Boston_House-Prices_Analysis-Databreed_Exercise.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# The Data Mining Process (Based on CRISPDM)\n",
    "* ** Business Objectives **\n",
    "* ** Data Understanding (EDA)**\n",
    "* ** Data Preparation **\n",
    "* ** Modeling **\n",
    "* ** Model Evaluation & Selection **\n",
    "* ** Deployment **"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Import the relevant python packages and load the dataset from standard datasets provided by scikit-learn package."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Boston House Prices dataset\n",
      "===========================\n",
      "\n",
      "Notes\n",
      "------\n",
      "Data Set Characteristics:  \n",
      "\n",
      "    :Number of Instances: 506 \n",
      "\n",
      "    :Number of Attributes: 13 numeric/categorical predictive\n",
      "    \n",
      "    :Median Value (attribute 14) is usually the target\n",
      "\n",
      "    :Attribute Information (in order):\n",
      "        - CRIM     per capita crime rate by town\n",
      "        - ZN       proportion of residential land zoned for lots over 25,000 sq.ft.\n",
      "        - INDUS    proportion of non-retail business acres per town\n",
      "        - CHAS     Charles River dummy variable (= 1 if tract bounds river; 0 otherwise)\n",
      "        - NOX      nitric oxides concentration (parts per 10 million)\n",
      "        - RM       average number of rooms per dwelling\n",
      "        - AGE      proportion of owner-occupied units built prior to 1940\n",
      "        - DIS      weighted distances to five Boston employment centres\n",
      "        - RAD      index of accessibility to radial highways\n",
      "        - TAX      full-value property-tax rate per $10,000\n",
      "        - PTRATIO  pupil-teacher ratio by town\n",
      "        - B        1000(Bk - 0.63)^2 where Bk is the proportion of blacks by town\n",
      "        - LSTAT    % lower status of the population\n",
      "        - MEDV     Median value of owner-occupied homes in $1000's\n",
      "\n",
      "    :Missing Attribute Values: None\n",
      "\n",
      "    :Creator: Harrison, D. and Rubinfeld, D.L.\n",
      "\n",
      "This is a copy of UCI ML housing dataset.\n",
      "http://archive.ics.uci.edu/ml/datasets/Housing\n",
      "\n",
      "\n",
      "This dataset was taken from the StatLib library which is maintained at Carnegie Mellon University.\n",
      "\n",
      "The Boston house-price data of Harrison, D. and Rubinfeld, D.L. 'Hedonic\n",
      "prices and the demand for clean air', J. Environ. Economics & Management,\n",
      "vol.5, 81-102, 1978.   Used in Belsley, Kuh & Welsch, 'Regression diagnostics\n",
      "...', Wiley, 1980.   N.B. Various transformations are used in the table on\n",
      "pages 244-261 of the latter.\n",
      "\n",
      "The Boston house-price data has been used in many machine learning papers that address regression\n",
      "problems.   \n",
      "     \n",
      "**References**\n",
      "\n",
      "   - Belsley, Kuh & Welsch, 'Regression diagnostics: Identifying Influential Data and Sources of Collinearity', Wiley, 1980. 244-261.\n",
      "   - Quinlan,R. (1993). Combining Instance-Based and Model-Based Learning. In Proceedings on the Tenth International Conference of Machine Learning, 236-243, University of Massachusetts, Amherst. Morgan Kaufmann.\n",
      "   - many more! (see http://archive.ics.uci.edu/ml/datasets/Housing)\n",
      "\n"
     ]
    }
   ],
   "source": [
    "import sklearn.datasets as dt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import sklearn as sk\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import statsmodels.api as sm\n",
    "import pylab\n",
    "import pandas.tseries\n",
    "import warnings\n",
    "\n",
    "\n",
    "%matplotlib inline\n",
    "\n",
    "warnings.simplefilter(action='ignore', category=FutureWarning)\n",
    "\n",
    "boston = dt.load_boston()\n",
    "\n",
    "# for key in boston.keys(): print(\"Key:\", key, \">>  Data Type:\", type(boston[key]))\n",
    "# print(type(boston))\n",
    "\n",
    "print(boston.DESCR)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data Understanding (Exploratory Data Analysis)\n",
    "* Review the dataset to understand the structure and format of the data available for modelling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(506, 15)\n",
      "=================================================================\n",
      "      CRIM    ZN  INDUS  CHAS    NOX     RM   AGE     DIS  RAD    TAX  \\\n",
      "0  0.00632  18.0   2.31   0.0  0.538  6.575  65.2  4.0900  1.0  296.0   \n",
      "1  0.02731   0.0   7.07   0.0  0.469  6.421  78.9  4.9671  2.0  242.0   \n",
      "2  0.02729   0.0   7.07   0.0  0.469  7.185  61.1  4.9671  2.0  242.0   \n",
      "3  0.03237   0.0   2.18   0.0  0.458  6.998  45.8  6.0622  3.0  222.0   \n",
      "4  0.06905   0.0   2.18   0.0  0.458  7.147  54.2  6.0622  3.0  222.0   \n",
      "\n",
      "   PTRATIO       B  LSTAT        Bk  MEDV  \n",
      "0     15.3  396.90   4.98  1.260000  24.0  \n",
      "1     17.8  396.90   9.14  1.260000  21.6  \n",
      "2     17.8  392.83   4.03  1.256762  34.7  \n",
      "3     18.7  394.63   2.94  1.258196  33.4  \n",
      "4     18.7  396.90   5.33  1.260000  36.2  \n"
     ]
    }
   ],
   "source": [
    "# Load the data from scikit-learn dataset Bunch object into a pandas DataFrame \n",
    "# print(type(boston))\n",
    "df_boston = pd.DataFrame(boston.data, index=np.arange(506), columns=boston.feature_names)  \n",
    "df_boston[\"Bk\"] = np.sqrt(df_boston[\"B\"] / 1000) + 0.63\n",
    "df_boston[\"MEDV\"] = pd.Series(boston.target, index=np.arange(506))\n",
    "\n",
    "sns.set_style(\"darkgrid\")\n",
    "\n",
    "print(df_boston.shape)\n",
    "print(\"=================================================================\")\n",
    "print(df_boston.head())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* To enhance understanding of the data, graphs and plots provide a visual aid."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x720 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def plotVariable(x, y, df):\n",
    "    plt.subplot(221)\n",
    "    p1 = sns.regplot(df[x], df[y])\n",
    "    plt.xlabel(x, fontsize=12, fontweight='bold')\n",
    "    plt.ylabel(\"MEDV (Target)\", fontsize=12, fontweight='bold')\n",
    "    p1.text(0.95, 0.95, 'Pearson Corr: ' + str(np.corrcoef(df[x], df[y])[0, 1]), verticalalignment='top', horizontalalignment='right',\n",
    "        transform=p1.transAxes, color='red', fontsize=12)\n",
    "\n",
    "    plt.subplot(222)\n",
    "    p2 = sns.regplot(np.log10(df[x]), df[y])\n",
    "    plt.xlabel(\"Log10(\" + x + \")\", fontsize=12, fontweight='bold')\n",
    "    plt.ylabel(\"MEDV (Target)\", fontsize=12, fontweight='bold')\n",
    "    p2.text(0.95, 0.95, 'Pearson Corr: ' + str(np.corrcoef(np.log10(df[x]), df[y])[0, 1]), verticalalignment='top', horizontalalignment='right',\n",
    "        transform=p2.transAxes, color='red', fontsize=12)\n",
    "\n",
    "    plt.subplot(223)\n",
    "    p3 = sns.regplot(np.reciprocal(df[x]), df[y])\n",
    "    plt.xlabel(\"1 / \" + x, fontsize=12, fontweight='bold')\n",
    "    plt.ylabel(\"MEDV (Target)\", fontsize=12, fontweight='bold')\n",
    "    p3.text(0.95, 0.95, 'Pearson Corr: ' + str(np.corrcoef(np.reciprocal(df[x]), df[y])[0, 1]), verticalalignment='top', horizontalalignment='right',\n",
    "        transform=p3.transAxes, color='red', fontsize=12)\n",
    "\n",
    "    plt.subplot(224)\n",
    "    p4 = sns.regplot(np.power(df[x], 2), df[y])\n",
    "    plt.xlabel(x + \" ^ 2\", fontsize=12, fontweight='bold')\n",
    "    plt.ylabel(\"MEDV (Target)\", fontsize=12, fontweight='bold')\n",
    "    p4.text(0.95, 0.95, 'Pearson Corr: ' + str(np.corrcoef(np.power(df[x], 2), df[y])[0, 1]), verticalalignment='top', horizontalalignment='right',\n",
    "        transform=p4.transAxes, color='red', fontsize=12)\n",
    "\n",
    "    fig = plt.gcf()\n",
    "    fig.set_size_inches(14, 10)\n",
    "    #fig.set_tight_layout(tight=True)\n",
    "    fig.suptitle(\"Scatter Plots & Correlation Co-eff for Variable MEDV/\" + x, fontsize=22, fontweight='bold')\n",
    "\n",
    "    plt.show()\n",
    "    \n",
    "    return 0\n",
    "\n",
    "# Execute this call for the input variables in dataset. plotVariable(input, target, dataframe)\n",
    "plotVariable(\"LSTAT\", \"MEDV\", df_boston) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Helper function to summarize important characteristics of variables in a dataset\n",
    "def summarize(data): \n",
    "    '''\n",
    "        Method to summarize important characteristics of the continuous variables in a DataFrame\n",
    "        \n",
    "        data: DataFrame containing the variables to be summarized\n",
    "    '''\n",
    "    col_count = 0\n",
    "    cols_df = pd.DataFrame({\"A\": 0}, index=np.array([0]))\n",
    "    for cols in data.columns:\n",
    "        col_type = str(data[cols].dtype)\n",
    "        if col_count == 0:  \n",
    "            cols_df = pd.DataFrame({\"VAR_NAME\": cols, \"TYPE\": data[cols].dtype, \"CORR_TARGET\": np.corrcoef(data[cols], data[\"MEDV\"])[0][1], \"MEAN\": np.mean(data[cols]), \"SKEW\": data[cols].skew(), \"KURT\": data[cols].kurt(), \"MIN\": np.min(data[cols]), \"MAX\": np.max(data[cols]), \"STDEV\": np.std(data[cols]), \"COUNT\":  len(data[cols]), \"MISSING\": np.sum(pd.isnull(data[cols])), \"TMIN\": np.mean(data[cols]) - (3 * np.std(data[cols])), \"TMAX\": np.mean(data[cols]) + (3 * np.std(data[cols]))}, index=np.array([col_count]))\n",
    "        else:\n",
    "            cols_df = cols_df.append(pd.DataFrame({\"VAR_NAME\": cols, \"TYPE\": data[cols].dtype, \"CORR_TARGET\": np.corrcoef(data[cols], data[\"MEDV\"])[0][1], \"MEAN\": np.mean(data[cols]), \"SKEW\": data[cols].skew(), \"KURT\": data[cols].kurt(), \"MIN\": np.min(data[cols]), \"MAX\": np.max(data[cols]), \"STDEV\": np.std(data[cols]), \"COUNT\":  len(data[cols]), \"MISSING\": np.sum(pd.isnull(data[cols])), \"TMIN\": np.mean(data[cols]) - (3 * np.std(data[cols])), \"TMAX\": np.mean(data[cols]) + (3 * np.std(data[cols]))}, index=np.array([col_count])))\n",
    "        col_count += 1\n",
    "    cols_df = cols_df[[\"VAR_NAME\", \"TYPE\", \"COUNT\", \"MISSING\", \"CORR_TARGET\", \"MEAN\", \"STDEV\", \"MIN\", \"MAX\", \"SKEW\", \"KURT\", \"TMIN\", \"TMAX\"]]\n",
    "    return cols_df\n",
    "\n",
    "def summarize_cat(data):\n",
    "    col_count = 0\n",
    "    cols_df = pd.DataFrame({\"A\": 0}, index=np.array([0]))\n",
    "    for cols in data.columns:\n",
    "        col_type = str(data[cols].dtype)\n",
    "        if col_count == 0: \n",
    "            continue\n",
    "    return cols_df        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "def r_squared_adj(r_squared, n, k):\n",
    "    '''\n",
    "        Method that computes the Adjusted R-Squared value of a regression model based on the R-Squared value.\n",
    "        \n",
    "        r_squared: The coefficient of determination \n",
    "        n: The total number of records in the sample\n",
    "        k: The total number of predictors in the sample\n",
    "    '''\n",
    "    val = 1 - ((1 - r_squared) * ((n - 1) / (n - k - 1)))\n",
    "    return val"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data Preparation\n",
    "* The tasks to be executed will be determined exploratory data analysis\n",
    "* The actual transformation tasks are executed below:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Functions for cleansing and transformations\n",
    "df_boston_regr = (df_boston\n",
    "                        .assign(CRIM=lambda x: np.log10(x['CRIM']),\n",
    "                                INDUS=lambda x: np.log10(x['INDUS']),\n",
    "                                LSTAT=lambda x: np.log10(x['LSTAT']),\n",
    "                                DIS=lambda x: np.reciprocal(x['DIS']),\n",
    "                                CHAS=lambda x: pd.Categorical(x['CHAS'].astype(int)))\n",
    "                 )\n",
    "\n",
    "# print(pd.crosstab(df_boston_regr.CHAS, columns=\"Count\"))\n",
    "# print(df_boston_regr.describe())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Modeling\n",
    "* First we build baseline model that is based on data in it's untransformed state. The resulting model will serve as a  baseline for comparison with other models.\n",
    "* Next we build a model using the sample machine learing algorithm using the transformed data. Compare this model with the baseline to see the effect of transformations."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Linear Regression Models\n",
    "* Build the baseline Linear Regression model. Use the dataset without the transformations (df_boston)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "#----------------------------- Results of the Recursive Feature Elimination Process------------------------#\n",
      "Optimal number of features: 6\n",
      "List of optimal features:  Index(['CHAS', 'NOX', 'RM', 'DIS', 'PTRATIO', 'LSTAT'], dtype='object')\n",
      "Assigned score of optimal features:  [-0.31637681 -0.08962896  0.17159141  0.29433175  0.39615242  0.50135307\n",
      "  0.43547489  0.36364582  0.3529999   0.36032063  0.38418399  0.38774601\n",
      "  0.39903343]\n",
      "\n",
      "#----------------------- Results of Model A - Developed with 6 Top Features from RFE ----------------------#\n",
      "Regression coefficients of Model A:  [[ 0.09705016 -0.24698783  0.33971453 -0.27245457 -0.22951849 -0.42696662]]\n",
      "Intercept coefficient of Model A:  [4.31040347e-16]\n",
      "Rank of Model A:  6\n",
      "R^2 Model A:  0.7257314647320705\n",
      "R^2 Adj. Model A:  0.7224336466727367\n",
      "\n",
      "#--------------------- Results of Model B - Developed with all 13 Features in Dataset ---------------------#\n",
      "Regression coefficients of Model B:  [[-0.10688508  0.07462915  0.03009218  0.07717747 -0.21745262  0.33696814\n",
      "  -0.01871893 -0.33079736  0.24126113 -0.18944484 -0.21885409  0.1221749\n",
      "  -0.388085  ]]\n",
      "Intercept coefficient of Model B:  [4.89061054e-16]\n",
      "Rank of Model B:  13\n",
      "R^2 Model B:  0.7508837786732914\n",
      "R^2 Adj. Model B:  0.7443014394918945\n",
      "\n",
      "#------------------ Results of Selected Model B on Previously Unseen Data (Test Dataset) ------------------#\n",
      "R^2 Model B (Validation):  0.6684825753971603\n",
      "R^2 Adj. Model B (Validation):  0.6597229686495243\n"
     ]
    }
   ],
   "source": [
    "from sklearn.model_selection import cross_val_score, KFold\n",
    "from sklearn.linear_model import LinearRegression\n",
    "from sklearn.feature_selection import RFECV\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import mean_squared_error, r2_score\n",
    "\n",
    "# split into input (X) and target (y) variables\n",
    "X = df_boston[[\"CRIM\", \"ZN\", \"INDUS\", \"CHAS\", \"NOX\", \"RM\", \"AGE\", \"DIS\", \"RAD\", \"TAX\", \"PTRATIO\", \"B\", \"LSTAT\" ]]\n",
    "y = df_boston[\"MEDV\"].values\n",
    "\n",
    "# Split data into train & test samples train 80% while test 20%\n",
    "X_train, X_valid, y_train, y_valid = train_test_split(X, y, test_size=0.20, random_state=42)\n",
    "\n",
    "# Scale the training data to reduce the effect of larger valued variables on smaller ones\n",
    "scalerX = StandardScaler().fit(X_train)\n",
    "Stan_X = pd.DataFrame(scalerX.transform(X_train), index=X_train.index, columns=X_train.columns)\n",
    "scalerY = StandardScaler().fit(y_train.reshape(y_train.shape[0], 1))\n",
    "Stan_y = scalerY.transform(y_train.reshape(y_train.shape[0], 1))\n",
    "\n",
    "# Create regression estimator and set hyper-parameters\n",
    "lin_reg = LinearRegression(fit_intercept = True, normalize = False)\n",
    "\n",
    "# Create the rfe object to recursively select best features by leveraging cross-validation \n",
    "kfold = KFold(n_splits=6, random_state=42, shuffle=False)\n",
    "rfecv = RFECV(estimator=lin_reg, step=1, cv=kfold)\n",
    "rfecv.fit(X, y)\n",
    "\n",
    "# Review the optimal features selected by the recursive elimination process\n",
    "print(\"\\n#----------------------------- Results of the Recursive Feature Elimination Process------------------------#\")\n",
    "print(\"Optimal number of features: %d\" % rfecv.n_features_)\n",
    "print(\"List of optimal features: \", X.columns[rfecv.support_])\n",
    "print(\"Assigned score of optimal features: \", rfecv.grid_scores_)\n",
    "\n",
    "# Select columns based on recursive feature extraction to be use for linear regression model\n",
    "cols = X.columns[rfecv.support_]\n",
    "\n",
    "# Build a model with the optimal columns selected earlier\n",
    "reg_model_a = lin_reg.fit(Stan_X[cols], Stan_y)\n",
    "print(\"\\n#----------------------- Results of Model A - Developed with 6 Top Features from RFE ----------------------#\")\n",
    "print(\"Regression coefficients of Model A: \", reg_model_a.coef_)\n",
    "print(\"Intercept coefficient of Model A: \", reg_model_a.intercept_)\n",
    "print(\"Rank of Model A: \", reg_model_a.rank_)\n",
    "print(\"R^2 Model A: \", reg_model_a.score(Stan_X[cols], Stan_y))\n",
    "print(\"R^2 Adj. Model A: \", r_squared_adj(reg_model_a.score(Stan_X[cols], Stan_y), len(y), len(cols)))\n",
    "\n",
    "# Build a model with the optimal columns selected earlier\n",
    "reg_model_b = lin_reg.fit(Stan_X, Stan_y)\n",
    "print(\"\\n#--------------------- Results of Model B - Developed with all 13 Features in Dataset ---------------------#\")\n",
    "print(\"Regression coefficients of Model B: \", reg_model_b.coef_)\n",
    "print(\"Intercept coefficient of Model B: \", reg_model_b.intercept_)\n",
    "print(\"Rank of Model B: \", reg_model_b.rank_)\n",
    "print(\"R^2 Model B: \", reg_model_b.score(Stan_X, Stan_y))\n",
    "print(\"R^2 Adj. Model B: \", r_squared_adj(reg_model_b.score(Stan_X, Stan_y), len(y), len(X.columns)))\n",
    "\n",
    "# Scale the hold-out sample to be used for testing using the same scaler used on the training data\n",
    "Stan_X_valid = pd.DataFrame(scalerX.transform(X_valid), index=X_valid.index, columns=X_valid.columns)\n",
    "Stan_y_valid = scalerY.transform(y_valid.reshape(y_valid.shape[0], 1))\n",
    "\n",
    "print(\"\\n#------------------ Results of Selected Model B on Previously Unseen Data (Test Dataset) ------------------#\")\n",
    "print(\"R^2 Model B (Validation): \", reg_model_b.score(Stan_X_valid, Stan_y_valid))\n",
    "print(\"R^2 Adj. Model B (Validation): \", r_squared_adj(reg_model_b.score(Stan_X_valid, Stan_y_valid), len(y), len(X.columns)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* Build the first iteration Linear Regression model. Use the dataset that has been transformed (df_boston_regr)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "#----------------------------- Results of the Recursive Feature Elimination Process------------------------#\n",
      "Optimal number of features: 7\n",
      "List of optimal features:  Index(['INDUS', 'CHAS', 'NOX', 'RM', 'DIS', 'PTRATIO', 'LSTAT'], dtype='object')\n",
      "Assigned score of optimal features:  [0.48485066 0.46812572 0.48224873 0.41364144 0.47495574 0.54790255\n",
      " 0.56882402 0.56086039 0.5165455  0.51202613 0.48903929 0.49873405\n",
      " 0.51666895]\n",
      "\n",
      "#----------------------- Results of Model A - Developed with 7 Top Features from RFE ----------------------#\n",
      "Regression coefficients of Model A:  [[-0.03662214  0.0840218  -0.23995846  0.27402621  0.33364548 -0.17383348\n",
      "  -0.56723011]]\n",
      "Intercept coefficient of Model A:  [5.39061964e-16]\n",
      "Rank of Model A:  7\n",
      "R^2 Model A:  0.7787119791619235\n",
      "R^2 Adj. Model A:  0.7756015049734365\n",
      "\n",
      "#--------------------- Results of Model B - Developed with all 13 Features in Dataset ---------------------#\n",
      "Regression coefficients of Model B:  [[-0.01729677 -0.07573965 -0.04011547  0.07003037 -0.21925309  0.2718801\n",
      "   0.01764014  0.3321882   0.16299153 -0.18191426 -0.18843121  0.09881807\n",
      "  -0.55629004]]\n",
      "Intercept coefficient of Model B:  [6.07870478e-16]\n",
      "Rank of Model B:  13\n",
      "R^2 Model B:  0.7985451335378196\n",
      "R^2 Adj. Model B:  0.7932221390987784\n",
      "\n",
      "#------------------ Results of Selected Model B on Previously Unseen Data (Test Dataset) ------------------#\n",
      "R^2 Model B (Validation):  0.7470127649229485\n",
      "R^2 Adj. Model B (Validation):  0.7403281428579045\n"
     ]
    }
   ],
   "source": [
    "from sklearn.model_selection import cross_val_score, KFold\n",
    "from sklearn.linear_model import LinearRegression\n",
    "from sklearn.feature_selection import RFECV\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import mean_squared_error, r2_score\n",
    "\n",
    "# split into input (X) and target (y) variables\n",
    "X = df_boston_regr[[\"CRIM\", \"ZN\", \"INDUS\", \"CHAS\", \"NOX\", \"RM\", \"AGE\", \"DIS\", \"RAD\", \"TAX\", \"PTRATIO\", \"B\", \"LSTAT\" ]]\n",
    "y = df_boston_regr[\"MEDV\"].values\n",
    "\n",
    "# Split data into train & test samples\n",
    "X_train, X_valid, y_train, y_valid = train_test_split(X, y, test_size=0.20, random_state=42)\n",
    "\n",
    "# Scale the training data to reduce the effect of larger valued variables on smaller ones\n",
    "scalerX = StandardScaler().fit(X_train)\n",
    "Stan_X = pd.DataFrame(scalerX.transform(X_train), index=X_train.index, columns=X_train.columns)\n",
    "scalerY = StandardScaler().fit(y_train.reshape(y_train.shape[0], 1))\n",
    "Stan_y = scalerY.transform(y_train.reshape(y_train.shape[0], 1))\n",
    "\n",
    "# Create regression estimator and set hyper-parameters\n",
    "lin_reg = LinearRegression(fit_intercept = True, normalize = False)\n",
    "\n",
    "# Create the rfe object to recursively select best features by leveraging cross-validation \n",
    "kfold = KFold(n_splits=6, random_state=42, shuffle=False)\n",
    "rfecv = RFECV(estimator=lin_reg, step=1, cv=kfold)\n",
    "rfecv.fit(X, y)\n",
    "\n",
    "# Review the optimal features selected by the recursive elimination process\n",
    "print(\"\\n#----------------------------- Results of the Recursive Feature Elimination Process------------------------#\")\n",
    "print(\"Optimal number of features: %d\" % rfecv.n_features_)\n",
    "print(\"List of optimal features: \", X.columns[rfecv.support_])\n",
    "print(\"Assigned score of optimal features: \", rfecv.grid_scores_)\n",
    "\n",
    "# Select columns based on recursive feature extraction to be use for linear regression model\n",
    "cols = X.columns[rfecv.support_]\n",
    "\n",
    "# Build a model with the optimal columns selected earlier\n",
    "reg_model_a = lin_reg.fit(Stan_X[cols], Stan_y)\n",
    "print(\"\\n#----------------------- Results of Model A - Developed with 7 Top Features from RFE ----------------------#\")\n",
    "print(\"Regression coefficients of Model A: \", reg_model_a.coef_)\n",
    "print(\"Intercept coefficient of Model A: \", reg_model_a.intercept_)\n",
    "print(\"Rank of Model A: \", reg_model_a.rank_)\n",
    "print(\"R^2 Model A: \", reg_model_a.score(Stan_X[cols], Stan_y))\n",
    "print(\"R^2 Adj. Model A: \", r_squared_adj(reg_model_a.score(Stan_X[cols], Stan_y), len(y), len(cols)))\n",
    "\n",
    "# Build a model with the optimal columns selected earlier\n",
    "reg_model_b = lin_reg.fit(Stan_X, Stan_y)\n",
    "print(\"\\n#--------------------- Results of Model B - Developed with all 13 Features in Dataset ---------------------#\")\n",
    "print(\"Regression coefficients of Model B: \", reg_model_b.coef_)\n",
    "print(\"Intercept coefficient of Model B: \", reg_model_b.intercept_)\n",
    "print(\"Rank of Model B: \", reg_model_b.rank_)\n",
    "print(\"R^2 Model B: \", reg_model_b.score(Stan_X, Stan_y))\n",
    "print(\"R^2 Adj. Model B: \", r_squared_adj(reg_model_b.score(Stan_X, Stan_y), len(y), len(X.columns)))\n",
    "\n",
    "# Scale the hold-out sample to be used for testing using the same scaler used on the training data\n",
    "Stan_X_valid = pd.DataFrame(scalerX.transform(X_valid), index=X_valid.index, columns=X_valid.columns)\n",
    "Stan_y_valid = scalerY.transform(y_valid.reshape(y_valid.shape[0], 1))\n",
    "\n",
    "#print(\"\\n#------------------ Results of Selected Model A on Previously Unseen Data (Test Dataset) ------------------#\")\n",
    "#print(\"R^2 Model A (Validation): \", reg_model_a.score(Stan_X_valid[cols], Stan_y_valid))\n",
    "#print(\"R^2 Adj. Model A (Validation): \", r_squared_adj(reg_model_a.score(Stan_X_valid[cols], Stan_y_valid), len(y), len(cols)))\n",
    "\n",
    "print(\"\\n#------------------ Results of Selected Model B on Previously Unseen Data (Test Dataset) ------------------#\")\n",
    "print(\"R^2 Model B (Validation): \", reg_model_b.score(Stan_X_valid, Stan_y_valid))\n",
    "print(\"R^2 Adj. Model B (Validation): \", r_squared_adj(reg_model_b.score(Stan_X_valid, Stan_y_valid), len(y), len(X.columns)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Decision Tree Regression Models\n",
    "* Apply the DecisionTreeRegressor object in the scikit-learn package to build models using the dataset above.\n",
    "* Build two models like we did for Linear Regression above. Baseline & First Iteration models."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Modelling\n",
    "Baseline with Tree Regression Model\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "#-------------------- Results of DecisionTreeRegressor model - Developed with all 13 Features in Dataset -------------------#\n",
      "\n",
      "Tuned DecisionTreeRegressor R squared on training data: 0.9187021181746381\n",
      "R^2 Adj. Model on training data :  0.9165540034109598\n",
      "Parameters of training data :  Pipeline(memory=None,\n",
      "     steps=[('scaler', StandardScaler(copy=True, with_mean=True, with_std=True)), ('DecisionTreeRegressor', DecisionTreeRegressor(criterion='mse', max_depth=6, max_features=None,\n",
      "           max_leaf_nodes=None, min_impurity_decrease=0.0,\n",
      "           min_impurity_split=None, min_samples_leaf=3,\n",
      "           min_samples_split=2, min_weight_fraction_leaf=0.0,\n",
      "           presort=False, random_state=None, splitter='best'))])\n",
      "********************************************************************************\n",
      "Tuned DecisionTreeRegressor R squared on testing data: 0.865524160124324\n",
      "R^2 Adj. Model on testing data :  0.8619709367129749\n",
      "\n",
      "\n",
      "#----------------------------- Results of the Recursive Feature Elimination Process------------------------#\n",
      "\n",
      "Optimal number of features: 11\n",
      "List of optimal features:  Index(['CRIM', 'INDUS', 'NOX', 'RM', 'AGE', 'DIS', 'RAD', 'TAX', 'PTRATIO',\n",
      "       'LSTAT', 'B'],\n",
      "      dtype='object')\n",
      "Assigned score of optimal features:  [0.31347209 0.65642206 0.65250247 0.67725294 0.67322175 0.65428953\n",
      " 0.64485207 0.66134198 0.68575189 0.69676804 0.71708292 0.69483677\n",
      " 0.70428628]\n",
      "\n",
      "\n",
      "#--------------------- Results of DecisionTreeRegressor Model - Developed with 10 Top Features from RFE --------------------#\n",
      "\n",
      "DecisionTreeRegressor R squared on training data_rfecv: 0.9772162012795118\n",
      "R^2 Adj. Model on training data_rfecv :  0.9766141903377102\n",
      "********************************************************************************\n",
      "DecisionTreeRegressor R squared on testing data_rfecv: 0.706356262923258\n",
      "R^2 Adj. Model on testing data_rfecv :  0.6985973836915556\n"
     ]
    }
   ],
   "source": [
    "from sklearn.model_selection import cross_val_score, KFold\n",
    "from sklearn.tree import DecisionTreeRegressor\n",
    "from sklearn.feature_selection import RFECV\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import mean_squared_error, r2_score\n",
    "from sklearn.pipeline import Pipeline\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "\n",
    "X=df_boston[[\"CRIM\", \"ZN\", \"INDUS\", \"CHAS\", \"NOX\", \"RM\",\"AGE\", \"DIS\",\"RAD\", \"TAX\", \"PTRATIO\", \"LSTAT\",\"B\" ]] #to get better result remove \"AGE\"\n",
    "y = df_boston[\"MEDV\"].values\n",
    "\n",
    "steps = [('scaler',StandardScaler()),\n",
    "         ('DecisionTreeRegressor',DecisionTreeRegressor())]\n",
    "\n",
    "\n",
    "pipeline=Pipeline(steps)\n",
    "\n",
    "parameters = {'DecisionTreeRegressor__max_depth':np.arange(1,9),\n",
    "             'DecisionTreeRegressor__min_samples_split':np.arange(2,10),\n",
    "             'DecisionTreeRegressor__min_samples_leaf':np.arange(1,5)}\n",
    "\n",
    "X_train, X_test, y_train, y_test =train_test_split(X,y,test_size=0.2,random_state=42)\n",
    "gm_cv = GridSearchCV(pipeline,parameters,cv=4)\n",
    "\n",
    "gm_cv.fit(X_train,y_train)\n",
    "######################################################################################\n",
    "#pipeline.fit(X_train,y_train)\n",
    "#r2 =pipeline.score(X_test, y_test)\n",
    "r2=gm_cv.score(X_train,y_train)\n",
    "r2_t=gm_cv.score(X_test, y_test)\n",
    "###################################################################################\n",
    "DecisionTreeRegressor=DecisionTreeRegressor(max_depth=8)\n",
    "kfold = KFold(n_splits=6, random_state=42, shuffle=False)\n",
    "rfecv = RFECV(estimator=DecisionTreeRegressor, step=1, cv=kfold)\n",
    "rfecv.fit(X_train, y_train)\n",
    "cols = X.columns[rfecv.support_]\n",
    "reg_model_a =DecisionTreeRegressor.fit(X_train[cols], y_train)\n",
    "\n",
    "#######################################################################################\n",
    "r2_rfecv=reg_model_a.score(X_train[cols],y_train)\n",
    "r2_rfecv_t=reg_model_a.score(X_test[cols], y_test)\n",
    "\n",
    "print(\"\\n#-------------------- Results of DecisionTreeRegressor model - Developed with all 13 Features in Dataset -------------------#\")\n",
    "print(\"\\nTuned DecisionTreeRegressor R squared on training data: {}\".format(r2))\n",
    "print(\"R^2 Adj. Model on training data : \", r_squared_adj(r2 , len(y), 13))\n",
    "print(\"Parameters of training data : \", gm_cv.best_estimator_)\n",
    "print(\"********************************************************************************\")\n",
    "print(\"Tuned DecisionTreeRegressor R squared on testing data: {}\".format(r2_t))\n",
    "print(\"R^2 Adj. Model on testing data : \", r_squared_adj(r2_t , len(y), 13))\n",
    "\n",
    "print(\"\\n\\n#----------------------------- Results of the Recursive Feature Elimination Process------------------------#\")\n",
    "print(\"\\nOptimal number of features: %d\" % rfecv.n_features_)\n",
    "print(\"List of optimal features: \", X.columns[rfecv.support_])\n",
    "print(\"Assigned score of optimal features: \", rfecv.grid_scores_)\n",
    "print(\"\\n\\n#--------------------- Results of DecisionTreeRegressor Model - Developed with 10 Top Features from RFE --------------------#\")\n",
    "print(\"\\nDecisionTreeRegressor R squared on training data_rfecv: {}\".format(r2_rfecv))\n",
    "print(\"R^2 Adj. Model on training data_rfecv : \", r_squared_adj(r2_rfecv , len(y), 13))\n",
    "print(\"********************************************************************************\")\n",
    "print(\"DecisionTreeRegressor R squared on testing data_rfecv: {}\".format(r2_rfecv_t))\n",
    "print(\"R^2 Adj. Model on testing data_rfecv : \", r_squared_adj(r2_rfecv_t , len(y), 13))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# First Iteration using transformed data df_boston_regr"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "#----------------------------- Results of the Recursive Feature Elimination Process------------------------#\n",
      "Optimal number of features: 4\n",
      "List of optimal features:  Index(['NOX', 'RM', 'DIS', 'LSTAT'], dtype='object')\n",
      "Assigned score of optimal features:  [-0.24958255  0.1206538   0.36272535  0.49831325  0.31763043  0.34887941\n",
      "  0.29933498  0.14623331  0.15240095  0.25243192  0.28938216  0.21718608\n",
      "  0.21256294]\n",
      "\n",
      "#----------------------- Results of Model A - Developed with 7 Top Features from RFE ----------------------#\n",
      "R^2 Model A:  0.9700857635690702\n",
      "R^2 Adj. Model A:  0.9698469273500607\n",
      "\n",
      "#------------------ Results of Selected Model B on Previously Unseen Data (Test Dataset) ------------------#\n",
      "R^2 Model B (Validation):  0.7470127649229485\n",
      "R^2 Adj. Model B (Validation):  0.7403281428579045\n"
     ]
    }
   ],
   "source": [
    "from sklearn.model_selection import cross_val_score, KFold\n",
    "from sklearn.tree import DecisionTreeRegressor\n",
    "from sklearn.feature_selection import RFECV\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import mean_squared_error, r2_score\n",
    "\n",
    "# split into input (X) and target (y) variables\n",
    "X = df_boston_regr[[\"CRIM\", \"ZN\", \"INDUS\", \"CHAS\", \"NOX\", \"RM\", \"AGE\", \"DIS\", \"RAD\", \"TAX\", \"PTRATIO\", \"B\", \"LSTAT\" ]]\n",
    "y = df_boston_regr[\"MEDV\"].values\n",
    "\n",
    "# Split data into train & test samples\n",
    "X_train, X_valid, y_train, y_valid = train_test_split(X, y, test_size=0.20, random_state=42)\n",
    "\n",
    "# Scale the training data to reduce the effect of larger valued variables on smaller ones\n",
    "scalerX = StandardScaler().fit(X_train)\n",
    "Stan_X = pd.DataFrame(scalerX.transform(X_train), index=X_train.index, columns=X_train.columns)\n",
    "scalerY = StandardScaler().fit(y_train.reshape(y_train.shape[0], 1))\n",
    "Stan_y = scalerY.transform(y_train.reshape(y_train.shape[0], 1))\n",
    "\n",
    "# Create regression estimator and set hyper-parameters\n",
    "DecisionTreeRegressor = DecisionTreeRegressor(max_depth=8)\n",
    "\n",
    "# Create the rfe object to recursively select best features by leveraging cross-validation \n",
    "kfold = KFold(n_splits=6, random_state=42, shuffle=False)\n",
    "rfecv = RFECV(estimator=DecisionTreeRegressor, step=1, cv=kfold)\n",
    "rfecv.fit(X, y)\n",
    "\n",
    "# Review the optimal features selected by the recursive elimination process\n",
    "print(\"\\n#----------------------------- Results of the Recursive Feature Elimination Process------------------------#\")\n",
    "print(\"Optimal number of features: %d\" % rfecv.n_features_)\n",
    "print(\"List of optimal features: \", X.columns[rfecv.support_])\n",
    "print(\"Assigned score of optimal features: \", rfecv.grid_scores_)\n",
    "\n",
    "# Select columns based on recursive feature extraction to be use for Decision Tree regression model\n",
    "cols = X.columns[rfecv.support_]\n",
    "\n",
    "# Build a model with the optimal columns selected earlier\n",
    "reg_model_a = DecisionTreeRegressor.fit(Stan_X[cols], Stan_y)\n",
    "print(\"\\n#----------------------- Results of Model A - Developed with 7 Top Features from RFE ----------------------#\")\n",
    "print(\"R^2 Model A: \", reg_model_a.score(Stan_X[cols], Stan_y))\n",
    "print(\"R^2 Adj. Model A: \", r_squared_adj(reg_model_a.score(Stan_X[cols], Stan_y), len(y), len(cols)))\n",
    "\n",
    "\n",
    "# Scale the hold-out sample to be used for testing using the same scaler used on the training data\n",
    "Stan_X_valid = pd.DataFrame(scalerX.transform(X_valid), index=X_valid.index, columns=X_valid.columns)\n",
    "Stan_y_valid = scalerY.transform(y_valid.reshape(y_valid.shape[0], 1))\n",
    "\n",
    "#print(\"\\n#------------------ Results of Selected Model A on Previously Unseen Data (Test Dataset) ------------------#\")\n",
    "#print(\"R^2 Model A (Validation): \", reg_model_a.score(Stan_X_valid[cols], Stan_y_valid))\n",
    "#print(\"R^2 Adj. Model A (Validation): \", r_squared_adj(reg_model_a.score(Stan_X_valid[cols], Stan_y_valid), len(y), len(cols)))\n",
    "\n",
    "print(\"\\n#------------------ Results of Selected Model B on Previously Unseen Data (Test Dataset) ------------------#\")\n",
    "print(\"R^2 Model B (Validation): \", reg_model_b.score(Stan_X_valid, Stan_y_valid))\n",
    "print(\"R^2 Adj. Model B (Validation): \", r_squared_adj(reg_model_b.score(Stan_X_valid, Stan_y_valid), len(y), len(X.columns)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Neural Network Models\n",
    "* Apply the MLPRegressor object in the scikit-learn package to build models using the dataset above.\n",
    "* Build two models like we did for Linear Regression above. Baseline & First Iteration models."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Baseline model\n",
    "*Multi-Layer Perceptron"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "#-------------------- Results of MLPRegressor model - Developed with all 13 Features in Dataset -------------------#\n",
      "\n",
      "Tuned MLPRegressor R squared on training data: 0.9933746383235316\n",
      "R^2 Adj. Model on training data :  0.9931995779540314\n",
      "Parameters of training data :  Pipeline(memory=None,\n",
      "     steps=[('scaler', StandardScaler(copy=True, with_mean=True, with_std=True)), ('MLPRegressor', MLPRegressor(activation='relu', alpha=0.01, batch_size='auto', beta_1=0.9,\n",
      "       beta_2=0.999, early_stopping=False, epsilon=1e-08,\n",
      "       hidden_layer_sizes=(100,), learning_rate='constant',\n",
      "       learni...True, solver='lbfgs', tol=0.0001, validation_fraction=0.1,\n",
      "       verbose=False, warm_start=False))])\n",
      "********************************************************************************\n",
      "Tuned MLPRegressorr R squared on testing data: 0.865524160124324\n",
      "R^2 Adj. Model on testing data :  0.7745749143821001\n"
     ]
    }
   ],
   "source": [
    "from sklearn.neural_network import MLPRegressor\n",
    "\n",
    "X=df_boston[[\"CRIM\", \"ZN\", \"INDUS\", \"CHAS\", \"NOX\", \"AGE\",\"RM\", \"DIS\",\"RAD\", \"TAX\", \"PTRATIO\", \"LSTAT\",\"B\" ]] #to get better result remove \"AGE\"\n",
    "y = df_boston[\"MEDV\"].values\n",
    "\n",
    "steps = [('scaler',StandardScaler()),\n",
    "         ('MLPRegressor',MLPRegressor(solver='lbfgs'))]\n",
    "\n",
    "\n",
    "pipeline=Pipeline(steps)\n",
    "\n",
    "parameters = {'MLPRegressor__alpha':np.linspace(0.0001,0.01,50)}\n",
    "\n",
    "X_train, X_test, y_train, y_test =train_test_split(X,y,test_size=0.2,random_state=42)\n",
    "gm_cv = GridSearchCV(pipeline,parameters,cv=4)\n",
    "\n",
    "gm_cv.fit(X_train,y_train)\n",
    "######################################################################################\n",
    "n2=gm_cv.score(X_train,y_train)\n",
    "n2_t=gm_cv.score(X_test, y_test)\n",
    "###################################################################################\n",
    "MLPRegressor=MLPRegressor(solver= 'lbfgs')\n",
    "kfold = KFold(n_splits=6, random_state=42, shuffle=False)\n",
    "reg_model_a =MLPRegressor.fit(X_train[cols], y_train)\n",
    "\n",
    "#######################################################################################\n",
    "\n",
    "print(\"\\n#-------------------- Results of MLPRegressor model - Developed with all 13 Features in Dataset -------------------#\")\n",
    "print(\"\\nTuned MLPRegressor R squared on training data: {}\".format(n2))\n",
    "print(\"R^2 Adj. Model on training data : \", r_squared_adj(n2 , len(y), 13))\n",
    "print(\"Parameters of training data : \", gm_cv.best_estimator_)\n",
    "print(\"********************************************************************************\")\n",
    "print(\"Tuned MLPRegressorr R squared on testing data: {}\".format(r2_t))\n",
    "print(\"R^2 Adj. Model on testing data : \", r_squared_adj(n2_t , len(y), 13))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* Iteration Model - MLP"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "#-------------------- Results of MLPRegressor model - Developed with all 13 Features in Dataset -------------------#\n",
      "\n",
      "Tuned MLPRegressor R squared on training data: 0.9907262238577587\n",
      "R^2 Adj. Model on training data :  0.9904811850572522\n",
      "Parameters of training data :  Pipeline(memory=None,\n",
      "     steps=[('scaler', StandardScaler(copy=True, with_mean=True, with_std=True)), ('MLPRegressor', MLPRegressor(activation='relu', alpha=0.006161224489795919, batch_size='auto',\n",
      "       beta_1=0.9, beta_2=0.999, early_stopping=False, epsilon=1e-08,\n",
      "       hidden_layer_sizes=(100,), learning_rate='constant...True, solver='lbfgs', tol=0.0001, validation_fraction=0.1,\n",
      "       verbose=False, warm_start=False))])\n",
      "********************************************************************************\n",
      "Tuned MLPRegressorr R squared on testing data: 0.865524160124324\n",
      "R^2 Adj. Model on testing data :  0.8744845720087442\n"
     ]
    }
   ],
   "source": [
    "from sklearn.neural_network import MLPRegressor\n",
    "\n",
    "X=df_boston_regr[[\"CRIM\", \"ZN\", \"INDUS\", \"CHAS\", \"NOX\", \"AGE\",\"RM\", \"DIS\",\"RAD\", \"TAX\", \"PTRATIO\", \"LSTAT\",\"B\" ]] #to get better result remove \"AGE\"\n",
    "y = df_boston_regr[\"MEDV\"].values\n",
    "\n",
    "steps = [('scaler',StandardScaler()),\n",
    "         ('MLPRegressor',MLPRegressor(solver='lbfgs'))]\n",
    "\n",
    "\n",
    "pipeline=Pipeline(steps)\n",
    "\n",
    "parameters = {'MLPRegressor__alpha':np.linspace(0.0001,0.01,50)}\n",
    "\n",
    "X_train, X_test, y_train, y_test =train_test_split(X,y,test_size=0.2,random_state=42)\n",
    "gm_cv = GridSearchCV(pipeline,parameters,cv=4)\n",
    "\n",
    "gm_cv.fit(X_train,y_train)\n",
    "######################################################################################\n",
    "n2=gm_cv.score(X_train,y_train)\n",
    "n2_t=gm_cv.score(X_test, y_test)\n",
    "###################################################################################\n",
    "MLPRegressor=MLPRegressor(solver= 'lbfgs')\n",
    "kfold = KFold(n_splits=6, random_state=42, shuffle=False)\n",
    "reg_model_a =MLPRegressor.fit(X_train[cols], y_train)\n",
    "\n",
    "#######################################################################################\n",
    "\n",
    "print(\"\\n#-------------------- Results of MLPRegressor model - Developed with all 13 Features in Dataset -------------------#\")\n",
    "print(\"\\nTuned MLPRegressor R squared on training data: {}\".format(n2))\n",
    "print(\"R^2 Adj. Model on training data : \", r_squared_adj(n2 , len(y), 13))\n",
    "print(\"Parameters of training data : \", gm_cv.best_estimator_)\n",
    "print(\"********************************************************************************\")\n",
    "print(\"Tuned MLPRegressorr R squared on testing data: {}\".format(r2_t))\n",
    "print(\"R^2 Adj. Model on testing data : \", r_squared_adj(n2_t , len(y), 13))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Support Vector Models\n",
    "* Apply the SVR object in the scikit-learn package to build models using the dataset above.\n",
    "* Build two models like we did for Linear Regression above. Baseline & First Iteration models."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "* SVR Baseline Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "# Tuning hyper-parameters\n",
      "0.5945302296947361\n",
      "\n",
      "#-------------------- Results of SVRmodel - Developed with all 13 Features in Dataset -------------------#\n",
      "\n",
      "Tuned SVR R squared on training data: 0.7148617022213721\n",
      "R^2 Adj. Model on training data :  0.7073275602068962\n",
      "********************************************************************************\n",
      "Tuned SVR R squared on testing data: 0.5945302296947361\n",
      "R^2 Adj. Model on testing data :  0.5838165975525239\n"
     ]
    }
   ],
   "source": [
    "from sklearn import svm\n",
    "from sklearn.svm import SVR\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "\n",
    "X=df_boston[[\"CRIM\", \"ZN\", \"INDUS\", \"CHAS\", \"NOX\", \"AGE\",\"RM\", \"DIS\",\"RAD\", \"TAX\", \"PTRATIO\", \"LSTAT\",\"B\" ]] #to get better result remove \"AGE\"\n",
    "y = df_boston[\"MEDV\"].values\n",
    "  \n",
    "# SVM regularization parameter\n",
    "svr_lin = SVR(kernel='linear')\n",
    "parameters = [{'kernel': ['linear']}]\n",
    "\n",
    "X_train, X_test, y_train, y_test =train_test_split(X,y,test_size=0.2,random_state=42)\n",
    "clf = GridSearchCV(svm.SVR(), parameters, cv=5)\n",
    "clf.fit(X_train, y_train)\n",
    "\n",
    "n2=clf.score(X_train, y_train)\n",
    "n2_t=clf.score(X_test, y_test)\n",
    "\n",
    "######################################################################################\n",
    "\n",
    "print(\"# Tuning hyper-parameters\")\n",
    "print(clf.score(X_test, y_test))\n",
    "print(\"\\n#-------------------- Results of SVRmodel - Developed with all 13 Features in Dataset -------------------#\")\n",
    "print(\"\\nTuned SVR R squared on training data: {}\".format(n2))\n",
    "print(\"R^2 Adj. Model on training data : \", r_squared_adj(n2 , len(y), 13))\n",
    "print(\"********************************************************************************\")\n",
    "print(\"Tuned SVR R squared on testing data: {}\".format(n2_t))\n",
    "print(\"R^2 Adj. Model on testing data : \", r_squared_adj(n2_t , len(y), 13))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#SVR Iteration model using transformed data df_boston_regr"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "# Tuning hyper-parameters\n",
      "0.6112893483715194\n",
      "\n",
      "#-------------------- Results of SVRmodel - Developed with all 13 Features in Dataset -------------------#\n",
      "\n",
      "Tuned SVR R squared on training data: 0.7299617616379817\n",
      "R^2 Adj. Model on training data :  0.7228266049332942\n",
      "********************************************************************************\n",
      "Tuned SVR R squared on testing data: 0.6112893483715194\n",
      "R^2 Adj. Model on testing data :  0.6010185384707669\n"
     ]
    }
   ],
   "source": [
    "from sklearn import svm\n",
    "from sklearn.svm import SVR\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "\n",
    "X=df_boston_regr[[\"CRIM\", \"ZN\", \"INDUS\", \"CHAS\", \"NOX\", \"AGE\",\"RM\", \"DIS\",\"RAD\", \"TAX\", \"PTRATIO\", \"LSTAT\",\"B\" ]] #to get better result remove \"AGE\"\n",
    "y = df_boston_regr[\"MEDV\"].values\n",
    "  \n",
    "# SVM regularization parameter\n",
    "svr_lin = SVR(kernel='linear')\n",
    "parameters = [{'kernel': ['linear']}]\n",
    "\n",
    "X_train, X_test, y_train, y_test =train_test_split(X,y,test_size=0.2,random_state=42)\n",
    "clf = GridSearchCV(svm.SVR(), parameters, cv=5)\n",
    "clf.fit(X_train, y_train)\n",
    "\n",
    "n2=clf.score(X_train, y_train)\n",
    "n2_t=clf.score(X_test, y_test)\n",
    "\n",
    "######################################################################################\n",
    "\n",
    "print(\"# Tuning hyper-parameters\")\n",
    "print(clf.score(X_test, y_test))\n",
    "print(\"\\n#-------------------- Results of SVRmodel - Developed with all 13 Features in Dataset -------------------#\")\n",
    "print(\"\\nTuned SVR R squared on training data: {}\".format(n2))\n",
    "print(\"R^2 Adj. Model on training data : \", r_squared_adj(n2 , len(y), 13))\n",
    "print(\"********************************************************************************\")\n",
    "print(\"Tuned SVR R squared on testing data: {}\".format(n2_t))\n",
    "print(\"R^2 Adj. Model on testing data : \", r_squared_adj(n2_t , len(y), 13))"
   ]
  }
 ],
 "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": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}