=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "cell_id": "00050-a0445189-7d2d-4431-96fa-20e6147b2da5",
    "deepnote_cell_type": "markdown",
    "deletable": false,
    "editable": false,
    "output_cleared": false,
    "run_control": {
     "frozen": true
    }
   },
   "source": [
    "### Adaptive Boosting"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.373169Z",
     "start_time": "2021-01-21T08:04:19.881Z"
    },
    "cell_id": "00051-2b9d38d8-36cc-4432-908a-a672e2fd1a3b",
    "deepnote_cell_type": "code",
    "execution_millis": 1,
    "output_cleared": false,
    "source_hash": "605ca1e4"
   },
   "outputs": [],
   "source": [
    "ada_params = {'ada__n_estimators': [500],\n",
    "              'ada__learning_rate': [0.05, 0.1, 0.2, 0.9],\n",
    "             }"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.373169Z",
     "start_time": "2021-01-21T08:04:19.884Z"
    },
    "cell_id": "00052-00963070-2a52-4a6b-8861-9f265cae8d4a",
    "deepnote_cell_type": "code",
    "execution_millis": 147596,
    "output_cleared": false,
    "scrolled": false,
    "source_hash": "3043ad5b"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 3 folds for each of 4 candidates, totalling 12 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 12 concurrent workers.\n",
      "[Parallel(n_jobs=-1)]: Done   2 out of  12 | elapsed:    4.3s remaining:   22.1s\n",
      "[Parallel(n_jobs=-1)]: Done  12 out of  12 | elapsed:    4.5s finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "### BEST PARAMS ###\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "AdaBoostClassifier(learning_rate=0.1, n_estimators=500, random_state=42)"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "### METRICS ###\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'model': 'ada',\n",
       " 'train_auc': 0.851758141618727,\n",
       " 'test_auc': 0.8482246891504692,\n",
       " 'precision': 0.786368843069874,\n",
       " 'specificity': 0.8095967330270546,\n",
       " 'recall': 0.7307078233102714,\n",
       " 'f_score': 0.7575172413793104}"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "True Negatives: 1586\n",
      "False Positives: 373\n",
      "False Negatives: 506\n",
      "True Positives: 1373\n"
     ]
    }
   ],
   "source": [
    "ada_gs = run_model('ada', mod_params=ada_params, grid_search=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "cell_id": "00053-59769990-73ce-4a7e-b426-b844f3519332",
    "deepnote_cell_type": "markdown",
    "output_cleared": false
   },
   "source": [
    "### Gradient Boosting"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.374171Z",
     "start_time": "2021-01-21T08:04:19.886Z"
    },
    "cell_id": "00054-59a272ab-9194-4dc6-b774-9163715b25c8",
    "deepnote_cell_type": "code",
    "execution_millis": 1,
    "output_cleared": false,
    "source_hash": "d4dc6cc7"
   },
   "outputs": [],
   "source": [
    "gb_params = {'gb__n_estimators': [500, 1000],\n",
    "             'gb__learning_rate': [0.1, 0.2, 0.3],\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.375168Z",
     "start_time": "2021-01-21T08:04:19.888Z"
    },
    "cell_id": "00055-5ff7ba13-4cdf-4ccf-88c6-194f5f4e2af5",
    "deepnote_cell_type": "code",
    "execution_millis": 397867,
    "output_cleared": false,
    "scrolled": false,
    "source_hash": "f23e8be3"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 3 folds for each of 6 candidates, totalling 18 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 12 concurrent workers.\n",
      "[Parallel(n_jobs=-1)]: Done  14 out of  18 | elapsed:   15.8s remaining:    4.5s\n",
      "[Parallel(n_jobs=-1)]: Done  18 out of  18 | elapsed:   22.3s finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "### BEST PARAMS ###\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "GradientBoostingClassifier(n_estimators=500, random_state=42)"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "### METRICS ###\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'model': 'gb',\n",
       " 'train_auc': 0.931495527345197,\n",
       " 'test_auc': 0.8660295504353347,\n",
       " 'precision': 0.7862554112554112,\n",
       " 'specificity': 0.7983665135273098,\n",
       " 'recall': 0.7732836615220862,\n",
       " 'f_score': 0.7797155889455325}"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "True Negatives: 1564\n",
      "False Positives: 395\n",
      "False Negatives: 426\n",
      "True Positives: 1453\n"
     ]
    }
   ],
   "source": [
    "gb_gs = run_model('gb', mod_params=gb_params, grid_search=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### XGBoost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.375168Z",
     "start_time": "2021-01-21T08:04:19.892Z"
    }
   },
   "outputs": [],
   "source": [
    "xgb_params = { \n",
    "              'xgb__learning_rate' : [0.05, 0.1, 0.3],\n",
    "              'xgb__max_depth' : [6, 7], \n",
    "              'xgb__n_estimators' : [100, 150],\n",
    "              'xgb__reg_alpha' : [0, 1, 2],\n",
    "              'xgb__reg_lambda' : [0, 1, 2],\n",
    "              'xgb__gamma' : [0, 1, 3],\n",
    "             }"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.376169Z",
     "start_time": "2021-01-21T08:04:19.894Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 3 folds for each of 324 candidates, totalling 972 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 12 concurrent workers.\n",
      "[Parallel(n_jobs=-1)]: Done  26 tasks      | elapsed:    5.6s\n",
      "[Parallel(n_jobs=-1)]: Done 176 tasks      | elapsed:   36.9s\n",
      "[Parallel(n_jobs=-1)]: Done 426 tasks      | elapsed:  1.5min\n",
      "[Parallel(n_jobs=-1)]: Done 776 tasks      | elapsed:  2.7min\n",
      "[Parallel(n_jobs=-1)]: Done 972 out of 972 | elapsed:  3.3min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "### BEST PARAMS ###\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
       "              colsample_bynode=1, colsample_bytree=1, eval_metric='auc',\n",
       "              gamma=0, gpu_id=-1, importance_type='gain',\n",
       "              interaction_constraints='', learning_rate=0.05, max_delta_step=0,\n",
       "              max_depth=6, min_child_weight=1, missing=nan,\n",
       "              monotone_constraints='()', n_estimators=100, n_jobs=12,\n",
       "              num_parallel_tree=1, random_state=42, reg_alpha=0, reg_lambda=0,\n",
       "              scale_pos_weight=1, seed=42, subsample=1, tree_method='exact',\n",
       "              use_label_encoder=False, validate_parameters=1, verbosity=None)"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "### METRICS ###\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'model': 'xgb',\n",
       " 'train_auc': 0.91806184709081,\n",
       " 'test_auc': 0.8727219603793683,\n",
       " 'precision': 0.7916666666666666,\n",
       " 'specificity': 0.8034711587544666,\n",
       " 'recall': 0.778605641298563,\n",
       " 'f_score': 0.7850818352562383}"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "True Negatives: 1574\n",
      "False Positives: 385\n",
      "False Negatives: 416\n",
      "True Positives: 1463\n"
     ]
    }
   ],
   "source": [
    "xgb_gs = run_model('xgb', mod_params=xgb_params, grid_search=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## ROC-AUC Evaluation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.376169Z",
     "start_time": "2021-01-21T08:04:19.898Z"
    },
    "cell_id": "00056-6e98fdee-308e-4356-92fb-506363492b9a",
    "deepnote_cell_type": "code",
    "execution_millis": 13,
    "output_cleared": false,
    "source_hash": "a3600245"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<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>model</th>\n",
       "      <th>train_auc</th>\n",
       "      <th>test_auc</th>\n",
       "      <th>precision</th>\n",
       "      <th>specificity</th>\n",
       "      <th>recall</th>\n",
       "      <th>f_score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>et</td>\n",
       "      <td>0.998875</td>\n",
       "      <td>0.872813</td>\n",
       "      <td>0.796154</td>\n",
       "      <td>0.810618</td>\n",
       "      <td>0.771155</td>\n",
       "      <td>0.783455</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>xgb</td>\n",
       "      <td>0.918062</td>\n",
       "      <td>0.872722</td>\n",
       "      <td>0.791667</td>\n",
       "      <td>0.803471</td>\n",
       "      <td>0.778606</td>\n",
       "      <td>0.785082</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>rf</td>\n",
       "      <td>0.991281</td>\n",
       "      <td>0.869670</td>\n",
       "      <td>0.792877</td>\n",
       "      <td>0.807044</td>\n",
       "      <td>0.770090</td>\n",
       "      <td>0.781317</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>gb</td>\n",
       "      <td>0.931496</td>\n",
       "      <td>0.866030</td>\n",
       "      <td>0.786255</td>\n",
       "      <td>0.798367</td>\n",
       "      <td>0.773284</td>\n",
       "      <td>0.779716</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>svc</td>\n",
       "      <td>0.905103</td>\n",
       "      <td>0.849053</td>\n",
       "      <td>0.788827</td>\n",
       "      <td>0.807044</td>\n",
       "      <td>0.751464</td>\n",
       "      <td>0.769692</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>lr</td>\n",
       "      <td>0.844156</td>\n",
       "      <td>0.848339</td>\n",
       "      <td>0.783858</td>\n",
       "      <td>0.800408</td>\n",
       "      <td>0.754657</td>\n",
       "      <td>0.768980</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>ada</td>\n",
       "      <td>0.851758</td>\n",
       "      <td>0.848225</td>\n",
       "      <td>0.786369</td>\n",
       "      <td>0.809597</td>\n",
       "      <td>0.730708</td>\n",
       "      <td>0.757517</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  model  train_auc  test_auc  precision  specificity    recall   f_score\n",
       "3    et   0.998875  0.872813   0.796154     0.810618  0.771155  0.783455\n",
       "6   xgb   0.918062  0.872722   0.791667     0.803471  0.778606  0.785082\n",
       "2    rf   0.991281  0.869670   0.792877     0.807044  0.770090  0.781317\n",
       "5    gb   0.931496  0.866030   0.786255     0.798367  0.773284  0.779716\n",
       "1   svc   0.905103  0.849053   0.788827     0.807044  0.751464  0.769692\n",
       "0    lr   0.844156  0.848339   0.783858     0.800408  0.754657  0.768980\n",
       "4   ada   0.851758  0.848225   0.786369     0.809597  0.730708  0.757517"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gs_df = pd.DataFrame(gs_list)\n",
    "gs_df.sort_values(by='test_auc', ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.377168Z",
     "start_time": "2021-01-21T08:04:19.900Z"
    }
   },
   "outputs": [],
   "source": [
    "gs_dict = {\n",
    "    xgb_gs: 'XGBoostClassifier',\n",
    "    et_gs: 'ExtraTreeClassifier',\n",
    "    rf_gs: 'RandomForest',\n",
    "    gb_gs: 'GradientBoostingClassifier',\n",
    "    svc_gs: 'SupportVectorMachineClf',\n",
    "    lr_gs: 'LogisticRegression',\n",
    "    ada_gs: 'AdaBoostClassifier',\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.378162Z",
     "start_time": "2021-01-21T08:04:19.902Z"
    }
   },
   "outputs": [],
   "source": [
    "def roc_curve_plotter(model_dict, plot_top=False):\n",
    "    fig, ax = plt.subplots(1, 1, figsize=(12,10))\n",
    "    axes = {}\n",
    "    for i, m in enumerate(model_dict.keys()):\n",
    "        axes[f'ax{i}'] = plot_roc_curve(m, X_test, y_test, ax=ax, name=model_dict[m])\n",
    "    if plot_top:\n",
    "        for i, a in enumerate(axes):\n",
    "            if i != 0:\n",
    "                axes[a].line_.set_color('gray')\n",
    "                axes[a].line_.set_alpha(0.25)\n",
    "                \n",
    "    plt.plot([0, 1], [0, 1], color='black', lw=2, linestyle='--', label='Random Guess')\n",
    "    plt.title('ROC-AUC Curve', fontsize=22)\n",
    "    plt.xlabel('False Positive Rate', fontsize=12)\n",
    "    plt.ylabel('True Positive Rate', fontsize=12)\n",
    "    plt.legend(fontsize=12)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Performance is pretty close between our ensemble tree models and XGBoost -- we can see that XGBoost has a TPR of 0.8 to a FPR of roughly 0.21."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.379165Z",
     "start_time": "2021-01-21T08:04:19.904Z"
    },
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "roc_curve_plotter(gs_dict, True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Testing against unseen data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "At this point, we're taking our data and testing it on \"unseen\" data, or the test set that we earlier created. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.380163Z",
     "start_time": "2021-01-21T08:04:19.907Z"
    }
   },
   "outputs": [],
   "source": [
    "final_actual = test['is_popular']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.380163Z",
     "start_time": "2021-01-21T08:04:19.909Z"
    }
   },
   "outputs": [],
   "source": [
    "test = test.drop(columns=['is_popular'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.381172Z",
     "start_time": "2021-01-21T08:04:19.912Z"
    }
   },
   "outputs": [],
   "source": [
    "def evaluate_model(model):\n",
    "    final_proba = model.predict_proba(test)\n",
    "    final_proba = final_proba[:, 1]\n",
    "    final_preds = model.predict(test)\n",
    "    print(f'Accuracy:', f'{accuracy_score(final_actual, final_preds):.3f}')\n",
    "    print(f'Precision:', f'{precision_score(final_actual, final_preds):.3f}')\n",
    "    print(f'Recall:', f'{recall_score(final_actual, final_preds):.3f}')\n",
    "    print(f'ROC-AUC:', f'{roc_auc_score(final_actual, final_proba):.3f}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.382170Z",
     "start_time": "2021-01-21T08:04:19.914Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.782\n",
      "Precision: 0.768\n",
      "Recall: 0.787\n",
      "ROC-AUC: 0.865\n"
     ]
    }
   ],
   "source": [
    "evaluate_model(xgb_gs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.382170Z",
     "start_time": "2021-01-21T08:04:19.915Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.784\n",
      "Precision: 0.775\n",
      "Recall: 0.781\n",
      "ROC-AUC: 0.864\n"
     ]
    }
   ],
   "source": [
    "evaluate_model(rf_gs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.383170Z",
     "start_time": "2021-01-21T08:04:19.917Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.778\n",
      "Precision: 0.777\n",
      "Recall: 0.760\n",
      "ROC-AUC: 0.866\n"
     ]
    }
   ],
   "source": [
    "evaluate_model(et_gs)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Refit models on entire dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.384171Z",
     "start_time": "2021-01-21T08:04:19.919Z"
    },
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "xgb_gs.fit(X, y)\n",
    "rf_gs.fit(X, y)\n",
    "et_gs.fit(X, y);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Performance for our top models is pretty close -- but looks like XGBoost wins. It has the highest accuracy and AUC-ROC score."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.385176Z",
     "start_time": "2021-01-21T08:04:19.925Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.786\n",
      "Precision: 0.775\n",
      "Recall: 0.785\n",
      "ROC-AUC: 0.869\n"
     ]
    }
   ],
   "source": [
    "evaluate_model(xgb_gs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.386169Z",
     "start_time": "2021-01-21T08:04:19.927Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.781\n",
      "Precision: 0.772\n",
      "Recall: 0.776\n",
      "ROC-AUC: 0.866\n"
     ]
    }
   ],
   "source": [
    "evaluate_model(rf_gs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.387163Z",
     "start_time": "2021-01-21T08:04:19.929Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.779\n",
      "Precision: 0.780\n",
      "Recall: 0.756\n",
      "ROC-AUC: 0.866\n"
     ]
    }
   ],
   "source": [
    "evaluate_model(et_gs)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Misclassification Analysis"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.387163Z",
     "start_time": "2021-01-21T08:04:19.931Z"
    }
   },
   "outputs": [],
   "source": [
    "final_proba = rf_gs.predict_proba(test)\n",
    "final_proba = final_proba[:, 1]\n",
    "final_preds = rf_gs.predict(test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.388164Z",
     "start_time": "2021-01-21T08:04:19.933Z"
    }
   },
   "outputs": [],
   "source": [
    "predictions = pd.DataFrame(data=[final_actual, final_preds, final_proba]).T\n",
    "predictions.columns = ['actual', 'predicted', 'proba']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.388164Z",
     "start_time": "2021-01-21T08:04:19.935Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<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>actual</th>\n",
       "      <th>predicted</th>\n",
       "      <th>proba</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.969706</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.988791</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.908532</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.994417</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.234660</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3990</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.847033</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3991</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.081849</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3992</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.519875</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3993</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.321361</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3994</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.201106</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>3995 rows × 3 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      actual  predicted     proba\n",
       "0        1.0        1.0  0.969706\n",
       "1        1.0        1.0  0.988791\n",
       "2        1.0        1.0  0.908532\n",
       "3        1.0        1.0  0.994417\n",
       "4        0.0        0.0  0.234660\n",
       "...      ...        ...       ...\n",
       "3990     1.0        1.0  0.847033\n",
       "3991     0.0        0.0  0.081849\n",
       "3992     1.0        1.0  0.519875\n",
       "3993     1.0        0.0  0.321361\n",
       "3994     1.0        0.0  0.201106\n",
       "\n",
       "[3995 rows x 3 columns]"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.389163Z",
     "start_time": "2021-01-21T08:04:19.937Z"
    }
   },
   "outputs": [],
   "source": [
    "wrong_predictions = predictions[predictions['actual'] != predictions['predicted']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.389163Z",
     "start_time": "2021-01-21T08:04:19.939Z"
    }
   },
   "outputs": [],
   "source": [
    "init_test = pd.read_pickle('./datasets/test.pickle').reset_index(drop=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.390164Z",
     "start_time": "2021-01-21T08:04:19.941Z"
    }
   },
   "outputs": [],
   "source": [
    "wrong_predictions = pd.merge(wrong_predictions, init_test, left_index=True, right_index=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.391163Z",
     "start_time": "2021-01-21T08:04:19.943Z"
    }
   },
   "outputs": [],
   "source": [
    "wrong_predictions = wrong_predictions[['actual', 'predicted', 'proba', 'newsdesk', 'section', 'subsection',\n",
    "                                       'headline', 'abstract', 'keywords']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.391163Z",
     "start_time": "2021-01-21T08:04:19.945Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.0    442\n",
       "1.0    432\n",
       "Name: actual, dtype: int64"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 1 = false positives, 0 = false negatives\n",
    "wrong_predictions['actual'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.392165Z",
     "start_time": "2021-01-21T08:04:19.947Z"
    },
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<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>actual</th>\n",
       "      <th>predicted</th>\n",
       "      <th>proba</th>\n",
       "      <th>newsdesk</th>\n",
       "      <th>section</th>\n",
       "      <th>subsection</th>\n",
       "      <th>headline</th>\n",
       "      <th>abstract</th>\n",
       "      <th>keywords</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>254</th>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.937960</td>\n",
       "      <td>OpEd</td>\n",
       "      <td>Opinion</td>\n",
       "      <td>NaN</td>\n",
       "      <td>What We Can Learn From the Rise and Fall of ‘Political Blackness’</td>\n",
       "      <td>Before “people of color” and BIPOC, Britain had its own inclusive concept for nonwhite people.</td>\n",
       "      <td>[Black People, Blacks, Race and Ethnicity, Minorities, Great Britain, Politics and Government]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>66</th>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.931563</td>\n",
       "      <td>Politics</td>\n",
       "      <td>U.S.</td>\n",
       "      <td>Politics</td>\n",
       "      <td>Democrats Mount an All-Out Effort to Get Detroit to Vote</td>\n",
       "      <td>Urgently trying to avoid a repeat of 2016, when Donald Trump won Michigan by a razor-thin margin, Democrats have bee...</td>\n",
       "      <td>[Presidential Election of 2020, Black People, Blacks, Biden, Joseph R Jr, Harris, Kamala D, Trump, Donald J, Detroit...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>321</th>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.844512</td>\n",
       "      <td>Washington</td>\n",
       "      <td>U.S.</td>\n",
       "      <td>Politics</td>\n",
       "      <td>Under Pence, Politics Regularly Seeped Into the Coronavirus Task Force</td>\n",
       "      <td>In taking a leading role in managing the White House’s response to the pandemic, the vice president and his team had...</td>\n",
       "      <td>[Pence, Mike, Coronavirus (2019-nCoV), United States Politics and Government, Presidential Election of 2020, Trump, ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>190</th>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.837098</td>\n",
       "      <td>The Upshot</td>\n",
       "      <td>The Upshot</td>\n",
       "      <td>NaN</td>\n",
       "      <td>The Pandemic Has Hindered Many of the Best Ideas for Reducing Violence</td>\n",
       "      <td>That may be one part of this year’s rise in violent crime.</td>\n",
       "      <td>[Crime and Criminals, Therapy and Rehabilitation, Nonprofit Organizations, Shutdowns (Institutional), Coronavirus (2...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>318</th>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.829171</td>\n",
       "      <td>Metropolitan</td>\n",
       "      <td>New York</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Their Buzzy Off Broadway Play Shut Down. Here’s What They Did Next.</td>\n",
       "      <td>When the virus forced New York theater to go dark, it upended thousands of lives, from actors to ticket takers. An a...</td>\n",
       "      <td>[Theater, New York City, Playwrights Horizons, Greenfield, Adam, Khoury, Sylvia, Selling Kabul (Play), Quarantine (L...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>227</th>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.826981</td>\n",
       "      <td>OpEd</td>\n",
       "      <td>Opinion</td>\n",
       "      <td>NaN</td>\n",
       "      <td>One Woman’s Revenge Against the Nazis</td>\n",
       "      <td>Her actions have become legend in a small Czech town.</td>\n",
       "      <td>[World War II (1939-45), Holocaust and the Nazi Era, Czech Republic, Documentary Films and Programs, Sex Crimes, Syp...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>423</th>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.821369</td>\n",
       "      <td>Washington</td>\n",
       "      <td>U.S.</td>\n",
       "      <td>Politics</td>\n",
       "      <td>Parties Offer Divergent Portraits of Barrett as Senate Opens Hearings</td>\n",
       "      <td>Democrats will portray President Trump’s Supreme Court nominee as an ideologue, while Republicans will paint her as ...</td>\n",
       "      <td>[Barrett, Amy Coney, Supreme Court (US), Appointments and Executive Changes, United States Politics and Government, ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>426</th>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.808347</td>\n",
       "      <td>Washington</td>\n",
       "      <td>U.S.</td>\n",
       "      <td>Politics</td>\n",
       "      <td>Trump’s Virus Treatment Revives Questions About Unchecked Nuclear Authority</td>\n",
       "      <td>Even before the president was given mood-altering drugs, there was a movement to end the commander in chief’s sole a...</td>\n",
       "      <td>[Trump, Donald J, White House Coronavirus Outbreak (2020), Coronavirus (2019-nCoV), Dexamethasone (Drug), Twenty-Fif...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>69</th>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.803129</td>\n",
       "      <td>Washington</td>\n",
       "      <td>U.S.</td>\n",
       "      <td>Politics</td>\n",
       "      <td>Peace Corps Faces Questions Over Another Volunteer Death</td>\n",
       "      <td>A 24-year-old volunteer died of undiagnosed malaria on the island nation of Comoros. It was one of at least three de...</td>\n",
       "      <td>[Volunteers and Community Service, United States Politics and Government, Peace Corps, Castle, Nicholas Matthew (199...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>462</th>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.800225</td>\n",
       "      <td>Foreign</td>\n",
       "      <td>World</td>\n",
       "      <td>Asia Pacific</td>\n",
       "      <td>‘We Were Bulletproof’: As Child Soldiers Grow Up, Legacy of War Lingers</td>\n",
       "      <td>Twin leaders of Myanmar’s God’s Army were once thought to have magical powers. Now adults, they are contending with ...</td>\n",
       "      <td>[Child Soldiers, Myanmar, Refugees and Displaced Persons, Karen National Union, Defense and Military Forces, God's A...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     actual  predicted     proba      newsdesk     section    subsection  \\\n",
       "254     0.0        1.0  0.937960          OpEd     Opinion           NaN   \n",
       "66      0.0        1.0  0.931563      Politics        U.S.      Politics   \n",
       "321     0.0        1.0  0.844512    Washington        U.S.      Politics   \n",
       "190     0.0        1.0  0.837098    The Upshot  The Upshot           NaN   \n",
       "318     0.0        1.0  0.829171  Metropolitan    New York           NaN   \n",
       "227     0.0        1.0  0.826981          OpEd     Opinion           NaN   \n",
       "423     0.0        1.0  0.821369    Washington        U.S.      Politics   \n",
       "426     0.0        1.0  0.808347    Washington        U.S.      Politics   \n",
       "69      0.0        1.0  0.803129    Washington        U.S.      Politics   \n",
       "462     0.0        1.0  0.800225       Foreign       World  Asia Pacific   \n",
       "\n",
       "                                                                        headline  \\\n",
       "254            What We Can Learn From the Rise and Fall of ‘Political Blackness’   \n",
       "66                      Democrats Mount an All-Out Effort to Get Detroit to Vote   \n",
       "321       Under Pence, Politics Regularly Seeped Into the Coronavirus Task Force   \n",
       "190       The Pandemic Has Hindered Many of the Best Ideas for Reducing Violence   \n",
       "318          Their Buzzy Off Broadway Play Shut Down. Here’s What They Did Next.   \n",
       "227                                        One Woman’s Revenge Against the Nazis   \n",
       "423        Parties Offer Divergent Portraits of Barrett as Senate Opens Hearings   \n",
       "426  Trump’s Virus Treatment Revives Questions About Unchecked Nuclear Authority   \n",
       "69                      Peace Corps Faces Questions Over Another Volunteer Death   \n",
       "462      ‘We Were Bulletproof’: As Child Soldiers Grow Up, Legacy of War Lingers   \n",
       "\n",
       "                                                                                                                    abstract  \\\n",
       "254                           Before “people of color” and BIPOC, Britain had its own inclusive concept for nonwhite people.   \n",
       "66   Urgently trying to avoid a repeat of 2016, when Donald Trump won Michigan by a razor-thin margin, Democrats have bee...   \n",
       "321  In taking a leading role in managing the White House’s response to the pandemic, the vice president and his team had...   \n",
       "190                                                               That may be one part of this year’s rise in violent crime.   \n",
       "318  When the virus forced New York theater to go dark, it upended thousands of lives, from actors to ticket takers. An a...   \n",
       "227                                                                    Her actions have become legend in a small Czech town.   \n",
       "423  Democrats will portray President Trump’s Supreme Court nominee as an ideologue, while Republicans will paint her as ...   \n",
       "426  Even before the president was given mood-altering drugs, there was a movement to end the commander in chief’s sole a...   \n",
       "69   A 24-year-old volunteer died of undiagnosed malaria on the island nation of Comoros. It was one of at least three de...   \n",
       "462  Twin leaders of Myanmar’s God’s Army were once thought to have magical powers. Now adults, they are contending with ...   \n",
       "\n",
       "                                                                                                                    keywords  \n",
       "254                           [Black People, Blacks, Race and Ethnicity, Minorities, Great Britain, Politics and Government]  \n",
       "66   [Presidential Election of 2020, Black People, Blacks, Biden, Joseph R Jr, Harris, Kamala D, Trump, Donald J, Detroit...  \n",
       "321  [Pence, Mike, Coronavirus (2019-nCoV), United States Politics and Government, Presidential Election of 2020, Trump, ...  \n",
       "190  [Crime and Criminals, Therapy and Rehabilitation, Nonprofit Organizations, Shutdowns (Institutional), Coronavirus (2...  \n",
       "318  [Theater, New York City, Playwrights Horizons, Greenfield, Adam, Khoury, Sylvia, Selling Kabul (Play), Quarantine (L...  \n",
       "227  [World War II (1939-45), Holocaust and the Nazi Era, Czech Republic, Documentary Films and Programs, Sex Crimes, Syp...  \n",
       "423  [Barrett, Amy Coney, Supreme Court (US), Appointments and Executive Changes, United States Politics and Government, ...  \n",
       "426  [Trump, Donald J, White House Coronavirus Outbreak (2020), Coronavirus (2019-nCoV), Dexamethasone (Drug), Twenty-Fif...  \n",
       "69   [Volunteers and Community Service, United States Politics and Government, Peace Corps, Castle, Nicholas Matthew (199...  \n",
       "462  [Child Soldiers, Myanmar, Refugees and Displaced Persons, Karen National Union, Defense and Military Forces, God's A...  "
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Looking at false positives\n",
    "wrong_predictions[wrong_predictions['proba'] > 0.80].head(10).sort_values(by='proba', ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.393164Z",
     "start_time": "2021-01-21T08:04:19.949Z"
    },
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<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>actual</th>\n",
       "      <th>predicted</th>\n",
       "      <th>proba</th>\n",
       "      <th>newsdesk</th>\n",
       "      <th>section</th>\n",
       "      <th>subsection</th>\n",
       "      <th>headline</th>\n",
       "      <th>abstract</th>\n",
       "      <th>keywords</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.094006</td>\n",
       "      <td>Weekend</td>\n",
       "      <td>Arts</td>\n",
       "      <td>Music</td>\n",
       "      <td>The Special Place Where Ella Fitzgerald Comes Alive</td>\n",
       "      <td>The singer’s concert recordings have always had a power that her studio outings could only imply. “Ella: The Lost Be...</td>\n",
       "      <td>[Jazz, Fitzgerald, Ella, Ella: The Lost Berlin Tapes (Album)]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.105682</td>\n",
       "      <td>RealEstate</td>\n",
       "      <td>Real Estate</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Real Estate Sales Continue to Stagnate in Manhattan</td>\n",
       "      <td>The number of apartments sold in the last three months was down by 46 percent compared to the same period in 2019.</td>\n",
       "      <td>[Real Estate and Housing (Residential), Quarantine (Life and Culture), Shutdowns (Institutional), Condominiums, Manh...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.237722</td>\n",
       "      <td>Well</td>\n",
       "      <td>Well</td>\n",
       "      <td>Live</td>\n",
       "      <td>Where You Carry Body Fat May Affect How Long You Live</td>\n",
       "      <td>Extra weight in some places may lower your risk of dying prematurely.</td>\n",
       "      <td>[Weight, Deaths (Fatalities)]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>60</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.247927</td>\n",
       "      <td>Styles</td>\n",
       "      <td>Style</td>\n",
       "      <td>NaN</td>\n",
       "      <td>‘Ridicule’: The French Reaction to ‘Emily in Paris’</td>\n",
       "      <td>Darren Star’s latest serial goes down like sour wine for actual Parisians.</td>\n",
       "      <td>[Television, Emily in Paris (TV Program), Paris (France), your-feed-fashion]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.298873</td>\n",
       "      <td>Well</td>\n",
       "      <td>Well</td>\n",
       "      <td>Mind</td>\n",
       "      <td>Laughter May Be Effective Medicine for These Trying Times</td>\n",
       "      <td>Doctors, nurses and therapists have a prescription for helping all of us to get through these difficult times: Try a...</td>\n",
       "      <td>[Content Type: Service, Anxiety and Stress, Laughter, Comedy and Humor, Emergency Medical Treatment, Hospitals, Coro...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>98</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.340034</td>\n",
       "      <td>Summary</td>\n",
       "      <td>Times Insider</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Following Falsehoods: A Reporter’s Approach on QAnon</td>\n",
       "      <td>Kevin Roose, a Times tech columnist, has watched the baseless conspiracy theory grow from fringe internet subculture...</td>\n",
       "      <td>[QAnon, Conspiracy Theories, News and News Media, Rumors and Misinformation, Fringe Groups and Movements, Trump, Don...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.349069</td>\n",
       "      <td>TStyle</td>\n",
       "      <td>T Magazine</td>\n",
       "      <td>NaN</td>\n",
       "      <td>On Long Island, a Beachfront Haven for Black Families</td>\n",
       "      <td>In the 1930s, a group of trailblazing African-Americans bought plots for themselves in Sag Harbor, establishing a cl...</td>\n",
       "      <td>[Black People, Blacks, Sag Harbor (NY), Families and Family Life, Historic Buildings and Sites, Real Estate and Hous...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.353092</td>\n",
       "      <td>Dining</td>\n",
       "      <td>Food</td>\n",
       "      <td>Wine, Beer &amp; Cocktails</td>\n",
       "      <td>Jim Murray, a Top Whiskey Critic, Faces Accusations of Sexism</td>\n",
       "      <td>Several distillers and retailers have severed ties to the British writer, after complaints surfaced about the treatm...</td>\n",
       "      <td>[Whiskey, Murray, Jim (1957- ), Discrimination, Women and Girls, Jim Murray's Whisky Bible (Book), Books and Literat...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>76</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.381530</td>\n",
       "      <td>Science</td>\n",
       "      <td>Health</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Trump Has the Coronavirus. What Risks Does He Face?</td>\n",
       "      <td>Older men have a greater chance of becoming severely ill from Covid-19, though many recover quickly. Experts say the...</td>\n",
       "      <td>[Trump, Donald J, United States, Presidential Election of 2020, Coronavirus (2019-nCoV), Dexamethasone (Drug), Obesi...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>102</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.384231</td>\n",
       "      <td>Express</td>\n",
       "      <td>World</td>\n",
       "      <td>Europe</td>\n",
       "      <td>Arctic Expedition’s Dress Code Raises Concerns About Sexism in Science</td>\n",
       "      <td>The scientific institute leading the trip denied that its policy had been applied to a specific sex, saying, “Women ...</td>\n",
       "      <td>[Wegener, Alfred, Institute for Polar and Marine Research, Science and Technology, Research, Workplace Hazards and V...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     actual  predicted     proba    newsdesk        section  \\\n",
       "34      1.0        0.0  0.094006     Weekend           Arts   \n",
       "50      1.0        0.0  0.105682  RealEstate    Real Estate   \n",
       "33      1.0        0.0  0.237722        Well           Well   \n",
       "60      1.0        0.0  0.247927      Styles          Style   \n",
       "17      1.0        0.0  0.298873        Well           Well   \n",
       "98      1.0        0.0  0.340034     Summary  Times Insider   \n",
       "28      1.0        0.0  0.349069      TStyle     T Magazine   \n",
       "19      1.0        0.0  0.353092      Dining           Food   \n",
       "76      1.0        0.0  0.381530     Science         Health   \n",
       "102     1.0        0.0  0.384231     Express          World   \n",
       "\n",
       "                 subsection  \\\n",
       "34                    Music   \n",
       "50                      NaN   \n",
       "33                     Live   \n",
       "60                      NaN   \n",
       "17                     Mind   \n",
       "98                      NaN   \n",
       "28                      NaN   \n",
       "19   Wine, Beer & Cocktails   \n",
       "76                      NaN   \n",
       "102                  Europe   \n",
       "\n",
       "                                                                   headline  \\\n",
       "34                      The Special Place Where Ella Fitzgerald Comes Alive   \n",
       "50                      Real Estate Sales Continue to Stagnate in Manhattan   \n",
       "33                    Where You Carry Body Fat May Affect How Long You Live   \n",
       "60                      ‘Ridicule’: The French Reaction to ‘Emily in Paris’   \n",
       "17                Laughter May Be Effective Medicine for These Trying Times   \n",
       "98                     Following Falsehoods: A Reporter’s Approach on QAnon   \n",
       "28                    On Long Island, a Beachfront Haven for Black Families   \n",
       "19            Jim Murray, a Top Whiskey Critic, Faces Accusations of Sexism   \n",
       "76                      Trump Has the Coronavirus. What Risks Does He Face?   \n",
       "102  Arctic Expedition’s Dress Code Raises Concerns About Sexism in Science   \n",
       "\n",
       "                                                                                                                    abstract  \\\n",
       "34   The singer’s concert recordings have always had a power that her studio outings could only imply. “Ella: The Lost Be...   \n",
       "50        The number of apartments sold in the last three months was down by 46 percent compared to the same period in 2019.   \n",
       "33                                                     Extra weight in some places may lower your risk of dying prematurely.   \n",
       "60                                                Darren Star’s latest serial goes down like sour wine for actual Parisians.   \n",
       "17   Doctors, nurses and therapists have a prescription for helping all of us to get through these difficult times: Try a...   \n",
       "98   Kevin Roose, a Times tech columnist, has watched the baseless conspiracy theory grow from fringe internet subculture...   \n",
       "28   In the 1930s, a group of trailblazing African-Americans bought plots for themselves in Sag Harbor, establishing a cl...   \n",
       "19   Several distillers and retailers have severed ties to the British writer, after complaints surfaced about the treatm...   \n",
       "76   Older men have a greater chance of becoming severely ill from Covid-19, though many recover quickly. Experts say the...   \n",
       "102  The scientific institute leading the trip denied that its policy had been applied to a specific sex, saying, “Women ...   \n",
       "\n",
       "                                                                                                                    keywords  \n",
       "34                                                             [Jazz, Fitzgerald, Ella, Ella: The Lost Berlin Tapes (Album)]  \n",
       "50   [Real Estate and Housing (Residential), Quarantine (Life and Culture), Shutdowns (Institutional), Condominiums, Manh...  \n",
       "33                                                                                             [Weight, Deaths (Fatalities)]  \n",
       "60                                              [Television, Emily in Paris (TV Program), Paris (France), your-feed-fashion]  \n",
       "17   [Content Type: Service, Anxiety and Stress, Laughter, Comedy and Humor, Emergency Medical Treatment, Hospitals, Coro...  \n",
       "98   [QAnon, Conspiracy Theories, News and News Media, Rumors and Misinformation, Fringe Groups and Movements, Trump, Don...  \n",
       "28   [Black People, Blacks, Sag Harbor (NY), Families and Family Life, Historic Buildings and Sites, Real Estate and Hous...  \n",
       "19   [Whiskey, Murray, Jim (1957- ), Discrimination, Women and Girls, Jim Murray's Whisky Bible (Book), Books and Literat...  \n",
       "76   [Trump, Donald J, United States, Presidential Election of 2020, Coronavirus (2019-nCoV), Dexamethasone (Drug), Obesi...  \n",
       "102  [Wegener, Alfred, Institute for Polar and Marine Research, Science and Technology, Research, Workplace Hazards and V...  "
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Looking at false negatives\n",
    "wrong_predictions[wrong_predictions['proba'] < 0.4].head(10).sort_values(by='proba')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Correct Prediction Analysis"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [],
   "source": [
    "predictions = pd.merge(predictions, init_test, left_index=True, right_index=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [],
   "source": [
    "predictions = predictions[['actual', 'predicted', 'proba', 'newsdesk', 'section', 'subsection',\n",
    "                                       'headline', 'abstract', 'keywords']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<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>actual</th>\n",
       "      <th>predicted</th>\n",
       "      <th>proba</th>\n",
       "      <th>newsdesk</th>\n",
       "      <th>section</th>\n",
       "      <th>subsection</th>\n",
       "      <th>headline</th>\n",
       "      <th>abstract</th>\n",
       "      <th>keywords</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1087</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>OpEd</td>\n",
       "      <td>Opinion</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Trump’s Army of Angry White Men</td>\n",
       "      <td>This group will continue to fight for Trump and he knows that.</td>\n",
       "      <td>[Presidential Election of 2020, Polls and Public Opinion, Whites, Race and Ethnicity, Right-Wing Extremism and Alt-R...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2146</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.998882</td>\n",
       "      <td>Games</td>\n",
       "      <td>Crosswords &amp; Games</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Baked Treat on a Stick</td>\n",
       "      <td>Kyle Dolan’s puzzle grows on us.</td>\n",
       "      <td>[Crossword Puzzles]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1054</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.998333</td>\n",
       "      <td>OpEd</td>\n",
       "      <td>Opinion</td>\n",
       "      <td>NaN</td>\n",
       "      <td>The Shrinking of the American Mind</td>\n",
       "      <td>What wasn’t said at the debates was as telling as what was.</td>\n",
       "      <td>[Presidential Election of 2020, European Union, Biden, Hunter, Trump, Donald J, North Atlantic Treaty Organization, ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>663</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.996100</td>\n",
       "      <td>OpEd</td>\n",
       "      <td>Opinion</td>\n",
       "      <td>NaN</td>\n",
       "      <td>How the G.O.P. Can Still Wreck America</td>\n",
       "      <td>Even if Trump loses, his party can do immense damage.</td>\n",
       "      <td>[Presidential Election of 2020, State Legislatures, Elections, Senate, Global Warming, National Debt (US), Coronavir...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1640</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.996069</td>\n",
       "      <td>OpEd</td>\n",
       "      <td>Opinion</td>\n",
       "      <td>NaN</td>\n",
       "      <td>We Waited in Vain for a Repudiation That Never Came</td>\n",
       "      <td>Trump may well end up losing to Biden, but Trumpism remains a viable political strategy.</td>\n",
       "      <td>[Presidential Election of 2020, United States Politics and Government, Black People, Blacks, Race and Ethnicity, Dem...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      actual  predicted     proba newsdesk             section subsection  \\\n",
       "1087     1.0        1.0  1.000000     OpEd             Opinion        NaN   \n",
       "2146     1.0        1.0  0.998882    Games  Crosswords & Games        NaN   \n",
       "1054     1.0        1.0  0.998333     OpEd             Opinion        NaN   \n",
       "663      1.0        1.0  0.996100     OpEd             Opinion        NaN   \n",
       "1640     1.0        1.0  0.996069     OpEd             Opinion        NaN   \n",
       "\n",
       "                                                 headline  \\\n",
       "1087                      Trump’s Army of Angry White Men   \n",
       "2146                               Baked Treat on a Stick   \n",
       "1054                   The Shrinking of the American Mind   \n",
       "663                How the G.O.P. Can Still Wreck America   \n",
       "1640  We Waited in Vain for a Repudiation That Never Came   \n",
       "\n",
       "                                                                                      abstract  \\\n",
       "1087                            This group will continue to fight for Trump and he knows that.   \n",
       "2146                                                          Kyle Dolan’s puzzle grows on us.   \n",
       "1054                               What wasn’t said at the debates was as telling as what was.   \n",
       "663                                      Even if Trump loses, his party can do immense damage.   \n",
       "1640  Trump may well end up losing to Biden, but Trumpism remains a viable political strategy.   \n",
       "\n",
       "                                                                                                                     keywords  \n",
       "1087  [Presidential Election of 2020, Polls and Public Opinion, Whites, Race and Ethnicity, Right-Wing Extremism and Alt-R...  \n",
       "2146                                                                                                      [Crossword Puzzles]  \n",
       "1054  [Presidential Election of 2020, European Union, Biden, Hunter, Trump, Donald J, North Atlantic Treaty Organization, ...  \n",
       "663   [Presidential Election of 2020, State Legislatures, Elections, Senate, Global Warming, National Debt (US), Coronavir...  \n",
       "1640  [Presidential Election of 2020, United States Politics and Government, Black People, Blacks, Race and Ethnicity, Dem...  "
      ]
     },
     "execution_count": 60,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "predictions.sort_values(by='proba', ascending=False).head(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<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>actual</th>\n",
       "      <th>predicted</th>\n",
       "      <th>proba</th>\n",
       "      <th>newsdesk</th>\n",
       "      <th>section</th>\n",
       "      <th>subsection</th>\n",
       "      <th>headline</th>\n",
       "      <th>abstract</th>\n",
       "      <th>keywords</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>3173</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.014920</td>\n",
       "      <td>Culture</td>\n",
       "      <td>Arts</td>\n",
       "      <td>Television</td>\n",
       "      <td>Jimmy Fallon: Trump’s Hanukkah Party Was a ‘Festival of Lies’</td>\n",
       "      <td>“Even the dreidel was looking at Trump like, ‘Damn, this guy spins more than I do,’” Fallon joked on Wednesday.</td>\n",
       "      <td>[Colbert, Stephen, Noah, Trevor (1984- ), Fallon, Jimmy, Trump, Donald J, The Tonight Show (TV Program), The Daily S...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3547</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.011199</td>\n",
       "      <td>RealEstate</td>\n",
       "      <td>Real Estate</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Homes That Sold for Around $600,000</td>\n",
       "      <td>Recent residential sales in New York City and the region.</td>\n",
       "      <td>[Real Estate and Housing (Residential), NoMad (Manhattan, NY), Norwalk (Conn), Midwood (Brooklyn, NY), Mamaroneck (N...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1647</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.010915</td>\n",
       "      <td>Podcasts</td>\n",
       "      <td>Podcasts</td>\n",
       "      <td>The Daily</td>\n",
       "      <td>Joe Biden Has the Edge</td>\n",
       "      <td>The Democratic candidate won Michigan and Wisconsin and seems to be well placed in other key states. President Trump...</td>\n",
       "      <td>[United States Politics and Government, Presidential Election of 2020, Absentee Voting, States (US), Trump, Donald J...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>412</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.008919</td>\n",
       "      <td>RealEstate</td>\n",
       "      <td>Real Estate</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Homes That Sold for Around $625,000</td>\n",
       "      <td>Recent residential sales in New York City and the region.</td>\n",
       "      <td>[Real Estate and Housing (Residential), Hastings-on-Hudson (NY), Midtown Area (Manhattan, NY), Park Slope (Brooklyn,...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>976</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.008611</td>\n",
       "      <td>Metro</td>\n",
       "      <td>New York</td>\n",
       "      <td>NaN</td>\n",
       "      <td>New York Lockdown Rules: What to Know</td>\n",
       "      <td>Schools and nonessential business can reopen in some zones that were deemed virus hot spots, but  restrictions remai...</td>\n",
       "      <td>[New York City, Coronavirus Reopenings, Brooklyn (NYC), Lawrence, Jacob, Queens (NYC), Cuomo, Andrew M, Metropolitan...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      actual  predicted     proba    newsdesk      section  subsection  \\\n",
       "3173     0.0        0.0  0.014920     Culture         Arts  Television   \n",
       "3547     0.0        0.0  0.011199  RealEstate  Real Estate         NaN   \n",
       "1647     0.0        0.0  0.010915    Podcasts     Podcasts   The Daily   \n",
       "412      0.0        0.0  0.008919  RealEstate  Real Estate         NaN   \n",
       "976      0.0        0.0  0.008611       Metro     New York         NaN   \n",
       "\n",
       "                                                           headline  \\\n",
       "3173  Jimmy Fallon: Trump’s Hanukkah Party Was a ‘Festival of Lies’   \n",
       "3547                            Homes That Sold for Around $600,000   \n",
       "1647                                         Joe Biden Has the Edge   \n",
       "412                             Homes That Sold for Around $625,000   \n",
       "976                           New York Lockdown Rules: What to Know   \n",
       "\n",
       "                                                                                                                     abstract  \\\n",
       "3173          “Even the dreidel was looking at Trump like, ‘Damn, this guy spins more than I do,’” Fallon joked on Wednesday.   \n",
       "3547                                                                Recent residential sales in New York City and the region.   \n",
       "1647  The Democratic candidate won Michigan and Wisconsin and seems to be well placed in other key states. President Trump...   \n",
       "412                                                                 Recent residential sales in New York City and the region.   \n",
       "976   Schools and nonessential business can reopen in some zones that were deemed virus hot spots, but  restrictions remai...   \n",
       "\n",
       "                                                                                                                     keywords  \n",
       "3173  [Colbert, Stephen, Noah, Trevor (1984- ), Fallon, Jimmy, Trump, Donald J, The Tonight Show (TV Program), The Daily S...  \n",
       "3547  [Real Estate and Housing (Residential), NoMad (Manhattan, NY), Norwalk (Conn), Midwood (Brooklyn, NY), Mamaroneck (N...  \n",
       "1647  [United States Politics and Government, Presidential Election of 2020, Absentee Voting, States (US), Trump, Donald J...  \n",
       "412   [Real Estate and Housing (Residential), Hastings-on-Hudson (NY), Midtown Area (Manhattan, NY), Park Slope (Brooklyn,...  \n",
       "976   [New York City, Coronavirus Reopenings, Brooklyn (NYC), Lawrence, Jacob, Queens (NYC), Cuomo, Andrew M, Metropolitan...  "
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "predictions.sort_values(by='proba', ascending=False).tail(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Model Insights"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Generally, our most important features regardless of model include newsdesk and section popularity. It goes to show that you can't ignore where an article is posted or placed when it comes to predicting popularity."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Random Forest Feature Importance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.393164Z",
     "start_time": "2021-01-21T08:04:19.952Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:>"
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x864 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(10,12))\n",
    "rf_feature_imp = pd.Series(rf_gs[1][1].feature_importances_, index=X.columns).sort_values(ascending=False)\n",
    "rf_feature_imp.sort_values().plot(kind = 'barh')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### XGBoost Feature Importance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.394165Z",
     "start_time": "2021-01-21T08:04:19.955Z"
    }
   },
   "outputs": [],
   "source": [
    "scores_dict = xgb_gs.steps[1][1]._Booster.get_score(importance_type='gain')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [],
   "source": [
    "total_cover = pd.DataFrame(xgb_gs.steps[1][1]._Booster.get_score(importance_type='total_cover'), index=[0]) \\\n",
    "            .T.reset_index()\n",
    "total_cover.columns = ['feature', 'total cover']\n",
    "total_cover = total_cover.sort_values(by='total cover', ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [],
   "source": [
    "total_gain = pd.DataFrame(xgb_gs.steps[1][1]._Booster.get_score(importance_type='total_gain'), index=[0]) \\\n",
    "            .T.reset_index()\n",
    "total_gain.columns = ['feature', 'total gain']\n",
    "total_gain = total_gain.sort_values(by='total gain', ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.395166Z",
     "start_time": "2021-01-21T08:04:19.956Z"
    }
   },
   "outputs": [],
   "source": [
    "gain_df = pd.DataFrame(xgb_gs.steps[1][1]._Booster.get_score(importance_type='gain'), index=[0]) \\\n",
    "            .T.reset_index()\n",
    "gain_df.columns = ['feature', 'gain']\n",
    "gain_df = gain_df.sort_values(by='gain', ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.395166Z",
     "start_time": "2021-01-21T08:04:19.958Z"
    }
   },
   "outputs": [],
   "source": [
    "weight_df = pd.DataFrame(xgb_gs.steps[1][1]._Booster.get_score(importance_type='weight'), index=[0]) \\\n",
    "            .T.reset_index()\n",
    "weight_df.columns = ['feature', 'weight']\n",
    "weight_df = weight_df.sort_values(by='weight', ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.396165Z",
     "start_time": "2021-01-21T08:04:19.960Z"
    }
   },
   "outputs": [],
   "source": [
    "cover_df = pd.DataFrame(xgb_gs.steps[1][1]._Booster.get_score(importance_type='cover'), index=[0]) \\\n",
    "            .T.reset_index()\n",
    "cover_df.columns = ['feature', 'cover']\n",
    "cover_df = cover_df.sort_values(by='cover', ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.397165Z",
     "start_time": "2021-01-21T08:04:19.962Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x864 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(8,12))\n",
    "sns.barplot(data=gain_df, y='feature', x='gain', orient='h', palette='Reds_r')\n",
    "plt.ylabel('');\n",
    "plt.title('Gain', fontsize=18);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.397165Z",
     "start_time": "2021-01-21T08:04:19.964Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x864 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(8,12))\n",
    "sns.barplot(data=weight_df, y='feature', x='weight', orient='h', palette='Blues_r')\n",
    "plt.ylabel('');\n",
    "plt.title('Weight', fontsize=18);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-21T08:04:51.398165Z",
     "start_time": "2021-01-21T08:04:19.965Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAkQAAALNCAYAAADOX35KAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8vihELAAAACXBIWXMAAAsTAAALEwEAmpwYAABmuklEQVR4nO3deZxdVYHt8d8iAQIkzEMjCEFGCZBACjBMBkRUHAANHQER0IbGtkW00eYpStCmFUVpAQUDQphEGgUEUQYZkjAFKnOYFXCCZp4CMiXr/XF34aWo4Valkpuqs76fT33uufvss4ebfrLe3vtWyTYRERERVbZMswcQERER0WwJRBEREVF5CUQRERFReQlEERERUXkJRBEREVF5CUQRERFReQlEERERUXkJRBFReZJWlHS0pKmSnpH0uqTHJf1W0qGSBjd7jBGxeOX/kUdEpUnaBLga2Az4PfAd4ClgbWBP4FxgS+CrzRpjRCx+ym+qjoiqkrQCMBPYGBhv+7IO6mwPbG/7J0t6fD0haZjtF5s9joj+KltmEVFl/wJsDvygozAEYPuu9mFI0r6SbpU0v/zcKmmfdnWmlW23t63ES/qAJEs6uq5Mkj4nabqklyW9KOkmSbu3e3Z4eXaCpPGl/t+B03r/MUREAlFEVNm48jqx0Qck/RtwObA68F/At8v1FZKOqKt6HrVttw920MyngTeAn9eVXQCcDvyB2vbcBGAV4HpJH+ugjX2BM4BrgKOA3zU6h4h4u2yZRURlSXoaWNb2yg3WXw34C/B/wHa2XyjlK1PbelsbeKft5yStDjwG/Nr2P9e1Maw8f4Ptj5Wy/YDLgH+1PbGu7mDgDmAN4F22LWk48DC1QLWN7XsX5TOIiJqsEEVEla0MvNCD+u8HVgJObQtDAOX6NGAotYPY2H4GuAr4mKRV69oYB6xIbQWpzaeAF6mtMq3Z9gOsWtoYDmzabixXJwxF9J0EooiosheAYT2ov1F5vbuDe/PK67vqys4Hlgf+ua7s08CzwG/qyt5dxvE48GS7nwmlzjrt+nugB+OOiG7ka/cRUWXzgN0kvcv2Qw3UVw/b/y21UPNpYKKkDYD3AmfafrVdu08CB3Yz1nov93AsEdGFBKKIqLJfAbtR+7bZ1xqo/8fyOgK4od29Lcvrm8HK9huSfg58UdK7gAOohZ/z2j37ILXfg3SH7fk9mkFE9IlsmUVElZ0N3A8c0/5r820kjS7fLAO4HngJ+EI5HN1WZxjwBWB+qVOvLfx8GjgYuN/2tHZ1zqf2v8ff6WQM7bfLIqKPZYUoIirL9suSPkLtN1VfIek6aoHmaWAtYHfgA8D3Sv3nJH0V+DEwTdKk0tShwCbUviX2fLs+ZkqaC3yJ2iHut61E2f6lpHOBf5e0HbXzRU8B6wNjStvvav9cRPSdfO0+IipP0orAvwKfoLYdNhR4BmgFfgH83PaCuvr7AV8BRpai2cD3bF/RSfv/AZwMLASG2/5LJ/UOBo4o7S5H7ev5M4BLbF9S6gyn9rX7E2xP6O2cI+KtEogiIiKi8nKGKCIiIiovgSgiIiIqL4EoIiIiKi+BKCIiIiovX7uviDXXXNPDhw9v9jAiIiKWmOnTpz9le61G6iYQVcTw4cNpbW1t9jAiIiKWGEl/arRutswiIiKi8vJ7iCpi0MrLe8h71mv2MCIiIt7mpesa+dvKPSdpuu2WRupmhSgiIiIqL4EoIiIiKi+BKCIiIiovgSgiIiIqL4EoIiIiKi+BKCIiIiovgagbkiZJGteL5yZIOmZxjCkiIiL6VgJRREREVF6/CkSShku6V9JZku6WdJ2kFSRtLOkaSdMlTZW0haRBkh5SzaqSFkrarbQzVdImkt4raVb5mSlpWKl/uqR7JF0NrF3X/2hJk0s/10pat5QfVerPkfSLDsZ9uKTfSVqhk3ndLOl/JN0maZ6kHUr56pKuKO3eIWmbUj5B0gWSbpT0oKTDO2n3CEmtklr9+oJF/vwjIiIGqv74t8w2BQ6wfbik/wU+ARwGHGn7QUk7Aj+xvYekB4AtgY2A6cCukqYB69v+g6RTgM/bvlXSUOAVYD9gc2BrYB3gHuAcScsCpwH72H5S0njgROAzwLHARrZflbRq/WAl/TuwF7Cv7Ve7mNdKtncqoe0cYCvgBGCm7X0l7QGcD4wq9bcB3gOsBMyUdLXtR+sbtD0RmAi131Td6AccERFRNf0xED1se1a5ng4MB3YCLpXUVmf58joV2I1aIPoOcDgwGbir3L8V+KGki4DLbP+1BJKLbS8AHpV0Y6m7ObWQcn3pZxDwWLk3B7hI0hXAFXVjPRj4K7Uw9Ho387oYwPYUSSuXYLULtcCH7RslrSFplVL/17b/Dvxd0k3ADu36joiIiAb1qy2zon6VZQGwOvCc7VF1P+8u96cCu1ILC78FVgXGAlMAbH8X+BdgBeAOSVuU5zpaTRFwd10fW9veq9z7MPBjYDQwXVJb0JxHLbCt38C82vfp0mdn9TqqHxEREb3QHwNRey8AD0vaH6CcARpZ7k2jtnq00PYrwCzgX6kFJSRtbHuu7ZOAVmALamHpk+UM0rrA7qWt+4G1JI0pzy4raYSkZYB32r4J+Cq10DW0PDOz9HelpHd0M4/xpd1dgOdtP1/GclApHws8ZfuFUn8fSUMkrUEt5N3VvsGIiIhoTH/cMuvIQcAZko4DlgV+AcwuZ3r+AtxR6k0FDgDmlvdHS9qd2krTPcDvgNeAPUqdB6htsWH7tfL1+1PLttVg4H9KnQtLmYBTbD/Xtn1n+5by9furJb3f9lOdzOFZSbcBK1M7lwQwAThX0hzgZeCQuvp3AlcDGwDfbn9+KCIiIhonOzstzSbpZuAY260N1p8AzLd9cqN9DFp5eQ95z3q9G2BERMRi9NJ1Dy2WdiVNt93SSN2BsGUWERERsUgGypZZvyDpx8DO7Yp/ZHtsT9qxPaGvxhQREREJREuU7c83q+9tN9ua1usa2pGLiIionGyZRUREROUlEEVERETlJRBFRERE5SUQRUREROXl9xBVxKBVh3iFsRs2exgR0Uvzr7i/2UOI6Hfye4giIiIieiCBKCIiIiovgSgiIiIqL4EoIiIiKi+BKCIiIiovgSgiIiIqb4kHIkm3dVI+SdK4XrY5QdIxizayN9s6VNLpfdFWX5M0VtJvmj2OiIiIgWaJByLbOy3pPvsrSYOaPYaIiIgqaMYK0fzyKkmnS7pH0tXA2nV1RkuaLGm6pGslrVvKD5d0l6TZkn4lacUG+7xZ0kmS7pT0gKRdG3zuw5Jul7SmpL3K9QxJl0oaKul9ki6vq/9+SZdJ+mdJPyxlX5T0ULneWNIt5fp9kmZKmivpHEnLl/JHJH2z1Ntf0gcl3Vfef7yur/dKmlV+Zkoa1sH4j5DUKqnVry1oZMoRERGV1MwzRPsBmwNbA4cDOwFIWhY4DRhnezRwDnBieeYy29vbHgncC3y2B/0Ntr0DcDRwfHeVJe0HHAvsXYqOA/a0vR3QCnwZuBF4t6S1Sp3DgHOBKUBb6NoVeFrSesAuwFRJQ4BJwHjbWwODgc/Vdf+K7V2AK4CzgI+Wdv6prs4xwOdtjyr3/t5+DrYn2m6x3aLlstgUERHRmWYGot2Ai20vsP0otXABtZC0FXC9pFnUgsj65d5WkqZKmgscBIzoQX+XldfpwPBu6u4O/CfwYdvPAu8BtgRuLWM6BNjQtb97cgHwKUmrAmOA39n+P2BoWbV5J/DzMt9dgalljg/bfqD0d1653+aS8rpFqfdg6evCujq3Aj+UdBSwqu03Gv0gIiIi4q0GN7n/jv6QmoC7bY/p4N4kYF/bsyUdCoztQV+vltcFdD/vh4B3AZtRWw0ScL3tAzqoey5wFfAKcGldMLmd2orR/dRC0GeoBab/ADbqpv+X6q47/GNztr9bthr3Bu6QtKft+7ppNyIiIjrQzBWiKcAnJQ0qZ4R2L+X3A2tJGgO1LTRJbStBw4DHyrbaQYtxbH+idl7n/NL3HcDOkjYpY1pR0mYAZXXrUWorWZPaze+Y8jqT2vxetf08cB8wvK094GBgcgfjuA/YSNLG5f2bgUzSxrbn2j6JWmjbYpFnHRERUVHNDESXAw8Cc4EzKIHA9mvAOOAkSbOBWZTzRcA3gGnA9dTCwmJj+35qoetSYGXgUOBiSXOoBaT6AHIR8Bfb99SVTaW2XTbF9gLgL8Atpe1XqK0eXVq2/xYCZ3YwhleAI4Cry6HqP9XdPlrSvPIZ/R343SJPOiIioqJUO5oSi6L83qKZtn/W7LF0ZtCqQ7zC2A2bPYyI6KX5V9zf7CFE9DuSpttuaaRus88Q9XuSplM78/MfzR5LRERE9M6ACkSSfgzs3K74R7bP7aDuB4CT2hU/bHu/nvRZfjVARERE9GMDKhDZ/nwP6l4LXLsYhxMRERH9xIAKRNG5bTfZitYrWps9jIiIiKVS/tp9REREVF4CUURERFReAlFERERUXn4PUUUMWmMFr/SBjbuvGBFLnRd+Pq/ZQ4jol3rye4iyQhQRERGVl0AUERERlZdAFBEREZWXQBQRERGVl0AUERERlZdAFBEREZU3IAKRpAmSjllMbY+StHfd+49JOnZx9BURERHNMSAC0WI2CngzENm+0vZ3mzeciIiI6GtLbSCStJKkqyXNljRP0nhJj0has9xvkXRz3SMjJd0o6UFJh5c660qaImlWaWPXUr6XpNslzZB0qaShpXx7SbeVPu+UtArwLWB8aWO8pEMlnV7qbyjpBklzyusGpXySpFNLWw9JGtfFPMeWMV4u6R5JZ0paptw7QNLcMvaT6p6ZL+kHZfw3SFqrLz/7iIiIqllqAxHwQeBR2yNtbwVc0039bYAPA2OAb0p6B3AgcK3tUcBIYFYJVMcBe9reDmgFvixpOeAS4Iu2RwJ7Ai8B3wQusT3K9iXt+jwdON/2NsBFwKl199YFdgE+AnS3orQD8B/A1sDGwMfL+E8C9qC2SrW9pH1L/ZWAGWX8k4HjO2pU0hGSWiW1+pUF3QwhIiKiupbmQDQX2FPSSZJ2tf18N/V/bfvvtp8CbqIWMu4CDpM0Adja9ovAe4AtgVslzQIOATYENgces30XgO0XbL/RTZ9jgJ+X6wuoBaA2V9heaPseYJ1u2rnT9kO2FwAXl3a2B262/WQZx0XAbqX+QmrhDeDCdv2+yfZE2y22WzRkUDdDiIiIqK7BzR5AZ2w/IGk0tfM735F0HfAG/whxQ9o/8vYmPEXSbtRWji6Q9H3gWeB62wfUV5a0TQdt9HjYddev1jffg+fa3nf3TFfPR0RERA8stStEZcvoZdsXAicD2wGPAKNLlU+0e2QfSUMkrQGMBe6StCHwhO2zgJ+VNu4Adpa0SelnRUmbAfcB75C0fSkfJmkw8CIwrJNh3gZ8slwfBNzSy+nuIGmjcnZofGlnGvBeSWtKGgQcQG17DGr/bm3nkg5chH4jIiKCpXiFiNp5mu9LWgi8DnwOWAH4maSvUQsM9e4ErgY2AL5t+1FJhwBfkfQ6MB/4tO0nJR0KXCxp+fLscWVFajxwmqQVgL9TO0d0E3Bs2V77Trs+jwLOkfQV4EngsF7O9XZq54y2BqYAl9teKOn/lf4F/Nb2r0v9l4ARkqYDz1MLUREREdFLsrPb0kySxgLH2P5ID56Zb3toT/oZtMYKXukDG/dwdBGxNHjh5/OaPYSIfknSdNstjdRdarfMIiIiIpaUpXnLbECRtDW1b6LVe9X2jsDNPWmrp6tDERER0bUEoiXE9lxqv08oIiIiljIJRBWx7UYjaP15a7OHERERsVTKGaKIiIiovASiiIiIqLwEooiIiKi8BKKIiIiovPxixooYvOaKXnmfLZo9jIgB6ZmfzWj2ECKiA/nFjBERERE9kEAUERERlZdAFBEREZWXQBQRERGVl0AUERERlZdAFBEREZWXQNQASbf14pmvLY6xRERERN9LIGqA7Z168ViHgUg1+dwjIiKWIvkPcwMkzS+v60qaImmWpHmSdu2k/neBFUq9iyQNl3SvpJ8AM4B3trVZ6o+TNKlcT5J0hqSbJD0k6b2SzinPT6ofk6QfSJoh6QZJa3UwjiMktUpq9Stv9OlnEhERMZAkEPXMgcC1tkcBI4FZHVWyfSzwd9ujbB9UijcHzre9re0/ddPPasAewJeAq4BTgBHA1pJGlTorATNsbwdMBo7vYBwTbbfYbtGQwY3PMiIiomISiHrmLuAwSROArW2/2INn/2T7jgbrXuXa31SZCzxue67thcDdwPBSZyFwSbm+ENilB2OJiIiIOglEPWB7CrAb8DfgAkmf7sHjL7Vvru56SLt7r5bXhXXXbe87W+rJH6WLiIjopQSiHpC0IfCE7bOAnwHbdVH9dUnLdnH/cUnvLges9+vFcJYBxpXrA4FbetFGRERE0PlqQ3RsLPAVSa8D84GuVogmAnMkzQC+3sH9Y4HfAH8B5gFDeziWl4ARkqYDzwPje/h8REREFKodVYn+RtJ82w2HqMFrruiV99licQ4porKe+dmMZg8hIjogabrtlkbqZsssIiIiKi9bZotI0jRg+XbFB9ueuzj77cnqUERERHQtgWgR2d6x2WNoxKjhW9L6s9ZmDyMiImKplC2ziIiIqLwEooiIiKi8BKKIiIiovASiiIiIqLz8HqKKWG7toV77n7dp9jAilhp/Pf22Zg8hIhaz/B6iiIiIiB5IIIqIiIjKSyCKiIiIyksgioiIiMpLIIqIiIjKSyCKiIiIyksg6oak4ZIOrHvfIunUZo4pIiIi+lYCUfeGA28GItutto9q3nAiIiKirw34QCRpJUlXS5otaZ6k8ZJGS5osabqkayWtW+puIun3pe4MSRsD3wV2lTRL0pckjZX0m1J/dUlXSJoj6Q5J25TyCZLOkXSzpIckdRqgygrUfZLOK+38UtKK5d77JM2UNLe0t3wpf0TSSZLuLD+bdNL2EZJaJbUu/PvrffvBRkREDCADPhABHwQetT3S9lbANcBpwDjbo4FzgBNL3YuAH9seCewEPAYcC0y1Pcr2Ke3aPgGYaXsb4GvA+XX3tgA+AOwAHC9p2S7GuDkwsbTzAvBvkoYAk4DxtrcGBgOfq3vmBds7AKcD/9NRo7Yn2m6x3bLMCl11HxERUW1VCERzgT3LisquwDuBrYDrJc0CjgPWlzQMWM/25QC2X7H9cjdt7wJcUOrfCKwhaZVy72rbr9p+CngCWKeLdv5i+9ZyfWFpd3PgYdsPlPLzgN3qnrm47nVMN+OMiIiILgxu9gAWN9sPSBoN7A18B7geuNv2W0KEpJV70bw66rK8vlpXtoCuP+v2f1DOnbTd2TP5g3QRERGLYMCvEEl6B/Cy7QuBk4EdgbUkjSn3l5U0wvYLwF8l7VvKly9neV4EhnXS/BTgoFJ/LPBUaaenNmgbD3AAcAtwHzC87nzQwcDkumfG173e3os+IyIiohjwK0TA1sD3JS0EXqd2DucN4NSyvTWY2hmcu6mFjp9K+lapuz8wB3hD0mxqZ3pm1rU9AThX0hzgZeCQXo7xXuAQST8FHgTOsP2KpMOASyUNBu4Czqx7ZnlJ06iF2gN62W9EREQAsrPb0kyShgO/KQe+G33mEaClnE9qyHJrD/Xa/7xNzwcYMUD99fTbmj2EiFjMJE233dJI3QG/ZRYRERHRnSpsmS0VJK0B3NDBrff1ZHUIwPbwPhlUREREAAlES4ztp4FRzR5HREREvF0CUUVss8EWtObMRERERIdyhigiIiIqL4EoIiIiKi+BKCIiIiovZ4gqYt5f72eTY3Zv9jAilrg/nHxTs4cQEf1AVogiIiKi8hKIIiIiovISiCIiIqLyEogiIiKi8hKIIiIiovISiCIiIqLyEogKSaMk7d2L594h6Zfd1BkuaV7vRxcRERGLUwLRP4wCehSIJA22/ajtcYtnSBEREbEkDKhAVFZi7pN0tqR5ki6StKekWyU9KGmH8nObpJnldXNJywHfAsZLmiVpvKSVJJ0j6a5Sd5/Sx6GSLpV0FXBd/epPuZ4qaUb52anBcR8q6deSrpF0v6Tj6+59ucxlnqSj283zPElzJP1S0op9/XlGRERUxUD8TdWbAPsDRwB3AQcCuwAfA74GfBrYzfYbkvYE/tv2JyR9E2ix/e8Akv4buNH2ZyStCtwp6feljzHANrafkTS8ru8ngPfbfkXSpsDFQEuD494B2Ap4GbhL0tWAgcOAHQEB0yRNBp4FNgc+a/tWSecA/wacXN+gpCPK58DgYcs3OIyIiIjqGYiB6GHbcwEk3Q3cYNuS5gLDgVWA80pgMbBsJ+3sBXxM0jHl/RBgg3J9ve1nOnhmWeB0SaOABcBmPRj39bafLuO+jFqIM3C57ZfqyncFrgT+YvvW8uyFwFG0C0S2JwITAYb80zD3YCwRERGVMhAD0at11wvr3i+kNt9vAzfZ3q+s7tzcSTsCPmH7/rcUSjsCL3XyzJeAx4GR1LYjX+nBuNsHFpcx9KR+RERE9MKAOkPUoFWAv5XrQ+vKXwSG1b2/FviCJAFI2rbBth+zvRA4GBjUg3G9X9LqklYA9gVuBaYA+0paUdJKwH7A1FJ/A0ljyvUBwC096CsiIiLqVDEQfQ/4jqRbeWtguQnYsu1QNbWVpGWBOeXQ9LcbaPsnwCGS7qC2XdbZSlJHbgEuAGYBv7LdansGMAm4E5gGnG17Zql/b+lrDrA6cEYP+oqIiIg6srPT0mySDqXuQHcD9YcDv7G9VaN9DPmnYV7/U42e744YOP5w8k3NHkJENImk6bYb+o9fFVeIIiIiIt5iIB6qXmpJ+gBwUrvih23vR21rrCG2H6H2Ff2IiIjoAwlES5Dta6kd1o6IiIilSAJRRWy1/ua05ixFREREh3KGKCIiIiovgSgiIiIqL4EoIiIiKi+BKCIiIiovh6or4p6/Pcg2x32w2cOI6JU5/3VNs4cQEQNcVogiIiKi8hKIIiIiovISiCIiIqLyEogiIiKi8hKIIiIiovISiCIiIqLyBnwgkjS/2WPoiqRRkvaue/8xScc2c0wRERFVM+ADUT8wCngzENm+0vZ3mzeciIiI6hlQgUjSFZKmS7pb0hF15T+QNEPSDZLWKmVHSbpH0hxJv+iizTUkXSdppqSfSvqTpDUlDZc0r67eMZImlOuNJV1TxjJV0halfH9J8yTNljRF0nLAt4DxkmZJGi/pUEmnl/obljHPKa8blPJJkk6VdJukhySN62TsR0hqldT6xsuvLfLnGxERMVANqEAEfMb2aKAFOErSGsBKwAzb2wGTgeNL3WOBbW1vAxzZRZvHA7fY3ha4EtiggXFMBL5QxnIM8JNS/k3gA7ZHAh+z/Vopu8T2KNuXtGvndOD8MsaLgFPr7q0L7AJ8BOhwRcn2RNsttlsGr7hcA8OOiIiopoEWiI6SNBu4A3gnsCmwEGgLGhdSCxEAc4CLJH0KeKOLNncrz2H7auDZrgYgaSiwE3CppFnAT6mFF4BbgUmSDgcGNTCfMcDPy/UFdWMHuML2Qtv3AOs00FZERER0YsD8LTNJY4E9gTG2X5Z0MzCkg6ourx+mFnY+BnxD0gjbnQUjd1D2Bm8NlG19LQM8Z3vU2xqxj5S0Y+l7lqS31elG/TherbtWD9uJiIiIOgNphWgV4NkShrYA3lPKlwHaztgcCNwiaRngnbZvAr4KrAoM7aTdKcBBAJI+BKxWyh8H1i5njJantnWF7ReAhyXtX56RpJHlemPb02x/E3iK2irWi8CwTvq+DfhkuT4IuKXRDyMiIiIaN2BWiIBrgCMlzQHup7ZtBvASMELSdOB5YDy17aoLJa1CbXXlFNvPddLuCcDFkmZQO4P0ZwDbr0v6FjANeBi4r+6Zg4AzJB0HLAv8ApgNfF/SpqXPG0rZn4Fjy/bad9r1fRRwjqSvAE8Ch/X0Q4mIiIjuye5oNyg6I+kRoMX2U80eS0+suO4q3uSzY5o9jIhemfNf1zR7CBHRD0mabrulkboDacssIiIiolcG0pbZIpF0GPDFdsW32v58fYHt4UtsUBEREbFEZMusIlpaWtza2trsYURERCwx2TKLiIiI6IEEooiIiKi8BKKIiIiovASiiIiIqLx8y6wi7n3sD+x44r7NHkbEm6Z9/YpmDyEi4k1ZIYqIiIjKSyCKiIiIyksgioiIiMpLIIqIiIjKSyCKiIiIyksgioiIiMpLIOoFSbc1oc+xknZa0v1GRERUQQJRL9heosFE0mBgLJBAFBERsRgkEPWCpPnldV1JUyTNkjRP0q5dPSPpB5JmSLpB0lql/HBJd0maLelXklYs5ZMk/VDSTcAlwJHAl0pfu0p6WNKype7Kkh5pe1/X5xGSWiW1vvHSa4vp04iIiOj/EogWzYHAtbZHASOBWV3UXQmYYXs7YDJwfCm/zPb2tkcC9wKfrXtmM2BP258AzgROsT3K9lTgZuDDpd4ngV/Zfr2+Q9sTbbfYbhm80nKLMM2IiIiBLYFo0dwFHCZpArC17Re7qLuQ2koPwIXALuV6K0lTJc0FDgJG1D1zqe0FnbR3NnBYuT4MOLcX44+IiAgSiBaJ7SnAbsDfgAskfbonj5fXScC/294aOAEYUlfnpS76vhUYLum9wCDb83oy9oiIiPiHBKJFIGlD4AnbZwE/A7brovoywLhyfSBwS7keBjxWzv8c1MXzL5a69c4HLiarQxEREYskgWjRjAVmSZoJfAL4URd1XwJGSJoO7AF8q5R/A5gGXA/c18XzVwH7tR2qLmUXAatRC0URERHRS7Ldfa1YZJLm2x7ax22OA/axfXB3dVdab1Vv9W9j+7L7iEUy7etXNHsIETHASZpuu6WRuoMX92Bi8ZB0GvAhYO9mjyUiIqK/SyDqY5KmAcu3Kz64r1eHbH+hL9uLiIiosgSiPmZ7x2aPISIiInomgagi3r3uJjmzERER0Yl8yywiIiIqL4EoIiIiKi+BKCIiIiovZ4gq4v7HH+K9Pxzf7GFEH5r85Uu6rxQREQ3JClFERERUXgJRREREVF4CUURERFReAlFERERUXgJRREREVF4CUURERFReAlGDJN22iM8fKun0Xj47StLede8nSDpmUcYTERER/5BA1CDbOzWx+1HA3t1VioiIiN5JIGqQpPnldV1JUyTNkjRP0q5dPHOYpAckTQZ2ritfS9KvJN1VfnYu5TtIuk3SzPK6uaTlgG8B40ufbb9dcUtJN0t6SNJRi2/mERERA19+U3XPHQhca/tESYOAFTuqJGld4ARgNPA8cBMws9z+EXCK7VskbQBcC7wbuA/YzfYbkvYE/tv2JyR9E2ix/e+l7QnAFsDuwDDgfkln2H693RiOAI4AWH61DocZERERJBD1xl3AOZKWBa6wPauTejsCN9t+EkDSJcBm5d6e1FZ42uquLGkYsApwnqRNAQPLdjGOq22/Crwq6QlgHeCv9RVsTwQmAgx75+ru0SwjIiIqJFtmPWR7CrAb8DfgAkmf7qp6J+XLAGNsjyo/69l+Efg2cJPtrYCPAkO6aPvVuusFJNxGRET0WgJRD0naEHjC9lnAz4DtOqk6DRgraY2ymrR/3b3rgH+va3NUuVyFWtACOLSu/ovUtsYiIiJiMUgg6rmxwCxJM4FPUDsP9Da2HwMmALcDvwdm1N0+CmiRNEfSPcCRpfx7wHck3QoMqqt/E7UttvpD1REREdFHZOdoSRUMe+fq3u5L72/2MKIPTf7yJc0eQkTEUk3SdNstjdTNClFERERUXg7i9gFJ04Dl2xUfbHtuM8YTERERPZNA1Ads79jsMURERETvJRBVxObrvCtnTiIiIjqRM0QRERFReQlEERERUXkJRBEREVF5OUNUEQ8++TAfPKOrvzISS4trPnd+s4cQEVE5WSGKiIiIyksgioiIiMpLIIqIiIjKSyCKiIiIyksgioiIiMpLIIqIiIjKq0QgknTbEujjSElv+167pOGS5i3OPiIiImLRVOL3ENneaQn0ceZA6CMiIqKKqrJCNL+8ritpiqRZkuZJ2rWLZ/aSdLukGZIulTS0lD8i6SRJd5afTUr5BEnHlOvRkmZLuh34fF2bgyR9X9JdkuZI+tdSPlbSZEn/K+kBSd+VdFBpf66kjTvoYxNJvy/9zGirExERET1XiUBU50DgWtujgJHArI4qSVoTOA7Y0/Z2QCvw5boqL9jeATgd+J8OmjgXOMr2mHblnwWet709sD1wuKSNyr2RwBeBrYGDgc1KH2cDX+igj4uAH9seCewEPNbBPI6Q1Cqp9bX5r3Y01YiIiKAiW2Z17gLOkbQscIXtWZ3Uew+wJXCrJIDlgNvr7l9c93pK/YOSVgFWtT25FF0AfKhc7wVsI2lceb8KsCnwGnCX7cdKG38Erit15gK7t+tjGLCe7csBbL/S0SRsTwQmAqyy4RruZK4RERGVV6lAZHuKpN2ADwMXSPq+7Y7+cJSA620f0FlTnVy3PdtZ+BDwBdvXvqVQGgvUL+EsrHu/kLf/O6mT9iMiIqIXKrVlJmlD4AnbZwE/A7brpOodwM5154NWlLRZ3f3xda/1K0fYfg54XtIupeigutvXAp8rK1RI2kzSSj2dh+0XgL9K2re0s7ykFXvaTkRERNRUaoUIGAt8RdLrwHygw6+w235S0qHAxZKWL8XHAQ+U6+UlTaMWKDtaRTqM2tbcy9RCUJuzgeHADNX24p4E9u3lXA4GfirpW8DrwP7AQ71sKyIiotJk52hJT0h6BGix/VSzx9ITq2y4hscc++FmDyMacM3nOtrFjYiInpI03XZLI3UrtWUWERER0ZGqbZm9Tdn6Wr5d8cG253ZU3/bwxT6oiIiIWKIqH4hs79jsMURERERzVT4QVcWma22UsykRERGdyBmiiIiIqLwEooiIiKi8BKKIiIiovASiiIiIqLwcqq6IPz79J/abdHizh1F5lx96VrOHEBERHcgKUURERFReAlFERERUXgJRREREVF4CUURERFReAlFERERUXgJRREREVF4C0SKQdKik05vY/yhJezer/4iIiIEigagHJA1q9hjaGQUkEEVERCyiygQiSV+VdFS5PkXSjeX6fZIulHSApLmS5kk6qe65+ZK+JWkaMEbSYZIekDQZ2LmbPteRdLmk2eVnp1L+5dLPPElHl7LhkubVPXuMpAnl+mZJJ0m6s/S9q6TlgG8B4yXNkjS+g/6PkNQqqfXVF19ZtA8wIiJiAKtMIAKmALuW6xZgqKRlgV2AB4GTgD2orbpsL2nfUnclYJ7tHYE/AidQC0LvB7bsps9Tgcm2RwLbAXdLGg0cBuwIvAc4XNK2DYx/sO0dgKOB422/BnwTuMT2KNuXtH/A9kTbLbZblh82pIEuIiIiqqlKgWg6MFrSMOBV4HZqwWhX4DngZttP2n4DuAjYrTy3APhVud6xrt5rwNtCSDt7AGcA2F5g+3lqAexy2y/Zng9cxj+CWlcuq5vH8AbqR0RERIMqE4hsvw48Qm115jZgKrA7sDHw5y4efcX2gvqmFnEo6qT8Dd7679F+SefV8rqA/A26iIiIPlWZQFRMAY4pr1OBI4FZwB3AeyWtWQ5OHwBM7uD5acBYSWuU7bb9u+nvBuBzUDuQLWnl0ve+klaUtBKwXxnL48Dape3lgY80MJ8XgWEN1IuIiIguVC0QTQXWBW63/TjwCjDV9mPA/wNuAmYDM2z/uv3Dpd4EatttvwdmdNPfF4HdJc2lttU1wvYMYBJwJ7WAdbbtmWUF61ul7DfAfQ3M5yZgy84OVUdERERjZC/qDlD0B6tttJbHHr9vs4dReZcfelazhxARURmSpttuaaRu1VaIIiIiIt4mh3P7gKSv8/bzRJfaPrEZ44mIiIieyZZZRbS0tLi1tbXZw4iIiFhismUWERER0QMJRBEREVF5CUQRERFReQlEERERUXn5lllFPPzMnzn44i80exj9zgUHnNbsIURExBKQFaKIiIiovASiiIiIqLwEooiIiKi8BKKIiIiovASiiIiIqLwEooiIiKi8JRKIJA2XNG8xtPuIpDXL9fzy+g5Jv+zrvhYnSYdKekfd+7MlbdnMMUVERFTJgPs9RLYfBcY1exw9dCgwD3gUwPa/NHU0ERERFbMkt8wGSTpL0t2SrpO0gqSNJV0jabqkqZK2AJD0UUnTJM2U9HtJ65TyNcqzMyX9FFD7TupXo8rKy2Wljwclfa+u3l6Sbpc0Q9KlkoZ2NnBJH5R0n6RbJJ0q6TelfIKkY+rqzZM0vFx/StKdkmZJ+qmkQeVnUqk3V9KXJI0DWoCLSt0VJN0sqaW0c0CpO0/SSXV9zZd0oqTZku5o+4zajfsISa2SWl998e89+9eKiIiokCUZiDYFfmx7BPAc8AlgIvAF26OBY4CflLq3AO+xvS3wC+Crpfx44JZSfiWwQQP9jgLGA1sD4yW9s2yzHQfsaXs7oBX4ckcPSxoCnAV8FNgV+KfuOpT07tLnzrZHAQuAg8pY1rO9le2tgXNt/7L0f5DtUbb/XtfOO4CTgD3Ks9tL2rfcXgm4w/ZIYApwePtx2J5ou8V2y/LDVuhu2BEREZW1JLfMHrY9q1xPB4YDOwGXSm8u9CxfXtcHLpG0LrAc8HAp3w34OIDtqyU920C/N9h+HkDSPcCGwKrAlsCtpe/lgNs7eX6LMvYHSxsXAkd00+f7gNHAXaX9FYAngKuAd0k6DbgauK6bdrYHbrb9ZOn7ImqfwRXAa8BvSr3pwPu7aSsiIiI6sSQD0at11wuAdYDnygpKe6cBP7R9paSxwIS6e17EfgdT22q73vYBDbbRWZ9v8NZVtiHlVcB5tv9f+wckjQQ+AHwe+GfgM130+7YtwTqv224bV9u8IiIiohea+bX7F4CHJe0PoJqR5d4qwN/K9SF1z0yhtvWEpA8Bq/Wy7zuAnSVtUtpaUdJmndS9D9hI0sblfX2IegTYrrSxHbBRKb8BGCdp7XJvdUkblq26ZWz/CvhG27PAi8CwDvqeBrxX0pqSBpW+J/d4thEREdGlZv8eooOAz0qaDdwN7FPKJ1DbSpsKPFVX/wRgN0kzgL2AP/em07IFdShwsaQ51ALSFp3UfYXaFtnVkm4B/lR3+1fA6pJmAZ8DHijP3EPtjNJ1pf3rgXWB9YCbS/1JQNsK0iTgzLZD1XV9P1bq3ATMBmbY/nVv5hwRERGd0z92XaIRZQvvGNsfafJQemSNd63tvU8c3+xh9DsXHHBas4cQERG9JGm67ZZG6jZ7hSgiIiKi6XIQt46ky/nHOaA2/2n72rY3tm8Gbl6Cw4qIiIjFLIGoju39mj2GiIiIWPISiCpio9U3yHmYiIiITuQMUURERFReAlFERERUXgJRREREVF7OEFXEn5/7K5+//CvNHsZS48f7fb/ZQ4iIiKVIVogiIiKi8hKIIiIiovISiCIiIqLyEogiIiKi8hKIIiIiovISiCIiIqLyBlQgknTbYmz7SEmf7qO25vdFOxEREdE3BtTvIbK90+JoV9Jg22cujrb7iqRBthc0exwRERH90UBbIZpfXteVNEXSLEnzJO3a1TOSfiBphqQbJK1Vym+W9N+SJgNflDRB0jF1904pfdwraXtJl0l6UNJ/1bX9KUl3lnH8VNKgunsnSpot6Q5J65SyDcsY5pTXDUr5JEnjOpjnWEk3Sfo5MLcvP8uIiIgqGVCBqM6BwLW2RwEjgVld1F0JmGF7O2AycHzdvVVtv9f2Dzp47jXbuwFnAr8GPg9sBRwqaQ1J7wbGAzuXcSwADqrr8w7bI4EpwOGl/HTgfNvbABcBpzYw1x2Ar9vesv0NSUdIapXU+vcXXm6gqYiIiGoaUFtmde4CzpG0LHCF7Vld1F0IXFKuLwQuq7t3ydurv+nK8joXuNv2YwCSHgLeCewCjAbukgSwAvBEeeY14Dflejrw/nI9Bvh4ub4A+F4X/be50/bDHd2wPRGYCLD2Jv/kBtqKiIiopAG5QmR7CrAb8Dfggh4ehq4PDi91Ue/V8rqw7rrt/WBAwHm2R5WfzW1PKHVet93WzwI6D6Ztdd6g/Fuplq6Wa3CMERER0YABGYgkbQg8Yfss4GfAdl1UXwZoO59zIHBLHw3jBmCcpLXLmFYv4+rKbcAny/VBdWN5hNpqE8A+wLJ9NMaIiIhg4G6ZjQW+Iul1YD7Q1QrRS8AISdOB56md+1lktu+RdBxwnaRlgNepnTP6UxePHUVtq+8rwJPAYaX8LODXku6kFrSyKhQREdGH9I+dm2qSNN/20GaPY3Fbe5N/8v7fP7jZw1hq/Hi/7zd7CBERsZhJmm67pZG6A3LLLCIiIqInBuqW2dtImgYs36744CqsDkVERETXKhOIbO/Y7DFERETE0qkygajqNlh1/ZybiYiI6ETOEEVERETlJRBFRERE5SUQRUREROUlEEVERETl5VB1Rfzt+Uc59nffaPYwmuq7H/p2s4cQERFLqawQRUREROUlEEVERETlJRBFRERE5SUQRUREROUlEEVERETlJRBFRERE5SUQdUDSzZJamj2OjkiaJGlcs8cRERExkCQQLcUkDWr2GCIiIqqgMoFI0nBJ90o6S9Ldkq6TtEI3zywj6TxJ/yVpkKTvS7pL0hxJ/1rqXCBpn7pnLpL0MUm/lbRNKZsp6Zvl+tuS/kU135c0T9JcSePL/bGSbpL0c2BuqXe6pHskXQ2sXdfXd0v5HEkndzD+IyS1Smp9+YWX+uJjjIiIGJAqE4iKTYEf2x4BPAd8oou6g4GLgAdsHwd8Fnje9vbA9sDhkjYCzgYOA5C0CrAT8FtgCrCrpJWBN4CdS7u7AFOBjwOjgJHAnsD3Ja1b6uwAfN32lsB+wObA1sDhpX0krV7ujbC9DfBf7Sdge6LtFtstK668Ug8+poiIiGqpWiB62Pascj0dGN5F3Z8C82yfWN7vBXxa0ixgGrAGsKntycAmktYGDgB+ZfsNaqFnN2oB6GpgqKQVgeG27y/lF9teYPtxYDK1oAVwp+2Hy/VudfUeBW4s5S8ArwBnS/o48HKvPpGIiIioXCB6te56AV3/LbfbgN0lDSnvBXzB9qjys5Ht68q9C4CDqK0UnVvK7gJagF2prRbNpLbCM72uvc60399y+woldO0A/ArYF7imi/YiIiKiC1ULRD3xM2pbX5dKGgxcC3xO0rIAkjaT1LYPNQk4GsD23eX1NeAvwD8Dd1BbMTqmvEItJI0vZ5PWorYSdGcH45gCfLLUWxfYvfQ/FFjF9m9L36P6auIRERFVk7923wXbPyzngtpWgIYDMyQJeJLaygy2H5d0L3BFuyamAu+z/bKkqcD6/CMQXQ6MAWZTWwH6qu3/k7RFuzYuB/YA5gIPUNtaAxgG/LqsYAn4Ul/MOSIioopkv203JnqonA2aC2xn+/lmj6cj6276Dh9y6mebPYym+u6Hvt3sIURExBIkabrthn6vYLbMFpGkPYH7gNOW1jAUERERXav0lpmkH/OPr8O3+ZHtczuq3xHbvwc26NOBRURExBKVLbOKaGlpcWtra7OHERERscRkyywiIiKiBxKIIiIiovISiCIiIqLyEogiIiKi8ir9LbMqeezFxzjxxv9u9jCa4ut7fK3ZQ4iIiKVcVogiIiKi8hKIIiIiovISiCIiIqLyEogiIiKi8hKIIiIiovISiCIiIqLyKhOIJI2StHfd+49JOnYx9zlW0k6Ls4+IiIhYdJUJRMAo4M1AZPtK299dzH2OBRKIIiIilnL9IhBJWknS1ZJmS5onabyk0ZImS5ou6VpJ65a6N0s6SdKdkh6QtKuk5YBvAeMlzSrPHyrp9PLMJElnSLpJ0kOS3ivpHEn3SppUN469JN0uaYakSyUNLeWPSDqhlM+VtIWk4cCRwJdKn7t2MrdJkk6VdFvpe1zdva9IukvSHEkn1JV/Q9J9kq6XdLGkYzpp+whJrZJaX3rupUX9Z4iIiBiw+kUgAj4IPGp7pO2tgGuA04BxtkcD5wAn1tUfbHsH4GjgeNuvAd8ELrE9yvYlHfSxGrAH8CXgKuAUYASwddluWxM4DtjT9nZAK/DluuefKuVnAMfYfgQ4Ezil9Dm1i/mtC+wCfAT4LtTCF7ApsAO11a3RknaT1AJ8AtgW+DjQ0lmjtifabrHdstKqK3XRfURERLX1lz/dMRc4WdJJwG+AZ4GtgOslAQwCHqurf1l5nQ4Mb7CPq2xb0lzgcdtzASTdXdpYH9gSuLX0uRxweyd9frwHcwO4wvZC4B5J65SyvcrPzPJ+KLWANAz4te2/l/Fd1cO+IiIiop1+EYhsPyBpNLUzQN8Brgfutj2mk0deLa8LaHyObc8srLtuez+4tHW97QP6sM/2zwKo7vU7tn9aX1HSl3rYdkRERHSjX2yZSXoH8LLtC4GTgR2BtSSNKfeXlTSim2ZepLa60lt3ADtL2qT0uaKkzRZjn9cCn6k7p7SepLWBW4CPShpS7n24l+1HRERE0S9WiICtge9LWgi8DnwOeAM4VdIq1ObxP8DdXbRxE3CspFnUVpl6xPaTkg4FLpa0fCk+Dnigi8euAn4paR/gC92cI2rf33WS3g3cXrbo5gOfsn2XpCuB2cCfqJ1ler6n84mIiIh/kO1mjyF6SNJQ2/MlrQhMAY6wPaOrZ9bbfD3/2xmfXzIDXMp8fY+vNXsIERHRBJKm2+70y0f1+ssKUbzVRElbAkOA87oLQxEREdG1BKIlRNLXgf3bFV9q+8SO6nfF9oF9M6qIiIiABKIlpgSfHoefiIiIWPwSiCpi3WHr5ixNREREJ/rF1+4jIiIiFqcEooiIiKi8BKKIiIiovJwhqogn5j/Oj279n2YPo899ceejmz2EiIgYALJCFBEREZWXQBQRERGVl0AUERERlZdAFBEREZWXQBQRERGVl0DUBJKGS5rX7HFERERETQLRACEpv0IhIiKilxKImmeQpLMk3S3pOkkrSBol6Q5JcyRdLmk1AEk3S2op12tKeqRcHyrpUklXAdc1byoRERH9WwJR82wK/Nj2COA54BPA+cB/2t4GmAsc30A7Y4BDbO+xuAYaEREx0CUQNc/DtmeV6+nAxsCqtieXsvOA3Rpo53rbz3R0Q9IRkloltc5/7qVFHnBERMRAlUDUPK/WXS8AVu2i7hv8499qSLt7nSYd2xNtt9huGbrqSr0aZERERBUkEC09ngeelbRreX8w0LZa9AgwulyPW8LjioiIGPDyzaSlyyHAmZJWBB4CDivlJwP/K+lg4MZmDS4iImKgSiBqAtuPAFvVvT+57vZ7Oqh/H7BNXdFxpXwSMGlxjDEiIqJKsmUWERERlZdAFBEREZWXQBQRERGVl0AUERERlZdD1RWx9tB1+OLORzd7GBEREUulrBBFRERE5SUQRUREROUlEEVERETlJRBFRERE5eVQdUU89dKTnH3XT5s9jIb9y/b/2uwhREREhWSFKCIiIiovgSgiIiIqL4EoIiIiKi+BKCIiIiovgSgiIiIqL4EoIiIiKq+ygUjSvpK2bPY4OiLpUEmnN3scERERVVHZQATsCyzRQKSaKn/mERERS6V++R9nScMl3SfpPElzJP1S0oqS3idppqS5ks6RtHyp/11J95S6J0vaCfgY8H1JsyRtLOmoujq/6KLvCZIukHSjpAclHV537yuS7iptnFA31nsl/QSYAbyzk3YPk/SApMnAznXlH5U0rczr95LWkbRM6XutUmcZSX+QtGa7No+Q1Cqp9cXn5vf6846IiBjo+mUgKjYHJtreBngB+DIwCRhve2tqv4X7c5JWB/YDRpS6/2X7NuBK4Cu2R9n+I3AssG2pc2Q3fW8DfBgYA3xT0jsk7QVsCuwAjAJGS9qtbqzn297W9p/aNyZpXeAEakHo/bx15eoW4D22twV+AXzV9kLgQuCgUmdPYLbtp+rbtT3RdovtlmGrDu1mShEREdXVnwPRX2zfWq4vBN4HPGz7gVJ2HrAbtbD0CnC2pI8DL3fS3hzgIkmfAt7opu9f2/57CSA3UQtBe5WfmdRWgragFpAA/mT7ji7a2xG42faTtl8DLqm7tz5wraS5wFeAEaX8HODT5fozwLndjDkiIiI60Z8DkRuqZL9BLbD8itq5oWs6qfph4MfAaGC6pK7+zlv7vg0I+E5ZcRplexPbPyv3X2pkqJ2UnwacXla9/hUYAmD7L8DjkvagFqh+10AfERER0YH+HIg2kDSmXB8A/B4YLmmTUnYwMFnSUGAV278Fjqa2nQXwIjAMamdwgHfavgn4KrAq0NUe0z6ShkhaAxgL3AVcC3ym9Iek9SSt3eBcpgFjJa0haVlg/7p7qwB/K9eHtHvubGqrY/9re0GDfUVEREQ7/fmv3d8LHCLpp8CDwBeBO4BLy+rOXcCZwOrAryUNobaK86Xy/C+AsyQdBXwS+JmkVUqdU2w/10XfdwJXAxsA37b9KPCopHcDt0sCmA98Cug2qNh+TNIE4HbgMWpbboPK7QllTn8r89uo7tErqW2VZbssIiJiEfTnQLTQdvvDzzcA27Yre4zaltlblPNH9YeXd+lB3w/YPqKDNn8E/KiD+lt116DtDoON7V8Dv+7ksZHUDlPf1137ERER0bn+HIgqTdKxwOf4xzfNIiIiopf6ZSCy/QgNrLosCkmHUduGq3er7c8vYrvTgOXbFR9se25P2rH9XeC7izKWiIiIqJHd0Je1op9raWlxa2trs4cRERGxxEiabrulkbr9+VtmEREREX0igSgiIiIqL4EoIiIiKi+BKCIiIiqvX37LLHrumb8/zc9nn9/sYTTkwJGf7r5SREREH8oKUURERFReAlFERERUXgJRREREVF4CUURERFReAlFERERUXgJRREREVF5lApGk25aCMUySNK6P2xwuaV5fthkREVE1lQlEtndq9hgiIiJi6VSZQCRpfnldV9IUSbMkzZO0ayf1/1nSD8v1FyU9VK43lnRLuR4tabKk6ZKulbRuXZ1rSvlUSVt00P63y4rRMpK+IukuSXMknVDuD5d0r6SzJN0t6TpJK9T1O1vS7cDnu5jzEZJaJbW++OyLi/T5RUREDGSVCUR1DgSutT0KGAnM6qTeFKAtLO0KPC1pPWAXYKqkZYHTgHG2RwPnACeW+hOBL5TyY4Cf1Dcs6XvA2sBhwJ7ApsAOwChgtKTdStVNgR/bHgE8B3yilJ8LHGV7TFcTtT3RdovtlmGrDeuqakRERKVV8U933AWcUwLNFbZndVTJ9v9JGippGPBO4OfAbtTC0WXA5sBWwPWSAAYBj0kaCuwEXFrKAZava/obwDTbRwBI2gvYC5hZ7g+lFoT+DDxcN77pwHBJqwCr2p5cyi8APtS7jyIiIiKggoHI9pSyAvNh4AJJ37fd2R/5up3aKs79wFTgM8AY4D+ADYC726/SSFoZeK6sQHXkLmqrQKvbfgYQ8B3bP23XznDg1bqiBcAKpb4bnG5EREQ0oHJbZpI2BJ6wfRbwM2C7LqpPobblNYXaCs7uwKu2n6cWktaSNKa0u6ykEbZfAB6WtH8pl6SRdW1eA3wXuLqsPl0LfKasLCFpPUlrdzYg288Bz0vapRQd1LNPICIiItqr3AoRMBb4iqTXgflAV39afSq17bIpthdI+gtwH4Dt18pX6E8t21iDgf8B7qYWUs6QdBywLPALYHZbo7YvLWHoSmBvattxt5cttvnAp6itCHXmMGrbfi9TC1QRERGxCGRn96UK3jViI//Xz09o9jAacuDIrjJqREREYyRNt93SSN3KbZlFREREtFfFLbO3kTSNt34TDOBg23ObMZ6IiIhYshKIANs7NnsMi9vqK6yRraiIiIhOZMssIiIiKi+BKCIiIiovgSgiIiIqL4EoIiIiKi+HqiviuVee5df3/rLZw2jIPu8e1+whRERExWSFKCIiIiovgSgiIiIqL4EoIiIiKi+BKCIiIiovgSgiIiIqL4EoIiIiKq/fByJJwyXNWwztPiJpzV48N38R+z1U0umL0kZERET0TL8PRBERERGLaqAEokGSzpJ0t6TrJK0gaWNJ10iaLmmqpC0AJH1U0jRJMyX9XtI6pXyN8uxMST8F1FWHkq4obd8t6Yh2934gaYakGyStVcqOknSPpDmSftHIpCStJelXku4qPzuX8gmSzpF0s6SHJB3VyfNHSGqV1PrCMy800mVEREQlDZRAtCnwY9sjgOeATwATgS/YHg0cA/yk1L0FeI/tbYFfAF8t5ccDt5TyK4ENuunzM6XtFuAoSWuU8pWAGba3AyaXdgGOBba1vQ1wZIPz+hFwiu3ty5zOrru3BfABYAfgeEnLtn/Y9kTbLbZbVl595Qa7jIiIqJ6B8qc7HrY9q1xPB4YDOwGXSm8u9CxfXtcHLpG0LrAc8HAp3w34OIDtqyU9202fR0nar1y/k1ooexpYCFxSyi8ELivXc4CLJF0BXNHgvPYEtqybw8qShpXrq22/Crwq6QlgHeCvDbYbERERdQZKIHq17noBtXDwnO1RHdQ9Dfih7SsljQUm1N1zI52V5/YExth+WdLNwJBOqre1+WFqoetjwDckjbD9RjddLVP6+Hu7/uHtcx4o/5YRERFL3EDZMmvvBeBhSfsDqGZkubcK8LdyfUjdM1OAg0r9DwGrddH+KsCzJQxtAbyn7t4yQNtfJz0QuEXSMsA7bd9EbYtuVWBoA/O4Dvj3tjeSRjXwTERERPTQQA1EUAs3n5U0G7gb2KeUT6C2lTYVeKqu/gnAbpJmAHsBf+6i7WuAwZLmAN8G7qi79xIwQtJ0YA/gW8Ag4EJJc4GZ1M4FPdfAHI4CWspB7Hto/OxRRERE9IDshnaJop/bZKuN/YNLT2r2MBqyz7vHdV8pIiKiG5Km225ppO5AXiGKiIiIaEgO4nahfJX+hg5uvc/204vY9mHAF9sV32r784vSbkRERPRcAlEXSugZtZjaPhc4d3G0HRERET2TQFQRqw5ZLWdzIiIiOpEzRBEREVF5CUQRERFReQlEERERUXk5Q1QRL7z6PNf/4bfNHka33r/J3s0eQkREVFBWiCIiIqLyEogiIiKi8hKIIiIiovISiCIiIqLyEogiIiKi8hKIIiIiovIGXCCSNFzSvMXQ7iOS1izX88vrOyT9so/7uVlSS1+2GREREV3L7yFaBLYfBfIHwiIiIvq5AbdCVAySdJakuyVdJ2kFSRtLukbSdElTJW0BIOmjkqZJminp95LWKeVrlGdnSvopoPad1K9GSTpU0mWljwclfa+u3l6Sbpc0Q9KlkoY2MonOniurVSeU8rltc4mIiIjeGaiBaFPgx7ZHAM8BnwAmAl+wPRo4BvhJqXsL8B7b2wK/AL5ayo8HbinlVwIbNNDvKGA8sDUwXtI7yzbbccCetrcDWoEvd9dQA889VcrPKPPpqI0jJLVKan3+mecbGH5EREQ1DdQts4dtzyrX04HhwE7ApdKbCz3Ll9f1gUskrQssBzxcyncDPg5g+2pJzzbQ7w22nweQdA+wIbAqsCVwa+l7OeD2Btp6TzfPXVY3v4931IDtidSCIJttvakb6DMiIqKSBmogerXuegGwDvCc7VEd1D0N+KHtKyWNBSbU3etpiGjf72BqW23X2z6gh21191xbX239RERERC8N1C2z9l4AHpa0P4BqRpZ7qwB/K9eH1D0zBTio1P8QsFov+74D2FnSJqWtFSVtthifi4iIiB6qSiCCWrj5rKTZwN3APqV8ArWttKnAU3X1TwB2kzQD2Av4c286tf0kcChwsaQ51IJOt4ege/tcRERE9JzsHC2pgs223tQ/vvxHzR5Gt96/yd7NHkJERAwQkqbbbuh3+1VphSgiIiKiQzmM2ySSLgc2alf8n7avbcZ4IiIiqiyBqEls79fsMURERERNAlFFrLz8KjmfExER0YmcIYqIiIjKSyCKiIiIyksgioiIiMpLIIqIiIjKy6Hqipj/2ovc8qebmj2MN+2y4e7NHkJERMSbskIUERERlZdAFBEREZWXQBQRERGVl0AUERERlZdAFBEREZWXQBQRERGVV7lAJGl+D+qOlbRTH/Z9tKQVu6nziKQ1+6rPiIiI6F7lAlEPjQU6DESSevM7nI4GugxEERERseQN6EAk6QpJ0yXdLemIuvIfSJoh6QZJa5WyoyTdI2mOpF9IGg4cCXxJ0ixJu0qaJOmHkm4CTpK0g6TbJM0sr5uXtgZJOlnS3NLeFyQdBbwDuKk838j4PyXpztL/TyUNKuXzJZ0oabakOySt08nzR0hqldT63DPPL8pHGRERMaAN6EAEfMb2aKAFOErSGsBKwAzb2wGTgeNL3WOBbW1vAxxp+xHgTOAU26NsTy31NgP2tP0fwH3Abra3Bb4J/HepcwSwUV17F9k+FXgU2N12t7+mWdK7gfHAzrZHAQuAg8rtlYA7bI8EpgCHd9SG7Ym2W2y3rLr6Kt1+WBEREVU10P90x1GS9ivX7wQ2BRYCl5SyC4HLyvUc4CJJVwBXdNHmpbYXlOtVgPMkbQoYWLaU7wmcafsNANvP9GLs7wNGA3dJAlgBeKLcew34TbmeDry/F+1HREREMWADkaSx1ILJGNsvS7oZGNJBVZfXDwO7AR8DviFpRCdNv1R3/W3gJtv7lS22m9u6r2u3twScZ/v/dXDvddtt7S9gAP87RkRELAkDectsFeDZEoa2AN5TypcBxpXrA4FbJC0DvNP2TcBXgVWBocCLwLBu+vhbuT60rvw64Mi2g9eSVi/l3bVX7wZgnKS129qQtGGDz0ZEREQPDORAdA0wWNIcais5d5Tyl4ARkqYDewDfAgYBF0qaC8ykdm7oOeAqYL+2Q9Ud9PE94DuSbi1ttDkb+DMwR9JsasELYCLwu0YOVdu+BzgOuK7M4Xpg3YZnHxEREQ3TP3ZeYiDbYpvNffZVZzZ7GG/aZcNuz5VHREQsEknTbbc0UncgrxBFRERENCSHcZtE0jRg+XbFB9ue24zxREREVFkCUZPY3nFJ9jd0uWHZpoqIiOhEtswiIiKi8hKIIiIiovISiCIiIqLyEogiIiKi8nKouiJefv0lZjw6rdnDeNN271iiZ8ojIiK6lBWiiIiIqLwEooiIiKi8BKKIiIiovASiiIiIqLwEooiIiKi8BKKIiIiovMoFIkmjJO1d9/5jko5dzH2OlbTT4uwjIiIieq9ygQgYBbwZiGxfafu7i7nPsUACUURExFKqXwUiSStJulrSbEnzJI2XNFrSZEnTJV0rad1S92ZJJ0m6U9IDknaVtBzwLWC8pFnl+UMlnV6emSTpDEk3SXpI0nslnSPpXkmT6saxl6TbJc2QdKmkoaX8EUknlPK5kraQNBw4EvhS6XPXTuY2SdKZkqaW8X6klA+RdG5pb6ak3Uv5iDK3WZLmSNq0gzaPkNQqqfXZp5/rw3+JiIiIgaVfBSLgg8Cjtkfa3gq4BjgNGGd7NHAOcGJd/cG2dwCOBo63/RrwTeAS26NsX9JBH6sBewBfAq4CTgFGAFuX7bY1geOAPW1vB7QCX657/qlSfgZwjO1HgDOBU0qfU7uY33DgvcCHgTMlDQE+D2B7a+AA4LxSfiTwI9ujgBbgr+0bsz3RdovtltXWWLWLbiMiIqqtv/3pjrnAyZJOAn4DPAtsBVwvCWAQ8Fhd/cvK63RqYaMRV9m2pLnA47bnAki6u7SxPrAlcGvpczng9k76/HgP5gbwv7YXAg9KegjYAtiFWujD9n2S/gRsVvr8uqT1gctsP9jDviIiIqLoV4HI9gOSRlM7A/Qd4HrgbttjOnnk1fK6gMbn2vbMwrrrtveDS1vX2z6gD/ts4w7eq8OK9s8lTaO2mnStpH+xfWMP+4uIiAj62ZaZpHcAL9u+EDgZ2BFYS9KYcn9ZSSO6aeZFYNgiDOMOYGdJm5Q+V5S0WR/1ub+kZSRtDLwLuB+YAhxU+toM2AC4X9K7gIdsnwpcCWzTq9lERERE/wpEwNbAnZJmAV+ndh5oHHCSpNnALLr/NtdNwJZth6p7OgDbTwKHAhdLmkMtIG3RzWNXAft1dai6uB+YDPwOONL2K8BPgEFlC+8S4FDbrwLjgXnls9gCOL+nc4mIiIga2e13aaIZyrfYfmP7l4uj/S1HvtsX/m7S4mi6V7Z7x47NHkJERAxwkqbbbmmkbn9bIYqIiIjoc/3qUPVAIOnrwP7tii+1fWgThhMREREkEC1xtk/krb8rKSIiIposgagiVlx2pZzbiYiI6ETOEEVERETlJRBFRERE5SUQRUREROXlDFFFvPLG37nnidlNHcOWa49sav8RERGdyQpRREREVF4CUURERFReAlFERERUXgJRREREVF4CUURERFReAlFERERU3hIJRJJGSdq77v3HJB27mPscK2mnxdnH0kLSI5LWbPY4IiIi+qsltUI0CngzENm+0vZ3F3OfY4FKBKKIiIhYNN0GIkkrSbpa0mxJ8ySNlzRa0mRJ0yVdK2ndUvdmSSdJulPSA5J2lbQc8C1gvKRZ5flDJZ1enpkk6QxJN0l6SNJ7JZ0j6V5Jk+rGsZek2yXNkHSppKGl/BFJJ5TyuZK2kDQcOBL4Uulz107mto6ky8vcZretKEn6cpnrPElHl7Lhku6TdHYpv0jSnpJulfSgpB1KvQmSLpB0Yyk/vJRL0vfLs3MljS/lYyX9pm5Mp0s6tLO5lfI1JF0naaaknwJq+F88IiIi3qaRFaIPAo/aHml7K+Aa4DRgnO3RwDnAiXX1B9veATgaON72a8A3gUtsj7J9SQd9rAbsAXwJuAo4BRgBbF2229YEjgP2tL0d0Ap8ue75p0r5GcAxth8BzgROKX1O7WRupwKTbY8EtgPuljQaOAzYEXgPcLikbUv9TYAfAdsAWwAHArsAxwBfq2t3G+DDwBjgm5LeAXyc2krZSGBP4PttQbIbb5lbKTseuMX2tsCVwAYdPSjpCEmtklqfefrZBrqKiIiopkb+dMdc4GRJJwG/AZ4FtgKulwQwCHisrv5l5XU6MLzBcVxl25LmAo/bngsg6e7SxvrAlsCtpc/lgNs76fPjDfYJtRD2aQDbC4DnJe0CXG77pTKGy4BdqQWPh9uN7Ya6cdfP9de2/w78XdJNwA7UgtPFpZ/HJU0Gtgde6GaMHc1tt7Zr21dL6jDt2J4ITATYatQIN/B5REREVFK3gcj2A2XVZG/gO8D1wN22x3TyyKvldUEj7bd7ZmHdddv7waWt620f0Id9dqar7af2Y6sfd32/7cOHu2j3Dd66Ujekkz7bzy0BJyIioo80coboHcDLti8ETqa2lbSWpDHl/rKSRnTTzIvAsEUY5x3AzpI2KX2uKGmzPujzBuBzpc1BklYGpgD7lj5WAvYDOtty68w+koZIWoPa4e67SrvjSz9rUVvluRP4E7ClpOUlrQK8r4H2pwAHlXF/iNqWY0RERPRSI2eItgbulDQL+Dq180DjgJMkzQZm0f23uW6i9h/9WW2HiXvC9pPAocDFkuZQC0hbdPPYVcB+XR2qBr4I7F62vKYDI2zPACZRCyvTgLNtz+zhkO8Eri7j/LbtR4HLgTnAbOBG4Ku2/8/2X4D/LfcuAhrp6wRgN0kzgL2AP/dwfBEREVFHdnZe+pKkCcB82yc3eyz1tho1wv973c+bOoYt1x7Z1P4jIqJaJE233dJI3fym6oiIiKi8RT2A3C9I+jqwf7viS22f2FH9RWF7Ql+3GREREYtXJQJRCT59Hn4iIiJiYKhEIAoYMniFnOGJiIjoRM4QRUREROUlEEVERETlJRBFRERE5SUQRUREROXlUHVFvPrGK/zxmfuWeL8br97dLxSPiIhovqwQRUREROUlEEVERETlJRBFRERE5SUQRUREROUlEEVERETlJRBFRERE5Q34QCRpgqRjFnMfW0iaJWmmpI0XZ191fd4sqWVJ9BURETHQDfhAtITsC/za9ra2/9jswURERETPDMhAJOnrku6X9Htg81J2uKS7JM2W9CtJK0oaJulhScuWOitLeqTtfQftjpJ0h6Q5ki6XtJqkvYGjgX+RdFMnz31V0lHl+hRJN5br90m6sFzvJel2STMkXSppaCkfLWmypOmSrpW0bru2l5F0nqT/6qDfIyS1Smp95ulne/dhRkREVMCAC0SSRgOfBLYFPg5sX25dZnt72yOBe4HP2n4RuBn4cKnzSeBXtl/vpPnzgf+0vQ0wFzje9m+BM4FTbO/eyXNTgF3LdQswtISuXYCpktYEjgP2tL0d0Ap8udQ5DRhnezRwDnBiXbuDgYuAB2wf175T2xNtt9huWX2N1ToZWkRERAzEP92xK3C57ZcBJF1ZyrcqqyirAkOBa0v52cBXgSuAw4DDO2pU0irAqrYnl6LzgEsbHNN0YLSkYcCrwAxqwWhX4CjgPcCWwK2SAJYDbqe2urUVcH0pHwQ8VtfuT4H/tV0fkiIiIqKHBmIgAnAHZZOAfW3PlnQoMBbA9q2Shkt6LzDI9rw+H4z9uqRHqAWu24A5wO7AxtRWqzYGrrd9QP1zkrYG7rY9ppOmbwN2l/QD26/09bgjIiKqYsBtmVHbntpP0gplReajpXwY8FjZhjqo3TPnAxcD53bWqO3ngWcltW19HQxM7qx+J+M6prxOBY4EZtk2cAews6RNAMr5ps2A+4G1JI0p5ctKGlHX5s+A3wKXShqo4TYiImKxG3CByPYM4BJgFvArauED4BvANOB6oP2ffb8IWI1aKOrKIcD3Jc0BRgHf6sHQpgLrArfbfhx4pW1stp8EDgUuLm3fAWxh+zVgHHCSpNllTju1m+8PqW3BXSBpwP17RkRELAmqLVBUm6RxwD62D272WBaXrUdt5Stu/OUS73fj1bdY4n1GREQASJpuu6Hf2Vf5bRZJpwEfAvZu9lgiIiKiOSofiGx/oX2ZpB8DO7cr/pHtTs8YlefWAG7o4Nb7bD/d+1FGRETE4lT5QNQR25/v5XNPUztbtNRZfvCQbF9FRER0IodwIyIiovISiCIiIqLyEogiIiKi8hKIIiIiovJyqLoiXl/wGn974U9LtM/1Vt5wifYXERHRW1khioiIiMpLIIqIiIjKSyCKiIiIyksgioiIiMpLIIqIiIjKSyCKiIiIyksg6gFJty3h/t4h6Zfd1Bkuad6SGlNERMRAlEDUA7Z3WpTnJQ3qYX+P2h63KH1GRERE9xKIekDS/PK6rqQpkmZJmidp166ekfQtSdOAMZK+Kemu8txESSr1NpH0e0mzJc2QtHH96k+5nlruzZDUbTiTdISkVkmtTz/9TB99ChEREQNPAlHvHAhca3sUMBKY1UXdlYB5tne0fQtwuu3tbW8FrAB8pNS7CPix7ZHATsBj7dp5Ani/7e2A8cCp3Q3S9kTbLbZb1lhj9cZnFxERUTH50x29cxdwjqRlgStsz+qi7gLgV3Xvd5f0VWBFYHXgbkk3A+vZvhzA9isAZfGozbLA6ZJGlTY365OZRERERFaIesP2FGA34G/ABZI+3UX1V2wvAJA0BPgJMM721sBZwBBAXTzf5kvA49RWpFqA5Xo/g4iIiKiXQNQLkjYEnrB9FvAzYLsGHx1SXp+SNBQYB2D7BeCvkvYt7S8vacV2z64CPGZ7IXAw0KMD2hEREdG5BKLeGQvMkjQT+ATwo0Yesv0ctVWhucAV1Lbe2hwMHCVpDnAb8E/tHv8JcIikO6htl73U++FHREREPdlu9hhiCRi57Tb+7eSrlmif66284RLtLyIiop6k6bZbGqmbFaKIiIiovHzLrI+U3zO0fLvig23PbcZ4IiIionEJRH3E9o7NHkNERET0TgJRRSw7aLmc6YmIiOhEzhBFRERE5SUQRUREROUlEEVERETl5QxRRbyx8HWefKn934tdfNZaad0l1ldERMSiygpRREREVF4CUURERFReAlFERERUXgJRREREVF4CUURERFReAlFERERUXgJRD0m6rUn9Hinp0x2UD5c0rxljioiIGCjye4h6yPZOTer3zGb0GxERUQVZIeohSfPL67qSpkiaJWmepF27eOaDkmZImi3phlK2uqQrJM2RdIekbSQtI+kRSavWPfsHSetImiDpmFI2urR1O/D5xTvjiIiIgS+BqPcOBK61PQoYCczqqJKktYCzgE/YHgnsX26dAMy0vQ3wNeB82wuBXwP7lWd3BB6x/Xi7Zs8FjrI9pqsBSjpCUquk1qeferoXU4yIiKiGBKLeuws4TNIEYGvbL3ZS7z3AFNsPA9h+ppTvAlxQym4E1pC0CnAJML7U+WR5/6ZSZ1Xbk0vRBZ0N0PZE2y22W9ZYc42ezi8iIqIyEoh6yfYUYDfgb8AFHR14LgS4k/K3NQvcDmxSVpb2BS5rsL2IiIjopQSiXpK0IfCE7bOAnwHbdVL1duC9kjYqz61eyqcAB5WyscBTtl+wbeBy4IfAvbbfstdl+zngeUm7lKKD+mpOERERVZVvmfXeWOArkl4H5gMdrhDZflLSEcBlkpYBngDeD0wAzpU0B3gZOKTusUuobckd2knfhwHnSHoZuHaRZxIREVFxqi1IxEA3aruRvn7qNUusv7VWWneJ9RUREdERSdNttzRSN1tmERERUXnZMutDkqYBy7crPtj23GaMJyIiIhqTQNSHbO/Y7DFEREREzyUQVcTgZZbNuZ6IiIhO5AxRREREVF4CUURERFReAlFERERUXgJRREREVF4OVVfEgoVv8Ozf+/Yv3q+2Qv5gbEREDAxZIYqIiIjKSyCKiIiIyksgioiIiMpLIIqIiIjKSyCKiIiIyqtMIJI0StLede8/JunYxdznWEk7Lc4+IiIiYtFVJhABo4A3A5HtK21/dzH3ORZIIIqIiFjK9YtAJGklSVdLmi1pnqTxkkZLmixpuqRrJa1b6t4s6SRJd0p6QNKukpYDvgWMlzSrPH+opNPLM5MknSHpJkkPSXqvpHMk3StpUt049pJ0u6QZki6VNLSUPyLphFI+V9IWkoYDRwJfKn3u2sncJkk6VdJtpe9xdfe+IukuSXMknVDKhkuaV1fnGEkT+vgjj4iIqJR+EYiADwKP2h5peyvgGuA0YJzt0cA5wIl19Qfb3gE4Gjje9mvAN4FLbI+yfUkHfawG7AF8CbgKOAUYAWxdttvWBI4D9rS9HdAKfLnu+adK+RnAMbYfAc4ETil9Tu1ifusCuwAfAb4LtfAFbArsQG11a7Sk3br/qP5B0hGSWiW1PvVU3/5SxoiIiIGkv/ym6rnAyZJOAn4DPAtsBVwvCWAQ8Fhd/cvK63RgeIN9XGXbkuYCj9ueCyDp7tLG+sCWwK2lz+WA2zvp8+M9mBvAFbYXAvdIWqeU7VV+Zpb3Q6kFpD832qjticBEgG23G+UejikiIqIy+kUgsv2ApNHUzgB9B7geuNv2mE4eebW8LqDxObY9s7Duuu394NLW9bYP6MM+2z8LoLrX79j+aX1FSevz1pW9IT3sKyIiItrpF1tmkt4BvGz7QuBkYEdgLUljyv1lJY3oppkXgWGLMIw7gJ0lbVL6XFHSZouxz2uBz9SdU1pP0trA48DaktaQtDy1bbaIiIhYBP1ihQjYGvi+pIXA68DngDeAUyWtQm0e/wPc3UUbNwHHSppFbZWpR2w/KelQ4OISRKB2puiBLh67CvilpH2AL3Rzjqh9f9dJejdwe9mimw98yvYTkr4FTAMeBu7r6VwiIiLirWTnaEkVbLvdKN946w192mb+2n1ERCzNJE233dJI3X6xZRYRERGxOPWXLbN+T9LXgf3bFV9q+8SO6kdERMSSk0C0hJTgk/ATERGxFEogqohBywzOmZ+IiIhO5AxRREREVF4CUURERFReAlFERERUXgJRREREVF4OVVfEQi9g/msv9GmbQ5dbuU/bi4iIaJasEEVERETlJRBFRERE5SUQRUREROUlEEVERETlJRBFRERE5SUQRUREROUNiEAkaYKkYxZzH1tImiVppqSNF2df7fpdVdK/1b0fK+k3S6r/iIiIKhgQgWgJ2Rf4te1tbf9xCfa7KvBv3VWKiIiI3uu3gUjS1yXdL+n3wOal7HBJd0maLelXklaUNEzSw5KWLXVWlvRI2/sO2h0l6Q5JcyRdLmk1SXsDRwP/IummTp4bLuk+SWdLmifpIkl7SrpV0oOSdij1Vpd0RWn/DknblPIJks6RdLOkhyQdVZr+LrBxWZ36fikbKumXpb+LJKmTMR0hqVVS61NPPd2rzzkiIqIK+mUgkjQa+CSwLfBxYPty6zLb29seCdwLfNb2i8DNwIdLnU8Cv7L9eifNnw/8p+1tgLnA8bZ/C5wJnGJ79y6GtgnwI2AbYAvgQGAX4Bjga6XOCcDM0v7XSn9ttgA+AOwAHF9C27HAH22Psv2VUm9bagFtS+BdwM4dDcb2RNsttlvWXHONLoYdERFRbf0yEAG7Apfbftn2C8CVpXwrSVMlzQUOAkaU8rOBw8r1YcC5HTUqaRVgVduTS9F5wG49GNfDtufaXgjcDdxg29SC1fBSZxfgAgDbNwJrlH4Brrb9qu2ngCeAdTrp507bfy39zKprOyIiInqhvwYiAHdQNgn4d9tbU1uJGQJg+1ZguKT3AoNsz1tMY3q17nph3fuF/OPvxnW0vdU2l/rnF9D535prtF5EREQ0oL8GoinAfpJWkDQM+GgpHwY8VraaDmr3zPnAxXSyOgRg+3ngWUm7lqKDgcmd1e+lKW1jkzQWeKqscnXmRWrzioiIiMWkX64s2J4h6RJq20V/AqaWW98AppWyubw1SFwE/Be1UNSVQ4AzJa0IPMQ/ttr6ygTgXElzgJdLf52y/XQ5mD0P+B1wdR+PJyIiovJUO+Iy8EkaB+xj++Bmj6UZthu9rafc3reLXUOXW7lP24uIiOhLkqbbbmmkbr9cIeopSacBHwL2bvZYIiIiYulTiUBk+wvtyyT9mLd/Xf1Htjs9Y1SeWwO4oYNb77OdX/YTERHRD1UiEHXE9ud7+dzTwKi+Hc3it4wGZYsrIiKiE/31W2YRERERfSaBKCIiIiqvMt8yqzpJLwL3N3scTbIm8FSzB9EkVZ17VecNmXsV517VeUP3c9/Q9lqNNFTZM0QVdH+jXz0caCS1Zu7VUtV5Q+ZexblXdd7Qt3PPlllERERUXgJRREREVF4CUXVMbPYAmihzr56qzhsy9yqq6ryhD+eeQ9URERFReVkhioiIiMpLIIqIiIjKSyCqAEkflHS/pD9IOrbZ4+lLkt4p6SZJ90q6W9IXS/nqkq6X9GB5Xa3umf9XPov7JX2geaNfdJIGSZop6TflfVXmvaqkX0q6r/zbj6nQ3L9U/m99nqSLJQ0ZqHOXdI6kJyTNqyvr8VwljZY0t9w7VZKW9Fx6qpO5f7/83/wcSZdLWrXu3oCYe0fzrrt3jCRLWrOurO/mbTs/A/gHGAT8EXgXsBwwG9iy2ePqw/mtC2xXrocBDwBbAt8Dji3lxwInlesty2ewPLBR+WwGNXseizD/LwM/B35T3ldl3ucB/1KulwNWrcLcgfWAh4EVyvv/BQ4dqHMHdgO2A+bVlfV4rsCdwBhAwO+ADzV7br2c+17A4HJ90kCce0fzLuXvBK4F/gSsuTjmnRWigW8H4A+2H7L9GvALYJ8mj6nP2H7M9oxy/SJwL7X/aOxD7T+alNd9y/U+wC9sv2r7YeAP1D6jfkfS+sCHgbPriqsw75Wp/Y/mzwBsv2b7OSow92IwsIKkwcCKwKMM0LnbngI80664R3OVtC6wsu3bXfsv5fl1zyy1Opq77etsv1He3gGsX64HzNw7+TcHOAX4KlD/TbA+nXcC0cC3HvCXuvd/LWUDjqThwLbANGAd249BLTQBa5dqA+nz+B9q/wOxsK6sCvN+F/AkcG7ZLjxb0kpUYO62/wacDPwZeAx43vZ1VGDudXo61/XKdfvy/u4z1FY+YIDPXdLHgL/Znt3uVp/OO4Fo4Oto33TA/a4FSUOBXwFH236hq6odlPW7z0PSR4AnbE9v9JEOyvrdvIvB1JbUz7C9LfASta2TzgyYuZfzMvtQ2x54B7CSpE919UgHZf1y7g3obK4D7jOQ9HXgDeCitqIOqg2IuUtaEfg68M2ObndQ1ut5JxANfH+ltvfaZn1qS+wDhqRlqYWhi2xfVoofL8umlNcnSvlA+Tx2Bj4m6RFq26B7SLqQgT9vqM3lr7anlfe/pBaQqjD3PYGHbT9p+3XgMmAnqjH3Nj2d61/5x9ZSfXm/JOkQ4CPAQWU7CAb23Dem9v8BmF3+9259YIakf6KP551ANPDdBWwqaSNJywGfBK5s8pj6TPnmwM+Ae23/sO7WlcAh5foQ4Nd15Z+UtLykjYBNqR2+61ds/z/b69seTu3f9Ebbn2KAzxvA9v8Bf5G0eSl6H3APFZg7ta2y90hasfzf/vuonZurwtzb9GiuZVvtRUnvKZ/Zp+ue6VckfRD4T+Bjtl+uuzVg5257ru21bQ8v/3v3V2pfpPk/+nrezT5Rnp/F/wPsTe3bV38Evt7s8fTx3HahthQ6B5hVfvYG1gBuAB4sr6vXPfP18lncz1L+jYsGP4Ox/ONbZpWYNzAKaC3/7lcAq1Vo7icA9wHzgAuofcNmQM4duJjaWanXy38IP9ubuQIt5fP6I3A65a80LM0/ncz9D9TOzLT9b92ZA23uHc273f1HKN8y6+t55093REREROVlyywiIiIqL4EoIiIiKi+BKCIiIiovgSgiIiIqL4EoIiIiKi+BKCIiIiovgSgiYilT/nBrRCxBCUQRET0g6dOS5kiaLekCSRtKuqGU3SBpA0mrSHpE0jLlmRUl/UXSspI2lnSNpOmSpkraotSZJOmHkm4CTmrqJCMqKP+/kIiIBkkaQe034+5s+ylJqwPnAefbPk/SZ4BTbe8raTbwXuAm4KPAtbZflzQRONL2g5J2BH4C7FG62AzY0/aCJT23iKpLIIqIaNwewC9tPwVg+xlJY4CPl/sXAN8r15cA46kFok8CP5E0lNofY7209ieWgNqf3mhzacJQRHMkEEVENE7U/nZeV9ruXwl8p6wijQZuBFYCnrM9qpNnX+qLQUZEz+UMUURE424A/lnSGgAl7NxGbQUI4CDgFgDb86n9ZfkfUfvjuwtsvwA8LGn/8rwkjVzCc4iIDmSFKCKiQbbvlnQiMFnSAmAmcBRwjqSvAE8Ch9U9cglwKTC2ruwg4AxJxwHLAr8AZi+B4UdEF/LX7iMiIqLysmUWERERlZdAFBEREZWXQBQRERGVl0AUERERlZdAFBEREZWXQBQRERGVl0AUERERlff/ASJueoxlCINHAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 576x864 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(8,12))\n",
    "sns.barplot(data=cover_df, y='feature', x='cover', orient='h', palette='Greens_r')\n",
    "plt.ylabel('');\n",
    "plt.title('Cover', fontsize=18);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjkAAANYCAYAAAAmNPwmAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8vihELAAAACXBIWXMAAAsTAAALEwEAmpwYAAB+pklEQVR4nOzdeZhdVZ318e8CwpgQZiGBJAKCDA2oQRxAULAZhHZiUqagEmlfxQFHRAUFHNoBEboxKIPMozSDKNI2CAhoQgMtiA2EhJCEQCaSQIAQ1vvH2aWHspJUpapyq06tz/Pcp+49e5+9f+eGx1rus+8t2SYiIiKiaVZqdQERERERvSEhJyIiIhopISciIiIaKSEnIiIiGikhJyIiIhopISciIiIaKSEnIqIHSLKk81tdB4CkUaWek1pdS0QrJeRERL9RfnF39jGqE+ONknSSpJ16v/p/mHtlSYdL+pWkpyS9JGm+pPsl/UTSziu6poimWaXVBUREdMER7V7vBowFxgG3t2t7phPjjQK+AUwC7uteaZ0naQPgWuDtwATgP4ApwOrAtsB7gU9Kepvtu5ZjisnAGsDLPVJwRD+VkBMR/Ybti+qvJa1CFXLuat/WV0kScCVVwPmU7TM76PMZ4GjgxeWZw9VX2b/QjTIjGiG3qyKicSStJenbkh6T9GK5HfQLSSNrfcYA/11enle7zXVraV9J0lcl/b52O+kJSf8haf1ulLc/sAdwSUcBB8D2y7bPsX1vrd4hkk6RdI+kmeW6HpX0HUlrtrv+f9iTUz8maX9Jf5L0gqTpkv6tBMaIRsl/1BHRKOWX9W+oVkquAn4AvA74V+CfJY22/STwe+A04ARefbtrRvm5KvAF4GrgP4HngJ2BjwK7SnqT7ZeWo8QDy8+fdfG84cDHSj2XUN2K2h34IvAGYO9OjrMf8AngbOBcqltjnwfmUL0fEY2RkBMRTXM0VcD5N9tfbDso6RbgBuDbwBG2J0r6LVXI6eh214vAJrYX1o6dLekPVAHlfcAVy1Hf9uXnfe0byl6duoW2nyvPJwKb2V5Uaz9L0reAEyW92fYfOzH/dsB2tieVOc8G/hf4FAk50TC5XRURTfN+4BWqMPM3tm+kChbvlbTM/+1zZSH87ZNQ65QQ8rvSZZflrG/t8nNe/aCkwVSbpeuP79bqeakt4EhaRdK6pZ5buljPtW0Bp4xrqtt2G5caIhojIScimua1wDTbczpoexAYArRfMemQpIMl3QMspLqd8wzVigrAustZX1u4Wbvd8YXAu8vj8CXU8wlJD1CtMs0u9dzaxXomdnBsVvnZnb1GEX1ObldFRNOoRwaRPgBcDvwR+DTVR7xfAFYGfs3y/5/EPwNvBHbi7xufsb2YsirT0Xf8SPoc1f6im4EzgGnAS1R7dc7vQj2Ll9LWI+9dRF+RkBMRTfMYsI+kdWzPbde2LdVKyszy2ksZ5wiqUPNO28+3HZT0+m7WdxVwJNUm4v9eRt/29UwC9rX9Sq2efbpZT0Rj5XZVRDTNtVT/2/bl+kFJ+1J9Cum6WkhYUH6u18E4i6lC0N/+d7J8x82J3azvBuA24MOSPrmEPh2tqLTV87e28kmyL3fQNyLISk5ENM/5wFHAl8ptn98DW1J9bHoG1aep2jwEzAc+Iel5YC7wtO3fUa24fBD4naRfAIOoPlH1qu+k6SrblnQgVRj7Sfm+nuupboetCWwBHFy6T6qdehXVZuqbJF1Dtafnw0D901YRUZOQExGNYnuRpL2pVlwOAT5AFV6uBE60PaXWd6GkQ4FTgNOB1ahWWX5n+zJJQ4DPAt+n2nh8PdXKySy6wfZMSXsAH6IKKp+gWk16AXicKgCdZ3t87bR/o1rF+SjwY+Apqj1D51GFtYhoR9WnByMiIiKaJXtyIiIiopESciIiIqKREnIiIiKikRJyIiIiopHy6arolg022MCjRo1qdRkREdFPTZgwYabtDXtj7ISc6JZRo0Yxfvz4ZXeMiIjogKTJvTV2bldFREREI+V7cqJbBg3eyOvteGCry4iIiBVkxp3/3qPjSZpge3SPDlpkJSciIiIaKSEnIiIiGikhJyIiIhopISciIiIaKSEnIiIiGikhJyIiIhopIaeHSRoj6Y5ujrGHpCd7qqaIiIiBKCEnIiIiGikhJyIiIhqpMSFH0iRJn5f0gKRnJV0uafXStr+k+yTNlfQHSTuU40dLur42xqOSrqi9niJpJ1V+JOnpMvYDkrYvfdaXdJ2keZL+CGzRrq7XS/qtpNmS/irp4FrbfpIekjRf0lRJn1/CtR1X+m26lOvfQ9KTkk6QNLO8H4fV2odK+oWkZyRNlnSipJVK2xhJd0r6Sbm+hyXt2cV/goiIiD6laX+g82BgH+AF4E5gTAke5wIHAOOBw4HrJG0N3Ab8qPyyfw0wCHg7gKTNgcHAA8A/A+8AtgKeBV4PzC1znlXm2wR4LfAb4PEyxlrAb4GvA/sCOwA3S3rQ9oPAz4GDbd8uad1y/qtI+hrwfmB3288s4/o3BjYAhgNvAX4labztvwI/AYYCmwPrAzcD00sNALsAV5XzPwBcI+m1tmd3UNNYYCzASqsOXkZJERERrdGYlZziDNvTyi/m64GdgGOAn9q+x/Zi2xcALwJvsT0RmF/67U4VUKZKen15fbvtV4BFwBCqcCPbf7E9XdLKwAeBr9t+zvafgQtq9ewPTLJ9nu2Xbd8LXA20/bGnRcC2kta2Pae0t5GkHwJ7A+/sRMBp8zXbL9q+DbgROLjUeQjwFdvzbU8CfgAcUTvvaeB024tsXw78FXhPRxPYHmd7tO3RKw1ao5NlRURErFhNCzlP1Z4/T7USMxI4vtyqmitpLrAZMKz0uw3Yg2ql5jbgVqqAs3t5je3fAWdSrdrMkDRO0trAhlSrYVNq89b/ZPxIYJd2cx9GteICVUDaD5gs6TZJb62duw7Vasm3bT/byeufY/u5drUMo1qdWbVdbZOpVnzaTPWr/1pr27kRERH9UtNCTkemAKfaXqf2WNP2paW9LeTsVp7fRruQA2D7DNtvArajum31BeAZ4GWq0NRmRLu5b2s392Db/1rG/JPt9wIbAdcCV9TOnUO1EnSepLd38lrXLbfI6rVMA2ZSrRqNbNc2tfZ6uCR1cG5ERES/NBBCzjnAsZJ2KRuI15L0HklDSvttwDuBNWw/CdxOta9nfeB/ACTtXM4fBDxHtQdnse3FwDXASZLWlLQtcFRt7huArSQdIWlQeewsaRtJq0o6TNJQ24uAecDieuG2b6Va+fmlpF06eb0nl7F3owpJV5Y6rwBOlTRE0kjgc8BFtfM2Ao4rNR4EbAP8qpNzRkRE9DmNDzm2x1PtyzmTanXkUWBMrf3/gAVU4Qbb84CJwJ0lHACsTRWW5lDdxpkFfL+0fZLqtthTwPnAebWx51NtWj6UalXkKeC7wGqlyxHAJEnzgGOpNkW3r/+3wNFUm6XftIzLfarUOA24GDjW9sOl7VNUAW0icAdwCdWG7Db3AK+jWvU5FTjQ9qxlzBcREdFn6dXbMKK/krQHcJHtJX7MfCnnjgE+ZnvXrp47aPBGXm/HA5fdMSIiGmHGnf/eo+NJmmB7dI8OWjR+JSciIiIGpoScfqR80d+CDh43tbq2iIiIvqZpXwbYaLZPA05bSpcu36oq455PtZ8oIiKiMbKSExEREY2UkBMRERGNlNtV0S07vn4E43t4p31ERERPyEpORERENFJCTkRERDRSQk5EREQ0UkJORERENFL+rEN0y6C1N/YGbz6y1WXEMky/5XutLiEiokP5sw4RERERXZSQExEREY2UkBMRERGNlJATERERjZSQExEREY2UkBMRERGNlJATERERjZSQ00KSbpX0sVbXERER0UQJOfEPJI2RdEer64iIiOiOhJwVQJW81xEREStQfvF2QNLRkq6vvX5U0hW111Mk7STpbZL+JOnZ8vNttT63SjpV0p3A88Dmkt4t6eHS/0xAnaznGEl/kTRf0kOS3liOb1PmmSvpQUn/0m7+j9Vev2p1RpIlHSvpEUlzJJ1Vwtg2wNnAWyUtkDS3g3rGShovafwrixZ27k2NiIhYwRJyOnYbsJuklSRtAgwC3g4gaXNgMPAEcCNwBrA+8EPgRknr18Y5AhgLDAGeBa4GTgQ2AB5rG3NpJB0EnAQcCawN/AswS9Ig4HrgZmAj4FPAxZK27sJ17g/sDOwIHAzsbfsvwLHAXbYH216n/Um2x9kebXv0SoPW6MJ0ERERK05CTgdsTwTmAzsBuwO/AaZKen15fTvwHuAR2xfaftn2pcDDwAG1oc63/aDtl4F9gYdsX2V7EXA68FQnyvkY8D3bf3LlUduTgbdQha3v2H7J9u+AG4APdeFSv2N7ru0ngP8u1xsREdEIq7S6gD7sNmAPYMvyfC5VwHlreT0MmNzunMnA8NrrKbXnw+qvbVtSvX1JNqNa9WlvGDDF9itLmX9Z6iHrearQFBER0QhZyVmytpCzW3l+G1XI2b08nwaMbHfOCGBq7bVrz6dTBRag2oxcf70UU4AtOjg+Ddis3Ybm+vzPAWvW2jbuxFxtvOwuERERfVtCzpLdBrwTWMP2k1S3qPah2n/zP8CvgK0kfVjSKpIOAbalumXUkRuB7SR9QNIqwHF0Lnj8DPi8pDeVjcFbShoJ3EMVZL4oaZCkPahulV1WzrsP+ICkNSVtCXy0C9c+A9hU0qpdOCciIqJPSchZAtv/ByygCjfYngdMBO60vdj2LKqNu8cDs4AvAvvbnrmE8WYCBwHfKf1fB9zZiTquBE4FLqHaJ3QtsJ7tl6g2Ie8LzAT+HTjS9sPl1B8BL1EFlguAi7tw+b8DHgSektTh9URERPR1snNnIpbfoLU39gZvPrLVZcQyTL/le60uISKiQ5Im2B7dG2NnJSciIiIaKSGnD5B0dvnivfaPs1tdW0RERH+Vj5D3AbaPpfoCvoiIiOghWcmJiIiIRspKTnTLjlttyvhsao2IiD4oKzkRERHRSAk5ERER0UgJOREREdFICTkRERHRSNl4HN3ywKPTGXbAt1pdRiNNu/5rrS4hIqJfy0pORERENFJCTkRERDRSQk5EREQ0UkJORERENFJCTkRERDRSQk5EREQ0UkLOCiTpQUl7tLqOiIiIgSDfk9NLJJ0PPGn7xLZjtrdrXUUREREDS1ZyIiIiopEScmokfUnSVEnzJf1V0p6SVpL0ZUmPSZol6QpJ69XO2VXSHyTNlTRF0hhJY4HDgC9KWiDp+tJ3kqS9yvPVJJ0uaVp5nC5ptdK2h6QnJR0v6WlJ0yUd3Yn6z5d0tqTflmu4TdLIWvvbJP1J0rPl59tqbbdK+rakP5b2/6xfZ0RERH+TkFNI2hr4JLCz7SHA3sAk4DjgfcDuwDBgDnBWOWcEcBPwE2BDYCfgPtvjgIuB79kebPuADqb8KvCWcs6OwJuBE2vtGwNDgeHAR4GzJK3biUs5DPgWsAFwX6mDElhuBM4A1gd+CNwoaf3auUcCHynX+XLpGxER0S8l5PzdYmA1YFtJg2xPsv0Y8HHgq7aftP0icBJwoKRVqALFLbYvtb3I9izb93VyvsOAb9p+2vYzwMnAEbX2RaV9ke1fAQuArTsx7o22f19q/SrwVkmbAe8BHrF9oe2XbV8KPAzUA9iFtv9s+znga8DBklZuP4GksZLGSxr/ykvPdfJyIyIiVqyEnML2o8BnqELM05IukzQMGAn8styOmgv8hSoQvQbYDHhsOaccBkyuvZ5cjrWZZfvl2uvngcGdGHdK2xPbC4DZZdz287XNObyjc0vbIKoVoVexPc72aNujV1p1rU6UFBERseIl5NTYvsT2rlTBxsB3qX7x72t7ndpjddtTS9sWSxpuGdNNK/O0GVGOdddmbU8kDQbWK+O2n69tzqkdnVvaFgEze6CmiIiIFS4hp5C0taR3lc2/LwALqVZszgZObdvAK2lDSe8tp10M7CXpYEmrSFpf0k6lbQaw+VKmvBQ4sYy3AfB14KIeuJT9ymboVan25txjewrwK2ArSR8utR4CbAvcUDv3cEnbSloT+CZwle3FPVBTRETECpeQ83erAd+hWrl4CtgIOAH4MXAdcLOk+cDdwC4Atp8A9gOOp7otdB/VJmKAn1Pt75kr6doO5jsFGA88APwvcG851l2XAN8o9byJau8PtmcB+5daZwFfBPa3XV+puRA4n+r6V6fadB0REdEvyV7WXZXoLzr6AsIunHsrcJHtn3XlvFXXGe4Ndju2q9NFJ0y7/mutLiEiotdJmmB7dG+MnZWciIiIaKT8WYd+RtKD/OMGYqg+6h4RERFFQk4/s4y/f3VxN8bdY3nPjYiI6ItyuyoiIiIaKSEnIiIiGim3q6JbdthyE8bnU0AREdEHZSUnIiIiGikhJyIiIhopISciIiIaKSEnIiIiGikbj6NbHpg4gxEH/aDVZfSaJ648vtUlRETEcspKTkRERDRSQk5EREQ0UkJORERENFJCTkRERDRSQk5EREQ0UkJORERENFJCTotIOkHSzzrZ93xJp/R2TREREU2S78lZDpLOB560feLyjmH7tJ6rKCIiItrLSk4LSEq4jIiI6GUDKuRImiTpC5IekPScpJ9Leo2kmyTNl3SLpHVL3yslPSXpWUm/l7RdOT4WOAz4oqQFkq4vx4dJulrSM5Iel3Rcbd6TJF0l6SJJ84Ax5dhFtT4dzteFa9tD0pPlNtjMcq2H1dqHSvpFqW+ypBMlrVTaxki6U9JPyvwPS9qzG291REREyw2okFN8EHg3sBVwAHATcAKwAdX70RZObgJeB2wE3AtcDGB7XHn+PduDbR9QwsL1wP3AcGBP4DOS9q7N+17gKmCdtrHa6XC+Ltq4XMdw4ChgnKStS9tPgKHA5sDuwJHA0bVzdwEmlvO/AVwjab3lqCEiIqJPGIgh5ye2Z9ieCtwO3GP7f2y/CPwSeAOA7XNtzy/HTwJ2lDR0CWPuDGxo+5u2X7I9ETgHOLTW5y7b19p+xfbC9gN0cb6l+ZrtF23fBtwIHCxpZeAQ4CtljknAD4Ajauc9DZxue5Hty4G/Au/paAJJYyWNlzT+lRefW44SIyIiet9A3Bsyo/Z8YQevB5dQcCpwELAh8Epp3wB4toMxRwLDJM2tHVuZKkS1mbKkgpZjviWZY7ueOiYDw8o4q5bX9bbhtddTbbuDc/9BWc0aB7Dqepu5oz4RERGtNhBXcjrjw1S3l/aiusUzqhxX+dn+F/sU4HHb69QeQ2zvV+uztDCwrPk6a11Ja9VejwCmATOBRVRhrN42tfZ6uCR1cG5ERES/lJDTsSHAi8AsYE2g/ce9Z1DtbWnzR2CepC9JWkPSypK2l7RzD83XFSdLWlXSbsD+wJW2FwNXAKdKGiJpJPA54KLaeRsBx0kaJOkgYBvgV92oIyIioqUScjr2C6rbNVOBh4C727X/HNhW0lxJ15YQcQCwE/A41crJz6hWZXpivs56CphDtQJzMXCs7YdL26eA56g2F98BXAKcWzv3HqqNzzOpbp0daHvWctYRERHRcnr1NozoryTtAVxke9PlOHcM8DHbu3b13FXX28wb7/mZrp7Wbzxx5fGtLiEiotEkTbA9ujfGzkpORERENFJCTj9SvuhvQQePm1pdW0RERF8zED9C3m+Vv3e1tE3JXb5VVcY9Hzh/ec6NiIjoq7KSExEREY2UkBMRERGNlNtV0S07bP4axucTSBER0QdlJSciIiIaKSEnIiIiGikhJyIiIhopISciIiIaKRuPo1v+/PjTvO7IM1tdBgCP/OKTrS4hIiL6kKzkRERERCMl5EREREQjJeREREREIyXkRERERCMl5EREREQjJeREREREIyXkdJOkPSQ92eo6lpckS9qy1XVERET0tISciIiIaKSEnAFCUr74MSIiBpSEnHYkTZL0eUkPSHpW0uWSVu/C+cdJekjSppJWk/R9SU9ImiHpbElrlH5/lnRA7bxBkmZK2knSBZKOL8eHl1tKnyivt5Q0W5LK62MkPVqOXSdpWG1MS/p/kh4BHinHviBpuqRpkj7Srvb9Su3zJU2V9PluvJUREREtlZDTsYOBfYDXAjsAYzpzkqSvlb67234S+C6wFbATsCUwHPh66f4L4PDa6fsB023fB9wG7FGO7w5MLD8B3gHcbtuS3gV8u9S7CTAZuKxdWe8DdgG2lbQP8Hng3cDrgL3a9f058HHbQ4Dtgd915rojIiL6ooScjp1he5rt2cD1VCFlaSTph8DewDttP1NWWo4BPmt7tu35wGnAoeWci4D9JK1dXh8BXFie3wbsJmklqlDzPeDtpW330g5wGHCu7Xttvwh8BXirpFG12r5d5l9IFYbOs/1n288BJ7W7jkVUYWht23Ns37uEix0rabyk8YtfXLCMtyYiIqI1EnI69lTt+fPA4GX0XwcYSxUoni3HNgTWBCZImitpLvDrchzb04A7gQ9KWgfYF7i4tD0GLKAKV7sBNwDTJG3Nq0POMKrVG8p5C4BZVCtGbabUng9r93oyr/ZBqhWlyZJuk/TWji7W9jjbo22PXnm1Zb01ERERrZGQ0zPmAPsD50lqW3GZCSwEtrO9TnkMtV1PBRdQ3bI6CLjL9tRa223AgcCq5fhtwJHAusB9pc80YGTbCZLWAtYH6uO49nw6sFnt9Yj6Rdj+k+33AhsB1wJXdOrqIyIi+qCEnB5i+1aq20e/lLSL7VeAc4AfSdoI/raJeO/aadcCbwQ+TbVHp+424JPA78vrW4FPAXfYXlyOXQIcXTYrr0Z1O+we25OWUOYVwBhJ20paE/hGW4OkVSUdJmmo7UXAPGDxEsaJiIjo8xJyepDt3wJHA9dJehPwJeBR4G5J84BbgK1r/RcCV1NtcL6m3XC3AUP4e8i5g+r21+9r5/8X8LUyxnRgC/6+56ej+m4CTqfaUPwo/7ix+AhgUqn1WF69MToiIqJfke1l94peI+nrwFa2+2WgWH39Ed7sPV9sdRkAPPKLT7a6hIiI6CJJE2yP7o2x8wVxLSRpPeCjVCsoERER0YNyu6oTJJ0gaUEHj5u6MeYxVJ90usn275fVPyIiIromKzmdYPs0qk29PTnmOVQbkyMiIqIXZCUnIiIiGikhJyIiIhopt6uiW7Z/7UaMz6eaIiKiD8pKTkRERDRSQk5EREQ0UkJORERENFJCTkRERDRSNh5Htzw4+Rm2O2Zcq8vgwXPGtrqEiIjoY7KSExEREY2UkBMRERGNlJATERERjZSQExEREY2UkBMRERGNlJATERERjZSQ049JmiRpr1bXERER0Rcl5EREREQjJeTEUknKF0ZGRES/lJDT/+0k6QFJz0q6XNLqAJKOkfSopNmSrpM0rBwfJcn18CLpVkkfK8/HSLpT0o8kzQZOasVFRUREdFdCTv93MLAP8FpgB2CMpHcB3y5tmwCTgcu6MOYuwERgI+DUHq02IiJiBcmtiP7vDNvTACRdD+wE7Ayca/vecvwrwBxJozo55jTbPynPX27fKGksMBZg0OD1ulV8REREb8lKTv/3VO3588BgYBjV6g0AthcAs4DhnRxzytIabY+zPdr26JVXH9zFciMiIlaMhJxmmgaMbHshaS1gfWAq8Fw5vGat/8btznevVhcREbECJOQ00yXA0ZJ2krQacBpwj+1Jtp+hCjuHS1pZ0keALVpZbERERG9IyGkg2/8FfA24GphOFWIOrXU5BvgC1S2s7YA/rOgaIyIiepvs3JmI5bfGhiO9+fu+2uoyePCcsa0uISIiloOkCbZH98bYWcmJiIiIRkrIiYiIiEZKyImIiIhGSsiJiIiIRkrIiYiIiEbKn3WIbtlu5IaMzyebIiKiD8pKTkRERDRSQk5EREQ0UkJORERENFJCTkRERDRSQk5EREQ0Uj5dFd3y0JRZvPG4C3pl7HvPOKpXxo2IiIEhKzkRERHRSAk5ERER0UgJOREREdFICTkRERHRSAk5ERER0UgJOREREdFICTm9RNKDkvZodR0REREDVb4np5fY3m55zpM0CngcGGT75R4tKiIiYgDJSk4/JCnhNCIiYhkScnqJpEmS9pL0ZknjJc2TNEPSD5dx6u/Lz7mSFkh6q6Qxku6U9CNJs4GTJJ0k6aLafKMkuS0ASbpV0imS/lDGuV7S+pIuLrX8qawatZ1vScdJmihppqR/k5T/PiIiot/KL7He92Pgx7bXBrYArlhG/3eUn+vYHmz7rvJ6F2AisBFwaifnPhQ4Ahhe5r4LOA9YD/gL8I12/d8PjAbeCLwX+EhHg0oaW4Lb+JcXzu9kKREREStWQk7vWwRsKWkD2wts372c40yz/RPbL9te2MlzzrP9mO1ngZuAx2zfUvb6XAm8oV3/79qebfsJ4HTgQx0Nanuc7dG2R6+yxpDlvJyIiIjelZDT+z4KbAU8XG4R7b+c40xZjnNm1J4v7OD14KXMMRkYthxzRkRE9AkJOb3M9iO2P0R1m+m7wFWS1lraKZ08/hywZu31xstf5d9sVns+ApjWA2NGRES0REJOL5N0uKQNbb8CzC2HFy/llGeAV4DNlzH0fcA7JI2QNBT4SndrBb4gaV1JmwGfBi7vgTEjIiJaIiGn9+0DPChpAdUm5ENtv7Ckzrafp9pYfKekuZLesoR+v6UKIQ8AE4AbeqDW/yxj3QfcCPy8B8aMiIhoCdlLujsSA4kkA6+z/WhXzlvzNa/16w85qVdquveMo3pl3IiI6DskTbA9ujfGzkpORERENFJCTgtIOqx8QV/7x4Otri0iIqIp8ucBWsD2xcDFra6jzrZaXUNERERPykpORERENFJWcqJbtt1sfcZng3BERPRBWcmJiIiIRkrIiYiIiEZKyImIiIhGSsiJiIiIRkrIiYiIiEbKp6uiWx5+cjZv/9KlPTrmnd/9UI+OFxERA1NWciIiIqKREnIiIiKikRJyIiIiopESciIiIqKREnIiIiKikRJyIiIiopESciIiIqKRBkTIkTRJ0l69PMcoSZa03N89JGkPSU/2YE09Ol5ERER/MiBCTkRERAw8CTkRERHRSAMp5Owk6QFJz0q6XNLqAJL2l3SfpLmS/iBph7YTJH1Z0mOS5kt6SNL7a20rS/q+pJmSJgLv6UwRko6W9Jcy5kRJH++gzwll3EmSDqsd36/UMV/SVEmf78obIGmYpKslPSPpcUnH1dpOknSFpF+U8R+UNHoJ44yVNF7S+EUL53elhIiIiBVmIIWcg4F9gNcCOwBjJL0ROBf4OLA+8FPgOkmrlXMeA3YDhgInAxdJ2qS0HQPsD7wBGA0c2Mk6ni7nrQ0cDfyo1NFmY2ADYDhwFDBO0tal7efAx20PAbYHftfZi5e0EnA9cH8Ze0/gM5L2rnX7F+AyYB3gOuDMjsayPc72aNujB60xpLMlRERErFADKeScYXua7dlUv+x3ogoqP7V9j+3Fti8AXgTeAmD7ynLOK7YvBx4B3lzGOxg43faUMua3O1OE7RttP+bKbcDNVEGq7mu2XyztN5a5ABYB20pa2/Yc2/d24fp3Bja0/U3bL9meCJwDHFrrc4ftX9leDFwI7NiF8SMiIvqUgRRynqo9fx4YDIwEji+3quZKmgtsBgwDkHRk7VbWXKrVkw3KGMOAKbUxJ3emCEn7Srpb0uwy5n61MQHm2H6u3bjDyvMPlv6TJd0m6a2dmbMYCQxrd60nAK+p9Wn/Hq3enU+LRUREtNJA/wU2BTjV9qntGySNpFrp2BO4y/ZiSfcBKl2mUwWiNiOWNVm5DXY1cCTwn7YXSbq2NibAupLWqgWdEcCfAWz/CXivpEHAJ4Er2tWwrGt93PbrOtk/IiKiXxtIKzkdOQc4VtIuqqwl6T2ShgBrAQaegWrDMNVKTpsrgOMkbSppXeDLnZhvVWC1MubLkvYF/rmDfidLWlXSblT7d64srw+TNNT2ImAesLgL1/pHYJ6kL0lao2yc3l7Szl0YIyIiot8Y0CHH9niqfTlnAnOAR4Expe0h4AfAXcAM4J+AO2unnwP8hmoj773ANZ2Ybz5wHFVAmgN8mGqDb91TpW0acDFwrO2HS9sRwCRJ84BjgcO7cK2LgQOo9iI9DswEfka1qToiIqJxZLvVNUQ/Nnjjzb3jUf9wt69b7vzuh3p0vIiI6LskTbDd4VeWdNeAXsmJiIiI5krI6QWSFizh0f6j4j0x1wlLmOumnp4rIiKiPxnon67qFbYHr8C5TgNOW1HzRURE9BcJOdEtr990veyhiYiIPim3qyIiIqKREnIiIiKikRJyIiIiopESciIiIqKREnIiIiKikfLpquiW/5s2hz2/cXWPjvlfJ3+wR8eLiIiBKSs5ERER0UgJOREREdFICTkRERHRSAk5ERER0UgJOREREdFICTkRERHRSAk5ERER0UgDNuRImiRpr16eY5QkS1qlvL5V0sfK88Mk3dzL81vSlr05R0RERF81YENOq9m+2PY/t7qOiIiIpkrIiYiIiEYa6CFnJ0kPSHpW0uWSVgeQtL+k+yTNlfQHSTu0nSDpy5IekzRf0kOS3l9rW1nS9yXNlDQReM+SJpY0RtIdtdeWdKykRyTNkXSWJNXaPyLpL6XtN5JGduVCJa1WantC0gxJZ0tao7TtIelJScdLelrSdElHL2WssZLGSxr/0vPzulJGRETECjPQQ87BwD7Aa4EdgDGS3gicC3wcWB/4KXCdpNXKOY8BuwFDgZOBiyRtUtqOAfYH3gCMBg7sYj37AzsDO5ba9gaQ9D7gBOADwIbA7cClXRz7u8BWwE7AlsBw4Ou19o2prmk48FHgLEnrdjSQ7XG2R9seveqaa3exjIiIiBVjoIecM2xPsz0buJ4qABwD/NT2PbYX274AeBF4C4DtK8s5r9i+HHgEeHMZ72DgdNtTypjf7mI937E91/YTwH+XeqAKXN+2/RfbLwOnUa1CdWo1p6wIHQN81vZs2/PLGIfWui0Cvml7ke1fAQuArbtYf0RERJ8x0EPOU7XnzwODgZHA8eVW1VxJc4HNgGEAko6s3cqaC2wPbFDGGAZMqY05uQfqodT049qcswFRrbp0xobAmsCE2hi/LsfbzCoBqqP5IyIi+p1VWl1AHzQFONX2qe0bysrJOcCewF22F0u6jypwAEynCkRtRvRwTRcv5/kzgYXAdran9lBNERERfdpAX8npyDnAsZJ2UWUtSe+RNARYCzDwDEDZnLt97dwrgOMkbVr2s3y5h2o6G/iKpO3KvEMlHdTZk22/QnVdP5K0URljuKS9e6i+iIiIPichpx3b46n2r5wJzAEeBcaUtoeAHwB3ATOAfwLurJ1+DvAb4H7gXuCaHqrpl1Qbhy+TNA/4M7BvF4f5EtW13F3GuIXsuYmIiAaT7VbXEP3Y2sO28M7HfK9Hx/yvkz/Yo+NFRETfJWmC7dG9MXZWciIiIqKRsvG4H5O0G3BTR22288moiIgY0BJy+jHbt5OPeUdERHQoISe6Zath62YPTURE9EnZkxMRERGNlJATERERjZSQExEREY2UkBMRERGNlJATERERjZRPV0W3PPrUXPb/znU9Nt4NX/6XHhsrIiIGtqzkRERERCMl5EREREQjJeREREREIyXkRERERCMl5EREREQjJeREREREI7Uk5EiaJGmvXp5jlCRLWqW8vlXSx8rzwyTd3Jvz9wWSdpP011bXERER0QoDciXH9sW2/7nVdfS0Euq2bHtt+3bbW7eypoiIiFYZkCEnIiIimq+VIWcnSQ9IelbS5ZJWB5C0v6T7JM2V9AdJO7SdIOnLkh6TNF/SQ5LeX2tbWdL3Jc2UNBF4z5ImljRG0h2115Z0rKRHJM2RdJYk1do/Iukvpe03kkYu6+IkvVvSw+X6zpR0W+122UmSLqr1bX9rbaikn0uaLmmqpFMkrVzatixjPVuu9fJy/PdluPslLZB0iKQ9JD1Zm2ebctturqQHJf1Lre38ct03lvf3HklbLOs6IyIi+qpWhpyDgX2A1wI7AGMkvRE4F/g4sD7wU+A6SauVcx4DdgOGAicDF0napLQdA+wPvAEYDRzYxXr2B3YGdiy17Q0g6X3ACcAHgA2B24FLlzaQpA2Aq4ETgQ1K3W/vQi0XAC8DW1Jdzz8DHytt3wJuBtYFNgV+AmD7HaV9R9uDbV/erqZBwPXl3I2ATwEXS6rfzvoQ1fu6LvAocOoSrm+spPGSxr/03LwuXFZERMSK08qQc4btabZnU/3y3YkqqPzU9j22F9u+AHgReAuA7SvLOa+UX+KPAG8u4x0MnG57Shnz212s5zu259p+AvjvUg9Ugevbtv9i+2XgNKpVqKWt5uwHPGT7KtuLgNOBpzpThKTXAPsCn7H9nO2ngR8Bh5Yui4CRwDDbL9i+YwlDtfcWYHC5zpds/w64gSrYtLnG9h/LdV7M39+DV7E9zvZo26NXXWvtTk4fERGxYrUy5NR/6T9P9Qt4JHB8uZ0yV9JcYDNgGICkI2u3suYC21OtlFD6TKmNObkH6qHU9OPanLMBAcOXMtararHtdrUtzUhgEDC9NudPqVZfAL5Y5v9jueX0kU6OOwyYYvuV2rHJvPo6lvQeRERE9Dt97a+QTwFOtf0Pt0nKysk5wJ7AXbYXS7qP6hc+wHSqQNRmRA/XdHEXznlVLWV/T72254A1a683bjffi8AGZUXlVWw/RbXihaRdgVsk/d72o8uoaRqwmaSVakFnBPB/nbukiIiI/qWvfbrqHOBYSbuospak90gaAqwFGHgGQNLRVCs5ba4AjpO0qaR1gS/3UE1nA1+RtF2Zd6ikg5Zxzo3AdpI+UDYTH8erg8x9wDskjZA0FPhKW4Pt6VT7Zn4gaW1JK0naQtLuZf6DJG1aus+hek8Wl9czgM2XUNM9VOHqi5IGSdoDOAC4bJnvQERERD/Up0KO7fFUqxRnUv0CfxQYU9oeAn4A3EX1y/yfgDtrp58D/Aa4H7gXuKaHavol8F3gMknzgD9T7ZlZ2jkzgYOA7wCzgNfVa7X9W+By4AFgAtXemLojgVWBh6jeh6uAtg3WOwP3SFoAXAd82vbjpe0k4IJym+vgdjW9BPxLqX0m8O/AkbYf7tQbERER0c+o2i4SvU3SrcBFtn/W6lp60jqbbuldP/nDHhvvhi//y7I7RUREY0iaYHt0b4zdp1ZyIiIiInpKX9t43G9I2g24qaM22/lUUkRERIsl5Cwn27fThY9Y296j96qJiIiI9nK7KiIiIhopKznRLVtuvE42C0dERJ+UlZyIiIhopISciIiIaKSEnIiIiGikhJyIiIhopISciIiIaKR8uiq6ZeLTz3LQjzv8TsQuu/LTS/2TYBEREV2SlZyIiIhopISciIiIaKSEnIiIiGikhJyIiIhopISciIiIaKSEnIiIiGikhJwGk7SHpCdbXUdEREQrJOQsg6TzJZ3S6joiIiKiaxJy+hlJ+QLHiIiITmhcyJE0SdJXJD0kaY6k8yStXtqOkfSopNmSrpM0rByXpB9JelrSs5IekLS9pLHAYcAXJS2QdH3p/yVJUyXNl/RXSXsuo6aTJF0l6fJyzr2Sdqy1D5N0taRnJD0u6bgOzr1I0jxgzFLmWaOsPM2R9BCwc7v2L0t6rNTwkKT3l+Orlffkn2p9N5K0UNKGnX7zIyIi+pDGhZziMGBvYAtgK+BESe8Cvg0cDGwCTAYuK/3/GXhH6bsOcAgwy/Y44GLge7YH2z5A0tbAJ4GdbQ8p80zqRE3vBa4E1gMuAa6VNEjSSsD1wP3AcGBP4DOS9m537lWltouXMsc3yjVvUeo6ql37Y8BuwFDgZOAiSZvYfrG8F4fX+n4IuMX2M+0nkTRW0nhJ419cMK8Tlx4REbHiNTXknGl7iu3ZwKlUv7APA861fW/5pf4V4K2SRgGLgCHA6wHZ/ovt6UsYezGwGrCtpEG2J9l+rBM1TbB9le1FwA+B1YG3UK22bGj7m7Zfsj0ROAc4tHbuXbavtf2K7YVLmeNg4FTbs21PAc6oN9q+0va0Ms7lwCPAm0vzBcCHS+gCOAK4sKNJbI+zPdr26NUGr92JS4+IiFjxmhpyptSeTwaGlcfktoO2FwCzgOG2fwecCZwFzJA0TlKHv71tPwp8BjgJeFrSZW23vTpbk+1XgCdLTSOBYZLmtj2AE4DXLOF6lmYY/3jtfyPpSEn31ebZHtig1HQP8Bywu6TXA1sC13Vy3oiIiD6nqSFns9rzEcC08hjZdlDSWsD6wFQA22fYfhOwHdVtqy+Urm4/uO1LbO9axjPw3a7UVFZLNi01TQEet71O7THE9n71KTsxPsB0/vHa2+YcSbVC9ElgfdvrAH8GVOt/AdUtqyOAq2y/0Ml5IyIi+pymhpz/J2lTSetRrYpcTrUP5mhJO0laDTgNuMf2JEk7S9pF0iCq1YwXqG5LAcwANm8bWNLWkt5VxngBWFjruzRvkvSB8umozwAvAncDfwTmlc3Ma0hauWx63nlpgy3BFcBXJK0raVPgU7W2tajC0jPlOo6mWsmpuxB4P1XQ+cVyzB8REdFnNDXkXALcDEwsj1Ns/xfwNeBqqhWPLfj7vpe1qVY55lDd4pkFfL+0/Zxq/81cSddS7cf5DjATeArYiCpILct/Um1onkO1UvIB24tsLwYOAHYCHi/j/oxqc3BXnVzqf7xc/9/21Nh+CPgBcBdVcPsn4M76ybafBO6lCkO3L8f8ERERfYbszt4J6R8kTQI+ZvuWVtfSRtJJwJa2D19W31aTdC4wzfaJnem/3ojXec/jz1h2x0648tP79sg4ERHRf0iaYHt0b4ydL5aLvymfNPsA8IYWlxIREdFtTb1dtcJJuql8YWD7R2duZbV8HknfotqI/G+2H++ZaiMiIlqncSs5tke1aN4Vcq+lt+ax/TWqPUsRERGNkJWciIiIaKTGreTEirX5RkOzYTgiIvqkrOREREREIyXkRERERCMl5EREREQjJeREREREIyXkRERERCPl01XRLZOemcdRZ3fvL2hccOxePVRNRETE32UlJyIiIhopISciIiIaKSEnIiIiGikhJyIiIhopISciIiIaKSEnIiIiGmlAhRxJkyS17PPKknaT9NdWzR8RETGQDKiQs6JJsqQt217bvt321q2sKSIiYqBIyImIiIhG6rchR9KXJE2VNF/SXyXtKel8SafU+uwh6cl2p+4s6SFJcySdJ2n10ncDSTdImitptqTbJa1U2oZJulrSM5Iel3RcbY6VJZ0g6bFSywRJm0n6felyv6QFkg5pX4+kbSTdWuZ8UNK/1NrOl3SWpBvLuPdI2qIT74slHSdpoqSZkv6tdh0rSTpR0mRJT0v6haShpW1UOXespGmSpks6vuv/MhEREX1Dvww5krYGPgnsbHsIsDcwqZOnH1b6bwFsBZxYjh8PPAlsCLwGOAFwCQjXA/cDw4E9gc9I2ruc9zngQ8B+wNrAR4Dnbb+jtO9oe7Dty9tdw6Ay7s3ARsCngIvLtbX5EHAysC7wKHBqJ6/x/cBo4I3Ae0tNAGPK453A5sBg4Mx2574TeB3wz8CXO9rDVILQeEnjX1jwbCdLioiIWLH6ZcgBFgOrAdtKGmR7ku3HOnnumban2J5NFRo+VI4vAjYBRtpeVPbPGNgZ2ND2N22/ZHsicA5waDnvY8CJtv/qyv22Z3WijrdQhYzvlHF/B9xQqwfgGtt/tP0ycDGwUyev8bu2Z9t+Aji9NuZhwA9tT7S9APgKcKik+t8wO9n2c7b/FzivXT0A2B5ne7Tt0asPHtrJkiIiIlasfhlybD8KfAY4CXha0mWShnXy9Cm155OBtvP+jWq15OZyq+fL5fhIYFi5pTRX0lyqVZ7XlPbNgM4GrLphwBTbr7SrZ3jt9VO1589ThaLOWNI1Diuv622r8PdrWdq5ERER/Uq/DDkAti+xvStVCDHwXeA5YM1at407OHWz2vMRwLQy3nzbx9veHDgA+JykPal+6T9ue53aY4jt/coYU6hufXXVNGCztv0ytXqmLsdY7XV4jeXnyHZtLwMzOnFuREREv9IvQ46krSW9S9JqwAvAQqpbWPcB+0laT9LGVKs97f0/SZtKWo9qRebyMub+kraUJGBeGW8x8EdgXtnovEbZaLy9pJ3LeD8DviXpdarsIGn90jaDau9LR+6hCmVflDRI0h5U4eqy5X5j/u4LktaVtBnw6bZrBC4FPivptZIGA6cBl5fbYW2+JmlNSdsBR9fOjYiI6Ff6Zcih2o/zHWAm1S2djagCy4VUG4QnUW3o7egX9CWlbWJ5tH0a63XALcAC4C7g323fansxVfjYCXi8zPkzoG0zyg+BK8qY84CfA2uUtpOAC8ptroPrRdh+CfgXYN8y5r8DR9p+uOtvxz/4T2ACVei7sdQEcC7Ve/T7ci0vUG14rruN6rbdfwHft31zD9QTERGxwqnaWxtNIcnA68q+pa6cN4oq+Axqt7KzVBuM3Mrv+cq/d63Idi44tmVfQh0RES0maYLt0b0xdn9dyYmIiIhYqlWW3SX6Ekm7ATd11Ga7s5++ioiIaLyEnH7G9u0s5aPktrWc404CluvciIiIvii3qyIiIqKRspIT3TJqw7WzcTgiIvqkrOREREREIyXkRERERCMl5EREREQjJeREREREIyXkRERERCPl01XRLU/MWsC/nv/7Lp/3H2Pe0QvVRERE/F1WciIiIqKREnIiIiKikRJyIiIiopESciIiIqKREnIiIiKikRJyIiIiopESciIiIqKREnK6QNJNko5qdR0RERGxbPkywCWQdBKwpe3D247Z3rdFtZwPPGn7xFbMHxER0R9lJSciIiIaqTEhR9KXJE2VNF/SXyXtKWklSV+W9JikWZKukLRe6T9KkiUdJekJSTMlfbW07QOcABwiaYGk+8vxWyV9rDwfI+lOST+SNFfSRElvK8enSHq6fmtL0mqSvl/mmiHpbElrlLY9JD0p6fhy3nRJR5e2scBhwBdLLdcv432YJOnzkh6Q9KykyyWtXmvfX9J9peY/SNqh1mZJW9Zeny/plA7mGCtpvKTxC+fP7fK/VURExIrQiJAjaWvgk8DOtocAewOTgOOA9wG7A8OAOcBZ7U7fFdga2BP4uqRtbP8aOA243PZg2zsuYepdgAeA9YFLgMuAnYEtgcOBMyUNLn2/C2wF7FTahwNfr421MTC0HP8ocJakdW2PAy4GvldqOaATb8nBwD7Aa4EdgDHlfXojcC7w8VLzT4HrJK3WiTH/xvY426Ntj15jyDpdOTUiImKFaUTIARYDqwHbShpke5Ltx6h+mX/V9pO2XwROAg6UVN+LdLLthbbvB+4HlhRoOvK47fNsLwYuBzYDvmn7Rds3Ay8BW0oScAzwWduzbc+nClGH1sZaVM5dZPtXwAKq8LU8zrA9zfZs4HqqYEWp4ae277G92PYFwIvAW5ZznoiIiD6rERuPbT8q6TNUIWY7Sb8BPgeMBH4p6ZVa98XAa2qvn6o9fx4YTOfNqD1fWGppf2wwsCGwJjChyjsACFi51neW7Ze7UUtd+2saVp6PBI6S9Kla+6q19oiIiMZoykoOti+xvSvVL3JT3R6aAuxre53aY3XbUzszZA+WN5Mq8GxXq2Oo7c6GmJ6qZQpwarv3Y03bl5b256nCWJuNe2jeiIiIFa4RIUfS1pLeVfaWvEAVKBYDZwOnShpZ+m0o6b2dHHYGMEpSt98j268A5wA/krRRqWW4pL27UMvm3a2j1HCspF1UWUvSeyQNKe33AR+WtHLZfL17D8wZERHREo0IOVT7cb5DtWLyFLAR1aejfgxcB9wsaT5wN9Vm4c64svycJeneHqjxS8CjwN2S5gG30Pk9Nz+n2m80V9K1y1uA7fFU+3LOpNqE/ShlU3LxaeAAYC7VJ7qWe66IiIhWk92Td2VioNnota/3B78xrsvn/ceYd/RCNRER0d9ImmB7dG+M3ZSVnIiIiIhXacSnqwYSSSOAh5bQvK3tJ1ZkPREREX1VQk4/U0LM8n60PCIiYsBIyIluGbH+4OyviYiIPil7ciIiIqKREnIiIiKikRJyIiIiopESciIiIqKREnIiIiKikfLpquiWJ2c/x+cvu6vL533/0Lf2QjURERF/l5WciIiIaKSEnIiIiGikhJyIiIhopISciIiIaKSEnIiIiGikhJyIiIhopIScLpB0k6SjWl1HRERELFu+J2cJJJ0EbGn78LZjtvdtUS3nA0/aPrEV80dERPRHWcmJiIiIRmpMyJH0JUlTJc2X9FdJe0paSdKXJT0maZakKyStV/qPkmRJR0l6QtJMSV8tbfsAJwCHSFog6f5y/FZJHyvPx0i6U9KPJM2VNFHS28rxKZKert/akrSapO+XuWZIOlvSGqVtD0lPSjq+nDdd0tGlbSxwGPDFUsv1y3gfJkn6vKQHJD0r6XJJq9fa95d0X6n5D5J2qLW9UdL/lPfwynLuKT3zLxQREbFiNSLkSNoa+CSws+0hwN7AJOA44H3A7sAwYA5wVrvTdwW2BvYEvi5pG9u/Bk4DLrc92PaOS5h6F+ABYH3gEuAyYGdgS+Bw4ExJg0vf7wJbATuV9uHA12tjbQwMLcc/CpwlaV3b44CLge+VWg7oxFtyMLAP8FpgB2BMeZ/eCJwLfLzU/FPguhLAVgV+CZwPrAdcCry/o8EljZU0XtL45+fP6UQ5ERERK14jQg6wGFgN2FbSINuTbD9G9cv8q7aftP0icBJwoKT6XqSTbS+0fT9wP7CkQNORx22fZ3sxcDmwGfBN2y/avhl4CdhSkoBjgM/anm17PlWIOrQ21qJy7iLbvwIWUIWv5XGG7Wm2ZwPXUwUrSg0/tX2P7cW2LwBeBN5SHquUcxfZvgb4Y0eD2x5ne7Tt0WsOWXc5S4yIiOhdjdh4bPtRSZ+hCjHbSfoN8DlgJPBLSa/Uui8GXlN7/VTt+fPAYDpvRu35wlJL+2ODgQ2BNYEJVd4BQMDKtb6zbL/cjVrq2l/TsPJ8JHCUpE/V2lct7Qam2natbcpyzh8REdFyTVnJwfYltnel+kVuqttDU4B9ba9Te6xue2pnhuzB8mZSBZ7tanUMtd3ZENNTtUwBTm33fqxp+1JgOjBctRRGtTIVERHRLzUi5EjaWtK7JK0GvEAVKBYDZwOnShpZ+m0o6b2dHHYGMEpSt98j268A5wA/krRRqWW4pL27UMvm3a2j1HCspF1UWUvSeyQNAe6ies8+KWmV8j69uQfmjIiIaIlGhByq/TjfoVoxeQrYiOrTUT8GrgNuljQfuJtqs3BnXFl+zpJ0bw/U+CXgUeBuSfOAW+j8npufU+03mivp2uUtwPZ4qn05Z1Jtwn6UsinZ9kvAB6g2Pc+l2jh9A9WenYiIiH5Hr96CEfF3ku4BzrZ93pL6bLz5Nj78tHO7PPb3D31rd0qLiIiGkDTB9ujeGLspKznRAyTtLmnjcrvqKKqPn/+61XVFREQsj0Z8umogkTQCeGgJzdvafqIbw28NXEH1qa7HgANtT+/GeBERES2TkNPPlBCzvB8tX9bY44BxvTF2RETEipbbVREREdFIWcmJbtl0vbWyiTgiIvqkrOREREREIyXkRERERCMl5EREREQjJeREREREIyXkRERERCPl01XRLdPmPs9J147vdP+T3tcr39wdERHxD7KSExEREY2UkBMRERGNlJATERERjZSQExEREY2UkBMRERGNlJATERERjdRnQo6kmyQd1eo64u8kWdKWra4jIiJiebTke3IknQRsafvwtmO2921RLecDT9o+sRXzR0RERO/oMys5ERERET2pUyFH0pckTZU0X9JfJe0paSVJX5b0mKRZkq6QtF7pP6rc6jhK0hOSZkr6amnbBzgBOETSAkn3l+O3SvpYeT5G0p2SfiRprqSJkt5Wjk+R9HT91pak1SR9v8w1Q9LZktYobXtIelLS8eW86ZKOLm1jgcOAL5Zarl/G+7CZpGskPVOu+cxyfCVJJ0qaXOb4haSh7d6Lo0vtcyQdK2lnSQ+U6zuzNkfbtf9E0rOSHpa0Z619mKTrJM2W9KikY2pt50s6pfZ6D0lP1l5PkvT5Mu+zki6XtHqt/Qvl/Zkm6SOd+W8jIiKir1pmyJG0NfBJYGfbQ4C9gUnAccD7gN2BYcAc4Kx2p+8KbA3sCXxd0ja2fw2cBlxue7DtHZcw9S7AA8D6wCXAZcDOwJbA4cCZkgaXvt8FtgJ2Ku3Dga/XxtoYGFqOfxQ4S9K6tscBFwPfK7UcsJT3YWXgBmAyMKqMdVlpHlMe7wQ2BwYDZ7YbYhfgdcAhwOnAV4G9gO2AgyXt3q7vRGAD4BvANW0BErgUeJLqPT8QOK0egjrhYGAf4LXADqXutvD5eeDdpc69ljSApLGSxksa//y8OV2YOiIiYsXpzErOYmA1YFtJg2xPsv0Y8HHgq7aftP0icBJwoKT6Pp+TbS+0fT9wP7CkQNORx22fZ3sxcDmwGfBN2y/avhl4CdhSkoBjgM/anm17PlWIOrQ21qJy7iLbvwIWUIWvrngzVbD4gu3nbL9g+47SdhjwQ9sTbS8AvgIc2u69+FY552bgOeBS20/bngrcDryh1vdp4PRS7+XAX4H3SNqMKjh+qYx1H/Az4IguXMcZtqfZng1cTxUMoQo/59n+s+3nqP49O2R7nO3Rtkevufa6XZg6IiJixVnmxmPbj0r6DNUvve0k/Qb4HDAS+KWkV2rdFwOvqb1+qvb8eaoVjs6aUXu+sNTS/thgYENgTWBClXcAELByre8s2y93oxaoQtbkduO0GUa1wtNmMtV7W38v2tfe0bW0mWrb7cYbVh5tQa7e1pW/etn+32RY7RomtBs3IiKi3+rUnhzbl9jelSrYmOr20BRgX9vr1B6rl5WJZQ65/CX/g5lUIWG7Wh1DbXc2xHS2linAiHarM22mUb03bUYAL/PqINMVw1VLbGW8aeWxnqQh7dra3vPnqAJfm427MOd0qiBXHzciIqLf6tSeHEnvkrQa8AJVoFgMnA2cKmlk6behpPd2ct4ZwChJ3f50l+1XgHOAH0naqNQyXNLeXahl8070+yNVEPiOpLUkrS7p7aXtUuCzkl5b9gm17TnqaNWnMzYCjpM0SNJBwDbAr2xPAf4AfLvMvwPVHqOLy3n3AftJWk/SxsBnujDnFcAYSdtKWpNqL1BERES/1ZmQsRrwHaoVk6eofgGfAPwYuA64WdJ84G6qDbOdcWX5OUvSvV2quGNfAh4F7pY0D7iFzu+5+TnVfqO5kq5dUqeyN+gAqo3NT1Bt/j2kNJ8LXAj8HnicKgx+quuX8Tf3UG3+nQmcChxoe1Zp+xDVxudpwC+Bb9j+bWm7kGrv0yTgZqq9TJ1i+yaqDdG/o3ovf9eN+iMiIlpOr976Ea0maQzwsXJ7sM8btuW2Hvv9X3S6/0nv68r2oYiIaDpJE2z3yi+HfBlgRERENFJL/qxDXyVpBPDQEpq3tf3EiqwnIiIill9CTk0JMV39aHlP13A+cH4ra4iIiGiC3K6KiIiIRspKTnTLsHXWzGbiiIjok7KSExEREY2UkBMRERGNlJATERERjZSQExEREY2UkBMRERGNlE9XRbfMmPc8P/j1fcvsd/w+O/V6LREREXVZyYmIiIhGSsiJiIiIRkrIiYiIiEZKyImIiIhGSsiJiIiIRkrIiYiIiEYaMCFH0vmSTmnR3KdIminpqVbM3xmSbpX0sVbXERER0VMGTMhpFUmbAccD29reuNX1AEg6SdJFra4jIiKiNyXk9L6RwCzbT7e6kIiIiIGksSFH0hsk3StpvqTLgdXL8XUl3SDpGUlzyvNNS9tBkia0G+d4SdcuY66hkn5Rxpws6URJK0naC/gtMEzSAknnL2WMUZIs6WhJU0ptx0raWdIDkuZKOrPWf6Uyz2RJT5f5h7Yb6yhJT5RbZV8tbfsAJwCHlJrur5UxUtKd5T27WdIGnX/HIyIi+pZGhhxJqwLXAhcC6wFXAh8szSsB51GtsIwAFgJt4eE64LWStqkNd3gZZ2l+AgwFNgd2B44EjrZ9C7AvMM32YNtjOlH+LsDrgEOA04GvAnsB2wEHS9q99BtTHu8s8w6uXUebXYGtgT2Br0vaxvavgdOAy0tNO9b6fxg4GtgIWBX4fEcFShorabyk8c89O7cTlxQREbHiNTLkAG8BBgGn215k+yrgTwC2Z9m+2vbztucDp1IFE2y/CFxOFWyQtB0wCrhhSRNJWpkqkHzF9nzbk4AfAEcsZ+3fsv2C7ZuB54BLbT9teypwO/CG0u8w4Ie2J9peAHwFOFRS/e+RnWx7oe37gfuBeqDpyHm2/8/2QuAKYKeOOtkeZ3u07dFrDV1nOS8zIiKidzU15AwDptp27dhkAElrSvppuc0zD/g9sE4JKwAXAB+WJKqgckUJP0uyAdWqx+R2cw1fztpn1J4v7OD14PJ8WAdzrgK8pnas/mmu52vnLklX+0dERPRZTQ0504HhJai0GVF+Hk91C2cX22sD7yjHBWD7buAlYDeq2zfLulU1E1hEdfurPtfU7lxAJ0zrYM6XeXUoWhIvu0tERET/1tSQcxfVL/zjJK0i6QPAm0vbEKoVkbmS1gO+0cH5v6Da3/Ky7TuWNpHtxVS3dk6VNETSSOBzQG9/RPtS4LOSXitpMH/fZ/NyJ86dAYyS1NR//4iIiGaGHNsvAR+g2pg7h2rPzDWl+XRgDaoVmLuBX3cwxIXA9ix7FafNp6j2z0wE7gAuAc5druI771yq+n4PPA68UOrojCvLz1mS7u2F2iIiIlpOr962EgCS1gCeBt5o+5FW19OXbbbVtv7MGZcss9/x++zU+8VERES/I2mC7dG9MXYjV3J6wL8Cf0rAiYiI6L9WWXaXgUXSJKpNyO9rd/xBXr3Rt83HbV/cybEPA37aQdNk29t1rdKIiIhYmoScdmyPWsLxboeQEoY6FYgiIiKie3K7KiIiIhopKznRLa9Ze81sKo6IiD4pKzkRERHRSAk5ERER0UgJOREREdFICTkRERHRSAk5ERER0Uj5dFV0yzPzFzLu1geX2W/sHvmuw4iIWLGykhMRERGNlJATERERjZSQExEREY2UkBMRERGNlJATERERjZSQExEREY2UkBMRERGNlJCzFJLGSLqj1XV0h6RJkvZqdR0RERErWkJOL+jtcCTpfEmn9Nb4ERERTZCQ0yKSVm51DREREU2WkANI+rKkxyTNl/SQpPe/ulk/kfSspIcl7VlrGCNpYjnvcUmHSdoGOBt4q6QFkuaWvudL+g9Jv5L0HPBOSe+R9D+S5kmaIumkdnXtKukPkuaW9jGSxgKHAV8s41/fhetcqXatsyRdIWm90jZKkiUdJekJSTMlfXUJ44yVNF7S+AXPzuns9BEREStUQk7lMWA3YChwMnCRpE1K2y7ARGAD4BvANZLWk7QWcAawr+0hwNuA+2z/BTgWuMv2YNvr1Ob5MHAqMAS4A3gOOBJYB3gP8K+S3gcgaQRwE/ATYENgpzL+OOBi4Htl/AO6cJ3HAe8DdgeGAXOAs9r12RXYGtgT+HoJba9ie5zt0bZHDx66bhemj4iIWHEScgDbV9qeZvsV25cDjwBvLs1PA6fbXlTa/koVSABeAbaXtIbt6baX9Zcq/9P2nWWeF2zfavt/y+sHgEupAghUqzW32L60zD3L9n3dvNSPA1+1/aTtF4GTgAMl1f9Q68m2F9q+H7gf2LGbc0ZERLREQg4g6UhJ95XbQnOB7alWbgCm2nat+2RgmO3ngEOoVm2mS7pR0uuXMdWUdvPuIum/JT0j6dkyVtu8m1GtMPWkkcAva9f5F2Ax8Jpan6dqz58HBvdwDRERESvEgA85kkYC5wCfBNYvt5f+DKh0GS5JtVNGANMAbP/G9ruBTYCHyzgA9VBU1/74JcB1wGa2h1Lt5WmbawqwRSfH6awpVLfX1qk9Vrc9dTnHi4iI6LMGfMgB1qIKDc8ASDqaaiWnzUbAcZIGSToI2Ab4laTXSPqXsjfnRWAB1aoIwAxgU0mrLmPuIcBs2y9IejPVnp02FwN7STpY0iqS1pe0U238zZfjWs8GTi3BDkkbSnrvcowTERHR5w34kGP7IeAHwF1U4eGfgDtrXe4BXgfMpNo0fKDtWVTv3fFUqzqzqfbSfKKc8zvgQeApSTOXMv0ngG9Kmg98HbiiVtcTwH5ljtnAffx9f8zPgW3Lbadru3C5P6ZaObq5zHk31cbqiIiIxtGrt5tEdM3IrbfzV396xTL7jd1juxVQTURE9DeSJtge3RtjD/iVnIiIiGimVZbdJfqy8n06Dy2hedty2ysiImLAScjp50qIyce8IyIi2knIiW7ZcMga2W8TERF9UvbkRERERCMl5EREREQjJeREREREIyXkRERERCMl5EREREQj5dNV0S2zn3uBi+/661L7HPbWrVdQNREREX+XlZyIiIhopISciIiIaKSEnIiIiGikhJyIiIhopISciIiIaKSEnIiIiGikhJwuknSTpKNaXUdEREQsXb4nZykknQRsafvwtmO2921RLecDT9o+sRXzR0RE9DdZyYmIiIhGalTIkfQlSVMlzZf0V0l7SlpJ0pclPSZplqQrJK1X+o+SZElHSXpC0kxJXy1t+wAnAIdIWiDp/nL8VkkfK8/HSLpT0o8kzZU0UdLbyvEpkp6u39qStJqk75e5Zkg6W9IapW0PSU9KOr6cN13S0aVtLHAY8MVSy/XLeB8mSfqKpIckzZF0nqTVa+3HSHpU0mxJ10kaVo6rXMvTkp6V9ICk7XvuXygiImLFaUzIkbQ18ElgZ9tDgL2BScBxwPuA3YFhwBzgrHan7wpsDewJfF3SNrZ/DZwGXG57sO0dlzD1LsADwPrAJcBlwM7AlsDhwJmSBpe+3wW2AnYq7cOBr9fG2hgYWo5/FDhL0rq2xwEXA98rtRzQibfksPIebFHmPLG8T+8Cvg0cDGwCTC41A/wz8I7Sfx3gEGBW+4EljZU0XtL4eXPmdKKUiIiIFa8xIQdYDKwGbCtpkO1Jth8DPg581faTtl8ETgIOlFTfj3Sy7YW27wfuB5YUaDryuO3zbC8GLgc2A75p+0XbNwMvAVtKEnAM8Fnbs23PpwpRh9bGWlTOXWT7V8ACqvC1PM60PcX2bOBU4EPl+GHAubbvLe/HV4C3ShpV5h8CvB6Q7b/Ynt5+YNvjbI+2PXrtddddzvIiIiJ6V2NCju1Hgc9QhZinJV1WbsOMBH5ZbifNBf5CFYheUzv9qdrz54HBdN6M2vOFpZb2xwYDGwJrAhNqtfy6HG8zy/bL3ailbkrt+WSqVSzKz8ltDbYXUK3WDLf9O+BMqpWuGZLGSVp7OeePiIhoqcaEHADbl9jelSrYmOr20BRgX9vr1B6r257amSF7sLyZVIFnu1odQ213NsR0tZbNas9HANPK82lU7w8AktaiutU2FcD2GbbfBGxHddvqC12cNyIiok9oTMiRtLWkd0laDXiBKlAsBs4GTpU0svTbUNJ7OznsDGCUpG6/T7ZfAc4BfiRpo1LLcEl7d6GWzbsw5f+TtGnZZH0C1a00qPYNHS1pp/JenQbcY3uSpJ0l7SJpEPAc1fu4uAtzRkRE9BmNCTlU+3G+Q7Vi8hSwEdUv9x8D1wE3S5oP3E21Wbgzriw/Z0m6twdq/BLwKHC3pHnALXR+z83PqfYbzZV0bSf6XwLcDEwsj1MAbP8X8DXgamA61cbktn1Ba1MFsTlUt7RmAd/vZH0RERF9iuyevCMTfYGkScDHbN/S23Ntvs32/ta5Vy+1z2FvXd690xER0XSSJtge3RtjN2klJyIiIuJv8mcd+iFJI4CHltC87YqsJSIioq9KyOmHbD/B0j9aPmoFlRIREdFn5XZVRERENFJWcqJb1ltr9WwsjoiIPikrOREREdFICTkRERHRSAk5ERER0UgJOREREdFICTkRERHRSPl0VXTL3Odf5NoJE//h+Pve1JW/JRoREdHzspITERERjZSQExEREY2UkBMRERGNlJATERERjZSQExEREY2UkBMRERGNlJBTSDpf0iktmvsUSTMlPdWK+Wt1nCTpolbWEBER0VPyPTktJmkz4HhgpO2nW11PREREU2Qlp/VGArMScCIiInrWgA05kt4g6V5J8yVdDqxejq8r6QZJz0iaU55vWtoOkjSh3TjHS7p2GXMNlfSLMuZkSSdKWknSXsBvgWGSFkg6fyljXCDp+PJ8uCRL+kR5vaWk2ZJUXu8v6T5JcyX9QdIOtXGGSbq61PK4pOOWMN8gSZeWvqsu8w2NiIjoYwZkyCm/tK8FLgTWA64EPliaVwLOo1phGQEsBM4sbdcBr5W0TW24w8s4S/MTYCiwObA7cCRwtO1bgH2BabYH2x6zlDFuA/Yoz3cHJpafAO8AbrdtSW8EzgU+DqwP/BS4TtJqklYCrgfuB4YDewKfkbR3fSJJa1C9Py8CB9t+qV37WEnjJY2fN2f2Mi49IiKiNQZkyAHeAgwCTre9yPZVwJ8AbM+yfbXt523PB06lhAnbLwKXUwUbJG0HjAJuWNJEklYGDgG+Ynu+7UnAD4AjuljzbcBuJai8A/ge8PbStntpBzgG+Knte2wvtn0BVVh5C7AzsKHtb9p+yfZE4Bzg0No8awO/Bh6jCmKL2xdie5zt0bZHr73uel28jIiIiBVjoIacYcBU264dmwwgaU1JPy23leYBvwfWKWEF4ALgw+XW0BHAFSX8LMkGwKpt49fmGt6Vgm0/BiwAdgJ2owpW0yRtzatDzkjg+HKraq6kucBm5ZpHUt0aq7edALymNtVbgB2A77R7fyIiIvqVgRpypgPD2/awFCPKz+OBrYFdbK9NtWoCIADbdwMvUQWND7PsW1UzgUVUAaM+19TlqPs24EBgVdtTy+sjgXWB+0qfKcCpttepPda0fWlpe7xd2xDb+9XmuBn4NvBfkurhJyIiol8ZqCHnLuBl4DhJq0j6APDm0jaEah/OXEnrAd/o4PxfUO3Tedn2HUubqNzuuQI4VdIQSSOBzwHL8300twGfpFpdArgV+BRwR+220jnAsZJ2UWUtSe+RNAT4IzBP0pckrSFpZUnbS9q5Xc3fAy6hCjobLEedERERLTcgQ07ZSPsBYAwwh2rPzDWl+XRgDaoVmLup9qe0dyGwPctexWnzKeA5qs3Cd1AFiHOXo/TbqEJYW8i5A1iz9hrb46n25ZxJdW2PUl1nW+A6gOqW1+NU1/gzqk3Rr2L7W1Sbj28pYS8iIqJfUbZddF359NHTwBttP9Lqelppy23/yd+/8D//4fj73rR5C6qJiIj+RtIE26N7Y+wBuZLTA/4V+NNADzgRERF9Wf6sQxdJmkS1Cfl97Y4/yKs3F7f5uO2LOzn2YVTfa9PeZNvbda3SiIiIgS0hp4tsj1rC8W6HkBKGOhWIIiIiYulyuyoiIiIaKSEnIiIiGim3q6Jb1llztXySKiIi+qSs5EREREQjJeREREREIyXkRERERCMl5EREREQjZeNxdMu8hS/x2/+d8rfX7/6nzVpYTURExN9lJSciIiIaKSEnIiIiGikhJyIiIhopISciIiIaKSEnIiIiGikhJyIiIhopIaeQNEbSHa2uo6dJOlvS11pdR0RExIqW78lpEEljgI/Z3rXtmO1jW1dRRERE62QlJyIiIhppwIUcSV+W9Jik+ZIekvT+VzfrJ5KelfSwpD1rDWMkTSznPS7psGXMs7Kk70uaWc77f5IsaZXSPknSXrX+J0m6qPb6LZL+IGmupPsl7bG0WiRtA5wNvFXSAklzS9/zJZ1SO/cYSY9Kmi3pOknDam2WdKykRyTNkXSWJHX9XY6IiGi9ARdygMeA3YChwMnARZI2KW27ABOBDYBvANdIWk/SWsAZwL62hwBvA+5bxjzHAPsDbwBGAwd2tkBJw4EbgVOA9YDPA1dL2nBJtdj+C3AscJftwbbX6WDcdwHfBg4GNgEmA5e167Y/sDOwY+m3dwfjjJU0XtL4Z+fM7uxlRURErFADLuTYvtL2NNuv2L4ceAR4c2l+Gjjd9qLS9lfgPaXtFWB7SWvYnm77wWVMdXAZa4rt2VThorMOB35l+1elzt8C44H9lrOWNocB59q+1/aLwFeoVn5G1fp8x/Zc208A/w3s1H4Q2+Nsj7Y9eui663XhsiIiIlacARdyJB0p6b5yG2gusD3Vyg3AVNuudZ8MDLP9HHAI1UrJdEk3Snr9MqYaBkypvZ7chTJHAge11Vjq3BXYZDlrqdf0tzpsLwBmAcNrfZ6qPX8eGNyFuiMiIvqMARVyJI0EzgE+Caxfbun8GWjbdzK83R6UEcA0ANu/sf1uqts8D5dxlmY6UP+T3CPatT8HrFl7vXHt+RTgQtvr1B5r2f7OMmqpB7SOTKMKUACUW1/rA1OXcV5ERES/M6BCDrAWVRB4BkDS0VQrOW02Ao6TNEjSQcA2wK8kvUbSv5RQ8CKwAFi8jLmuKGNtKmld4Mvt2u8DDi1ztd+zcxFwgKS9ywbm1SXtUcZaWi0zgE0lrbqEmi4Bjpa0k6TVgNOAe2xPWsa1RERE9DsDKuTYfgj4AXAXVSD4J+DOWpd7gNcBM4FTgQNtz6J6n46nWgmZDewOfGIZ050D/Aa4H7gXuKZd+9eALYA5VBugL6nVOQV4L3ACVSCbAnyh1LG0Wn4HPAg8JWlmB9f/X2Xeq6lWmrYADl3GdURERPRLevUWlOgtZXPv48Ag2y+3uJwes9V2O/isy2782+t3/9NmS+kdERHxapIm2B7dG2MPqJWciIiIGDgScrqh/F2oBR08zm51bREREQNd/nZVN5S/C9Wpvw1VNvfm24MjIiJWkKzkRERERCMl5EREREQj5XZVdMvaa6yaT1RFRESflJWciIiIaKSEnIiIiGikhJyIiIhopISciIiIaKRsPI5uWfDCIu58ePrfXr/99Zu0sJqIiIi/y0pORERENFJCTkRERDRSQk5EREQ0UkJORERENFJCTkRERDRSQk5EREQ0Up8IOZIelLRHB8f3kPRkD81hSVv2xFgrcuzeJulWSR9rdR0RERE9rU98T47t7VpdQ0RERDRLn1jJid6nSv69IyJiwOgTv/QkTZK0l6Q1JJ0vaY6kh4Cd2/UbJulqSc9IelzScbW2N0u6S9JcSdMlnSlp1S7Wcb6ksyTdKGm+pHskbdHFMXaVNEXSO8vrj0j6S7mm30gaWY6fJekH7c69XtJnJB0t6fra8UclXVF7PUXSTuX52yT9SdKz5efbav1ulXSqpDuB54HNJb1b0sOl/5mAav23lHRbaZsp6fKuXHtERERf0idCTs03gC3KY2/gqLaGsgpxPXA/MBzYE/iMpL1Ll8XAZ4ENgLeW9k8sRw0fAk4G1gUeBU7t7ImllkuBD9r+b0nvA04APgBsCNxe2gEuAD7UtroiaYNS86XAbcBuklaStAkwCHh76bc5MBh4QNJ6wI3AGcD6wA+BGyWtXyvrCGAsMAR4FrgaOJHqfXqsbdziW8DN5do3BX6yhOscK2m8pPFz58zq7NsTERGxQvW1kHMwcKrt2banUP3ybrMzsKHtb9p+yfZE4BzgUADbE2zfbftl25OAnwK7L0cN19j+o+2XgYuBnTp53kHAOGA/238sxz4OfNv2X8p4pwE7SRpZ+jxLFWwo13Gr7Rnl2uaXuXcHfgNMlfT68vp2268A7wEesX1hue5LgYeBA2p1nW/7wTL/vsBDtq+yvQg4HXiq1ncRMBIYZvsF23d0dKG2x9kebXv0Ouuu31GXiIiIlutrIWcYMKX2enLt+UhgWLkdNVfSXKpVktcASNpK0g2SnpI0jypQbLAcNdR/6T9PtWrSGZ8BrrD9v+1q/nGt3tlUt4eGl/YLgMPL88OBC2vn3gbsAbyjPL+VKuDsXl5D9X7V3yPK6+G11/X381Xvr223a/9iqe+P5RNvH1nK9UZERPRpfS3kTAc2q70eUXs+BXjc9jq1xxDb+5X2/6BaxXid7bWpApBYcQ4C3ifpM+1q/ni7mtew/YfSfhHwXkk7AtsA19bObQs5u5Xnt/GPIWcaVZCqGwFMrb127fmr3l9Jqr+2/ZTtY2wPo1qF+vf++tH4iIiIvhZyrgC+ImldSZsCn6q1/RGYJ+lLZYPyypK2l9S2OXkIMA9YUG7r/OuKLZ1pVLeejpPUthfobKrr2Q5A0lBJB7WdYPtJ4E9UKzhX215YG+824J3AGqXf7cA+VHtv/qf0+RWwlaQPS1pF0iHAtsANS6jxRmA7SR+QtApwHLBxW6Okg8r7DjCHKiAtXo73IiIiouX6Wsg5mep2y+NUG2D/dvvG9mKqvSY7lfaZwM+AoaXL54EPU+1lOQdY4Z8Msv0EVdD5kqSP2f4l8F3gsnIL7c9U+2LqLgD+iVffqsL2/wELqMINtucBE4E7y3uB7VnA/sDxwCyq20372565hPpmUq04faf0fx1wZ63LzsA9khYA1wGftv34crwVERERLadqW0a0iqR3UN22GlU2E/crr99+R//8ql//7fXbX79JC6uJiIj+RtIE26N7Y+y+tpIzoEgaBHwa+Fl/DDgRERF92YALOeVTQws6eBy2lHN2W8I5C7pRxzbAXGATqo9yR0RERA/qE3+7akVanr+TZft2Ov9R8s6O+RdgrZ4cMyIiIv5uwK3kRERExMCQkBMRERGNNOBuV0XPGrz6oHyiKiIi+qSs5EREREQjJeREREREIyXkRERERCMl5EREREQjJeREtzz/4svc+9hM7n2swz+XFRER0TIJOREREdFICTkRERHRSAk5ERER0UgJOREREdFICTkRERHRSAk5ERER0UgJOSuQpAcl7dHqOiIiIgaC/IHOFcj2dq2uoY0kA6+z/Wira4mIiOgNWckZYCQl2EZExICQkLMCSZokaS9Jb5Y0XtI8STMk/XAZ542SZEljJU2TNF3S8bX2N0u6S9Lc0nampFVr7Zb0/yQ9Ajwi6fel6X5JCyQdIunPkg6onTNI0kxJO/Xw2xAREbFC5P/Vt8aPgR/bvlDSYGD7Tp73TuB1wObA7yTdb/sWYDHwWWA8sClwE/AJ4PTaue8DdgEW2l5Yblft2Ha7StJI4HDg+tJ/P2C67fvaFyFpLDAWYONhm3ay9IiIiBUrKzmtsQjYUtIGthfYvruT551s+znb/wucB3wIwPYE23fbftn2JOCnwO7tzv227dm2Fy5h7IuA/SStXV4fAVzYUUfb42yPtj163fXW72TpERERK1ZCTmt8FNgKeFjSnyTt38nzptSeTwaGAUjaStINkp6SNA84DdhgKef+A9vTgDuBD0paB9gXuLiTdUVERPQ5CTktYPsR2x8CNgK+C1wlaa1OnLpZ7fkIYFp5/h/Aw1SfllobOAFQ+2k7Mf4FVLesDgLusj21E+dERET0SQk5LSDpcEkb2n4FmFsOL+7EqV+TtKak7YCjgcvL8SHAPGCBpNcD/9qJsWZQ7e2puxZ4I/Bp4BedGCMiIqLPSshpjX2AByUtoNqEfKjtFzpx3m3Ao8B/Ad+3fXM5/nngw8B84Bz+Hn6W5iTggvKJrIMByn6dq4HXAtd0/nIiIiL6nny6agWyPao8vWU5hzjX9rgOxv098Pp2h79ea29/6wrbZwNndzDHE8AvbS9YzhojIiL6hISc+BtJ61Ftij6i1bVERER0V25X9RGSDitfzNf+8eAKmv8Yqk9g3VRWhiIiIvq1rOT0EbYvZukf2f6HW049PP85VPt5IiIiGiErOREREdFICTkRERHRSLldFd2y5mqr8MYt2n+5ckREROtlJSciIiIaKSEnIiIiGikhJyIiIhopISciIiIaKSEnumXhS4v58xNz+PMTc1pdSkRExKsk5EREREQjJeREREREIyXkRERERCMl5EREREQjJeREREREIyXkRERERCMl5EREREQjJeSsQJIelLRHq+sAkLSbpL8upf18SaesyJoiIiJ6Uv4K+Qpke7tW19DG9u3A1q2uIyIiordkJSciIiIaKSFnBZI0SdJekt4sabykeZJmSPphJ87dVdIfJM2VNEXSmHJ8qKRfSHpG0mRJJ0paSdJqpe/2tTE2lLRQ0kaS9pD0ZK3tDZLulTRf0uXA6r3xHkRERKwoCTmt8WPgx7bXBrYArlhaZ0kjgJuAnwAbAjsB95XmnwBDgc2B3YEjgaNtvwhcA3yoNtTBwG22n243/qrAtcCFwHrAlcAHl1LP2BLSxs+ZPXPZVxsREdECCTmtsQjYUtIGthfYvnsZ/Q8DbrF9qe1FtmfZvk/SysAhwFdsz7c9CfgBcEQ57xJeHXI+XI619xZgEHB6Gf8q4E9LKsb2ONujbY9ed70NOnO9ERERK1xCTmt8FNgKeFjSnyTtv4z+mwGPdXB8A2BVYHLt2GRgeHn+O2ANSbtIGkm1AvTLDsYZBky17XbjRERE9FsJOS1g+xHbHwI2Ar4LXCVpraWcMoXqtlZ7M6lWhUbWjo0AppZ5XqG6FfYhqlWcG2zP72Cc6cBwSWo3TkRERL+VkNMCkg6XtGEJIXPL4cVLOeViYC9JB0taRdL6knayvZgqxJwqaUhZrfkccFHt3EuobmkdRse3qgDuAl4GjivjfwB483JfYERERB+QkNMa+wAPSlpAtQn5UNsvLKmz7SeA/YDjgdlUm453LM2fAp4DJgJ3UAWZc2vn3lPah1FtXu5o/JeADwBjgDlUoeia5b24iIiIvkCv3oYR0TXb7fAGX37D7wDYfsS6La4mIiL6G0kTbI/ujbGzkhMRERGNlJDTR0g6TNKCDh4Ptrq2iIiI/ih/u6qPsH0x1QbjiIiI6AFZyYmIiIhGykpOdMsaq66cDccREdEnZSUnIiIiGikhJyIiIhopISciIiIaKSEnIiIiGikhJ7rlhUWL+b9pz/J/055tdSkRERGvkpATERERjZSQExEREY2UkBMRERGNlJATERERjZSQExEREY2UkBMRERGNlJDTjqQHJe2xguccIWmBpJWX0H6SpIt6ae4FkjbvjbEjIiJaKX+gsx3b27VgzieAwSt63jJ3S+aNiIjobVnJiYiIiEZKyGlH0iRJe0l6s6TxkuZJmiHph5049y2S/iBprqT767e9JN0q6duS/ijpWUn/KWm90jZKkiWtUl6/VtJtkuZL+i2wQRfnOaW0L5B0vaT1JV1cruVPkkbV+lvSluX5GpJ+IGlyqfEOSWt05/2MiIholYScJfsx8GPbawNbAFcsrbOk4cCNwCnAesDngaslbVjrdiTwEWAY8DJwxhKGuwSYQBVuvgUc1cV5DgWOAIaX2u8Cziv9/wJ8Ywnzfh94E/C20veLwCtLu+6IiIi+KiFnyRYBW0rawPYC23cvo//hwK9s/8r2K7Z/C4wH9qv1udD2n20/B3wNOLj9ZmNJI4Cdga/ZftH274HruzjPebYfs/0scBPwmO1bbL8MXAm8oX3xklaiCmCftj3V9mLbf7D9Ygd9x5ZVrvFzZs1axtsSERHRGgk5S/ZRYCvg4XKLZ/9l9B8JHFRuIc2VNBfYFdik1mdK7flkYBDtbkVRrfLMKUGo3rcr88yoPV/YweuONhtvAKwOPNbx5f2d7XG2R9seve766y+re0REREvk01VLYPsR4ENlheMDwFWS1m8XPuqmUK3UHLOUYTerPR9BtVo0s93x6cC6ktaqzTUCcBfmWR4zgReobm/d38NjR0RErHBZyVkCSYdL2tD2K8DccnjxUk65CDhA0t6SVpa0uqQ9JG1a63O4pG0lrQl8E7jK9qvGtD2Z6vbTyZJWlbQrcEAX5+mycp3nAj+UNKyM/VZJq3Vn3IiIiFZJyFmyfYAHJS2g2oR8qO0XltTZ9hTgvcAJwDNUKy5f4NXv8YXA+cBTVLeGjlvCcB8GdgFmU20S/kUX51lenwf+F/hTmfu7PTRuRETECifby+4V3SbpVuAi2z9rdS09afsd3+BrbroVgK2GDW1tMRER0e9ImmB7dG+Mnf+XHhEREY2UkNMFkg4rX7DX/vFgq2uLiIiIV8unq7rA9sXAxct57h49W01EREQsTVZyIiIiopESciIiIqKRcrsqumX1QSvnU1UREdEnZSUnIiIiGikhJyIiIhopISciIiIaKSEnIiIiGikhJ7rlpZdfYdLTC5j09IJWlxIREfEqCTkRERHRSAk5ERER0UgJOREREdFICTkRERHRSAk5ERER0UgJOREREdFIAy7kSHpQ0h4raK4TJP2sl8a2pC17Y+yIiIgmGHB/oNP2ditwrtNW1Fzx/9u79yitqvOO498fjIAwgMog7ZgI8UIudClRvKSJkayQqAmpqSYGRbs0Jmps4mqWWblpWnLRKE2pdJkmwcSoEa8ksZrExrZRjFESwBpdRNR4QRBBQUYYLirw9I+zp54ZZ+Z9h+G9nfl91jqL9z377H328+6ZWQ97n/MeMzOzzgbcTE61SCpEAlmUOMzMbOAZcEmOpGckTZN0pKQlkjZKWitpTol6E9IS0TmSVkt6XtKFufJZkhZIul7SRuDMtO/6LvXPkrRS0gZJ50k6QtLDktokXdnlnJ+U9Gg69teSxnfp1jRJT6Ty70pSqjdI0sWSVkh6QdJ1kkansqmSVnX3mfQUx6590mZmZrU14JKcnLnA3IgYBRwI3FJmvfcBBwMfBL7ckRwkJwILgL2A+T3UPyrV/wRwBXARMA2YBJwi6VgASR8FvgqcBIwFfgvc2KWt6cARwKHAKcBxaf+ZaXsfcADQDFxJ+cqJw8zMrK4N5CTnNeAgSS0R0R4Ri8qs9/WI2BwRjwA/Bk7NlT0QEbdFxM6I2NpD/W9GxLaIuAvYDNwYES9ExHNkicw703HnAt+OiEcjYjtwKTC5y2zOZRHRFhHPAncDk9P+mcCciHgqItqBrwAz+rD01GscaTZriaQl69evK7NJMzOz6hrISc7ZwERguaTFkqaXWW9l7vUKoLWHsp6szb3e2s375vR6PDA3LWO1AS8BAvbLHb8m93pLrm5r6lu+n03AuDL6ByXiiIh5ETElIqaMGdNSZpNmZmbVNWCTnIh4IiJOBfYFLgcWSBpRRtU3517vD6zON7sbu7gSODci9spte0bE/WXUXU2WJOX7uZ0sodoMDO8okDSYbDksb3fGYWZmVhMDNsmRdLqksRGxE2hLu3eUUfVrkoZLmgScBdxcoS5+H/hKOg+SRkv6eJl1bwQ+L+ktkprJlrpuTstejwPDJH1Y0h7AxcDQCvTfzMyspgby7cHHA3MkDSdbzpkREdvKqLcQ+DNZgviddG3NbhcRP08Jyk3pOpyXgf8Cbi2j+tVkS1b3AsOAXwOfS+2+LOl84IfAYGA2sKqHdszMzBqWIrwyUQ5JE4CngT3SjIgBh0w+LG6/614AJuzbXOJoMzOzziQtjYgplWh7wC5XmZmZWbE5ycmRNFNSezfbslr3zczMzPpmIF+T8wYRMZ/ev/xO1eqLmZmZ9Y9ncszMzKyQnOSYmZlZIXm5yvplSNMg31VlZmZ1yTM5ZmZmVkhOcszMzKyQnOSYmZlZITnJMTMzs0JykmP98tqOnazesJXVG7bWuitmZmadOMkxMzOzQnKSY2ZmZoXkJMfMzMwKyUmOmZmZFZKTHDMzMyskJzlmZmZWSE5yzMzMrJCc5FSYpGWSpu6mtp6RNG13tFXNts3MzGrBTyGvsIiYVOs+dCXpGmBVRFxc676YmZlVimdyzMzMrJCc5FRYxzKQpCMlLZG0UdJaSXPKqHuGpBWS1ku6qEvZIElflvRkKr9F0j658lslrZH0sqR7JU1K+88BZgJflNQu6Y5cs5MlPZzq3Cxp2G76GMzMzKrOSU71zAXmRsQo4EDglt4OlvQO4HvAGUArMAZ4U+6QC4CPAsem8g3Ad3PldwIHA/sCDwLzASJiXno9OyKaI+IjuTqnAMcDbwEOAc7soW/npIRtyfp160rFbWZmVhNOcqrnNeAgSS0R0R4Ri0oc/zHgFxFxb0S8AnwN2JkrPxe4KCJWpfJZwMckNQFExNURsSlXdqik0SXO+W8RsToiXgLuACZ3d1BEzIuIKRExZUxLS4kmzczMasNJTvWcDUwElktaLGl6ieNbgZUdbyJiM7A+Vz4e+LmkNkltwKPADmCcpMGSLktLWRuBZ1KdUhnJmtzrLUBziePNzMzqlu+uqpKIeAI4VdIg4CRggaQxKXnpzvPA2zveSBpOtmTVYSXwyYj4XdeKks4ATgSmkSU4o8mWs9TRnf5FY2ZmVv88k1Mlkk6XNDYidgJtafeOXqosAKZLeo+kIcA36Dxe3wcukTQ+tT9W0ompbCTwCtnMz3Dg0i5trwUO6E88ZmZm9c5JTvUcDyyT1E52EfKMiNjW08ERsQz4e+AGslmdDcCq3CFzgduBuyRtAhYBR6Wy64AVwHPAn1JZ3o+Ad6Slrtv6GZeZmVldUoRXLmzXHfrOw+LO32QrZq1771nj3piZWaORtDQiplSibc/kmJmZWSE5yakhSTPTF/J13ZbVum9mZmaNzndX1VBEzCd9SZ+ZmZntXp7JMTMzs0LyTI71yx6DB/mCYzMzq0ueyTEzM7NCcpJjZmZmheQkx8zMzArJSY6ZmZkVkpMcMzMzKyQnOdYv23f6sSBmZlafnOSYmZlZITnJMTMzs0JykmNmZmaF5CTHzMzMCslJjpmZmRWSkxwzMzMrJCc5VSJpmaSpNe7D/pLaJQ0u49ipklZVo19mZmaV4KeQV0lETKqDPjwLNNe6H2ZmZtXgmZwGI8mJqZmZWRmc5FSJpGckTZN0pKQlkjZKWitpTol6EySFpLMlPQv8Ju2/VdIaSS9LulfSpFydPSX9i6QVqfy+tK+jraZ03FmSHpW0SdJTks6t6IdgZmZWRU5yqm8uMDciRgEHAreUWe9Y4O3Acen9ncDBwL7Ag8D83LHfAQ4H/hrYB/gisLObNl8ApgOjgLOAf5V0WKmOSDonJWpL1q9bV2b3zczMqstLH9X3GnCQpJaIWAcsKrPerIjY3PEmIq7ueC1pFrBB0mhgE/BJ4OiIeC4dcn86rlODEfHL3NuFku4CjiFLmnoUEfOAeQCTDzvcD68yM7O65Jmc6jsbmAgsl7RY0vQy663seCFpsKTLJD0paSPwTCpqSdsw4MlSDUo6QdIiSS9JagM+lOqbmZk1PCc5VRYRT0TEqWTLTJcDCySNKKdq7vVpwInANGA0MCHtF7AO2Ea2FNYjSUOBn5ItbY2LiL2AX6U2zMzMGp6TnCqTdLqksRGxE2hLu3f0sZmRwCvAemA4cGlHQWr3amCOpNY06/OulNTkDQGGAi8C2yWdAHywzwGZmZnVKSc51Xc8sExSO9lFyDMiYlsf27gOWAE8B/yJN17X8wXgEWAx8BLZjFGnsY6ITcAFZBc+byCbHbq9j/0wMzOrW4rwdaO26yYfdng89ODSWnfDzMwalKSlETGlEm17JsfMzMwKyUlOHZA0Mz1Tquu2rNZ9MzMza1T+npw6EBHz6fxlfmZmZtZPnskxMzOzQnKSY/3SNMhfq2NmZvXJSY6ZmZkVkpMcMzMzKyQnOWZmZlZITnLMzMyskJzkmJmZWSE5ybF+2enHgpiZWZ1ykmNmZmaF5CTHzMzMCslJjpmZmRWSkxwzMzMrJCc5ZmZmVkhOcszMzKyQnOSYmZlZITnJ6YakZZKm1rof3ZE0S9L1VTjPNZK+VenzmJmZVUpTrTtQjyJiUq37YGZmZv3jmRwzMzMrJCc53ZD0jKRpko6UtETSRklrJc0pUe9aSRem1/tJCknnp/cHSXpJktL76ZIektQm6X5Jh+TaaZX0U0kvSnpa0gU9nG8PSTemY4f0Vi8tc90i6TpJm9KS3JRc+TslPZjKbgaG9RLnOelzWbLuxXVlfqpmZmbV5SSnd3OBuRExCjgQuKXE8QuBqen1scBT6V+A9wK/jYiQdBhwNXAuMAb4AXC7pKGSBgF3AH8E9gPeD/yDpOPyJ5K0J3Ab8ApwCrC9jHp/A9wE7AXcDlyZ2hqS2voJsA9wK3ByT0FGxLyImBIRU1rGtpT4SMzMzGrDSU7vXgMOktQSEe0RsajE8QuBY1Ki8l5gNvDuVHZsKgf4NPCDiPh9ROyIiGvJkpWjgSOAsRHxjYh4NSKeAq4CZuTOMwr4T+BJ4KyI2FFmvfsi4lfp+J8Ah6b9RwN7AFdExGsRsQBY3IfPyczMrO44yend2cBEYLmkxZKm93ZwRDwJtAOTgWOAXwCrJb2VzknOeODCtFTVJqkNeDPQmspau5R9FRiXO9XRwCHAZRH//xjwcuqtyb3eAgyT1JTO+1yuLYAVvX80ZmZm9c13V/UiIp4ATk0zMycBCySNiYjNvVRbCHwMGBIRz0laCPwdsDfwUDpmJXBJRFzStbKkdwFPR8TBvZzjLuBh4H8kTY2ItanNUvV68jywnyTlEp39yWaKzMzMGpJncnoh6XRJYyNiJ9CWdu8oUW0h8Fng3vT+HuBzZEtFHXWvAs6TdJQyIyR9WNJI4A/ARklfkrSnpMGS/krSEfmTRMRs4AayRKel3Ho9eIDsmp4LJDVJOgk4sox6ZmZmdctJTu+OB5ZJaie7CHlGRGwrUWchMJLXk5z7gOG590TEErLrcq4ENgB/Bs5MZTuAj5AteT0NrAN+CIzueqKI+CbZBcP/ncrLqtdNO6+SzVSdmfrzCeBnpeqZmZnVM3W+DMOsbw47/PB4cOnSWnfDzMwalKSlETGl9JF955kcMzMzKyQnOX0kaaak9m62ZbXum5mZmb3Od1f1UUTMB+bXuh9mZmbWO8/kWL8Myp5SYWZmVnec5JiZmVkhOckxMzOzQvIt5NYvkjYBj9W6HxXWQva9Q0XmGIvBMRbDQItxfESMrcRJfOGx9ddjlfp+g3ohaYljbHyOsRgcYzFUK0YvV5mZmVkhOckxMzOzQnKSY/01r9YdqALHWAyOsRgcYzFUJUZfeGxmZmaF5JkcMzMzKyQnOWZmZlZITnLMzMyskJzk2C6RtI+kn0vaLGmFpNNq3adySLpH0rbc0+Mfy5W9X9JySVsk3S1pfK5Mki6XtD5ts6XXH9wlaUKqsyW1Ma2KMX1W0hJJr0i6pktZxWKSdFoa+82SbpO0TzXjS/2L3Fi2S/pao8WXzjVU0o/S+TZJ+l9JJ+TKizCOPcZYsLG8XtLzkjZKelzSp3JlDT+OvcVYl+MYEd689XkDbgRuBpqB9wAvA5Nq3a8y+n0P8Klu9rekGD4ODAP+GViUKz+X7Jud3wTsB/wJOC9X/gAwB9gTOBloA8ZWKaaTgI8C3wOuqUZMwCRgE/De9DNwA3BTleObAATQ1EO9hogvnW8EMCvFNAiYns4/oUDj2FuMRRrLScDQ9PptwBrg8KKMY4kY624cK/IBeCv2lv5YvQpMzO37CXBZrftWRt/vofsk5xzg/i4xbgXelt7fD5yTKz+74w8UMBF4BRiZK/9t/pe3SrF9i85JQMViAi4FbsiVHZh+JkbuzphKxFfqD2pDxddN/x9Of+gLNY49xFjIsQTeCjwPnFLUcewSY92No5erbFdMBHZExOO5fX8ky7QbwbclrZP0O0lT075JZDEAEBGbgSd5PaZO5XSOdxLwVERs6qG8VioZU9e2nyQlvrux/+VaIWmVpB9Lasntb9j4JI1L51rWTV8KMY5dYuxQiLGU9O+StgDLyRKAX3XTj4Yexx5i7FA34+gkx3ZFM9m0a97LwMga9KWvvgQcQDZVOg+4Q9KBlI6pa/nLQHNaT67Xz6OSMdVDzOuAI4DxZFPlI4H5ufKGjE/SHmRxXBsRy8voS8PF2U2MhRrLiDg/tX8M8DOyGYpCjWMPMdbdODrJsV3RDozqsm8U2XppXYuI30fEpoh4JSKuBX4HfIjSMXUtHwW0RzZvWq+fRyVjqnnMEdEeEUsiYntErAU+C3xQUke/Gi4+SYPIln5fJYunnL40VJzdxVjEsYyIHRFxH9n1J5/ZhX42XIz1OI5OcmxXPA40STo4t+9QOk87N4oARNb3Qzt2ShpBtubbEVOncjrHuww4QNLIHsprpZIxdW37AGAo2c9GrXR8fXvH3RoNFV/63+yPgHHAyRHxWg99adhx7CXGrhp6LLto4vXxKsQ4dqMjxq5qP46VvijJWzE34CayO6xGAO+mAe6uAvYCjiO7s6EJmAlsJrtwbmyK4eRUfjmd73w4D3iUbJmrNf3C5e8KWAR8J9X9W6p7d1VTOu+3yf6H3BFfxWIiWx/fSDZVPQK4nsrdldNTfEelsRsEjCG72+/uRosv15/vpz41d9lfiHEsEWMhxhLYF5hBtrQymOzvzWbgxKKMY4kY624cK/YL663YG7APcFv64X4WOK3WfSqjz2OBxWTTm23pF+oDufJpZBfRbSW7C2tCrkzAbOCltM0mPfstlU9IdbaS3SI5rYpxzSL7H1N+m1XpmIDT0thvBv4D2Kea8QGnAk+n8z8PXAf8RaPFl841PsW1jWxavmObWaBx7DHGoowl2d+YhWR/XzYCjwCfzpUXYRx7jLEex9EP6DQzM7NC8jU5ZmZmVkhOcszMzKyQnOSYmZlZITnJMTMzs0JykmNmZmaF5CTHzMzMCslJjpnZbiQpJB1Uwfbb07e9mlkJTnLMbEBIyUHHtlPS1tz7mT3UmSpp1W7uxwck3S1pk6T1kh6S9CVJw8qpHxHNEfHU7uyTWVE5yTGzASElB80R0Uz2rakfye2bX6r+7iDp48AC4AZgfESMAT5B9oDDN1ejD2YDiZMcMxvQJA2VdIWk1Wm7Iu0bAdwJtOZmfFolHSnpAUltkp6XdKWkIWWcR8Ac4BsRcVVEvAQQEY9FxOci4ol0XK/t55fDJF0j6buSfplmhn4vqbsHJZoNSE5yzGyguwg4GphM9pTjI4GLI2IzcAKwOjfjsxrYAXweaAHeBbwfOL+M87yVbMbmpyWO62v7pwJfB/YG/gxcUkZfzAYEJzlmNtDNJJtdeSEiXiRLGM7o6eCIWBoRiyJie0Q8A/wAOLaM87Skf9d07JB0U5qx2SLpjF1s/2cR8YeI2A7MJ0vWzAxoqnUHzMxqrBVYkXu/Iu3rlqSJZMtOU4DhZH9Hl5ZxnvXp378ke1IzETEjtXkfMHgX21+Te70FaC6jL2YDgmdyzGygWw2Mz73fP+0DiG6O/x6wHDg4IkYBXwVUxnmWA88BJ5U4blfbN7MunOSY2UB3I3CxpLGSWoB/BK5PZWuBMZJG544fCWwE2iW9DfhMOSeJiAAuBP5J0qcl7a3MwcC4/rZvZm/kJMfMBrpvAUuAh4FHgAfTPiJiOVkS9FS6dqYV+AJwGrAJuAq4udwTRcTNwCnA6cBKYB1wCzAPuDUdtsvtm1lnyv5zYWZmZlYsnskxMzOzQnKSY2ZmZoXkJMfMzMwKyUmOmZmZFZKTHDMzMyskJzlmZmZWSE5yzMzMrJCc5JiZmVkh/R/Y5HJkoQ2CEwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 576x864 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(8,12))\n",
    "sns.barplot(data=total_gain, y='feature', x='total gain', orient='h', palette='Blues_r')\n",
    "plt.ylabel('');\n",
    "plt.xlabel('Total Gain', fontsize=12)\n",
    "plt.yticks(fontsize=12)\n",
    "plt.xticks(fontsize=12)\n",
    "plt.title('Total Gain', fontsize=18)\n",
    "plt.tight_layout()\n",
    "plt.savefig(fname='total_gain', dpi=180)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x864 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(8,12))\n",
    "sns.barplot(data=total_cover, y='feature', x='total cover', orient='h', palette='Purples_r')\n",
    "plt.ylabel('');\n",
    "plt.xlabel('Total Cover', fontsize=12)\n",
    "plt.yticks(fontsize=12)\n",
    "plt.xticks(fontsize=12)\n",
    "plt.title('Total Cover', fontsize=18);"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.5"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": false,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {
    "height": "calc(100% - 180px)",
    "left": "10px",
    "top": "150px",
    "width": "384px"
   },
   "toc_section_display": true,
   "toc_window_display": true
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}