07_Simple_and_Multiple_Regression.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "07 Simple and Multiple Regression.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "authorship_tag": "ABX9TyOAJPmXXYrhwvtGuE3Xfsil",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/sandipanpaul21/Machine-Learning-in-Python-Code/blob/master/07_Simple_and_Multiple_Regression.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "O1w94qbJnRld"
      },
      "source": [
        "**LINEAR REGRESSION**\n",
        "\n",
        "Linear regression is studied as a model for understanding the relationship between input and output numerical variables.\n",
        "\n",
        "**Linear regression is a linear model**, e.g. a model that assumes a linear relationship between the input variables (x) and the single output variable (y). More specifically, that y can be calculated from a linear combination of the input variables (x).\n",
        "\n",
        "When there is a single input variable (x), the method is referred to as **simple linear regression**. When there are multiple input variables, literature from statistics often refers to the method as **multiple linear regression**\n",
        "\n",
        "Different techniques can be used to prepare or train the linear regression equation from data, the most common of which is called **Ordinary Least Squares**. It is common to therefore refer to a model prepared this way as **Ordinary Least Squares Linear Regression** or just **Least Squares Regression**."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "a7ibydXynwkK"
      },
      "source": [
        "The representation is a linear equation that combines a specific set of **input values (x)** the solution to which is the predicted output for that set of **output values (y)**. As such, both the input values (x) and the output value are numeric.\n",
        "\n",
        "The linear equation assigns one scale factor to each input value or column, called a **coefficient** and represented by the capital Greek letter **Beta (B)**. One additional coefficient is also added, giving the line an additional degree of freedom (e.g. moving up and down on a two-dimensional plot) and is often called the **intercept or the bias coefficient.**\n",
        "\n",
        "For example, in a simple regression problem (a single x and a single y), the form of the model would be:\n",
        "\n",
        "y = B0 + B1 * x1\n",
        "\n",
        "In higher dimensions when we have more than one input (x), the line is called a **plane or a hyper-plane**. The representation therefore is the form of the equation and the specific values used for the coefficients (e.g. B0 and B1 in the above example).\n",
        "\n",
        "It is common to talk about the complexity of a regression model like linear regression. This refers to the **number of coefficients used in the model.**\n",
        "\n",
        "When a **coefficient becomes zero,** it effectively removes the influence of the input variable on the model and therefore from the prediction made from the model (0 * x = 0). This becomes  relevant if you look at **regularization methods** that change the learning algorithm to reduce the complexity of regression models by putting pressure on the absolute size of the coefficients, driving some to zero."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "R7lnhXAzoleK"
      },
      "source": [
        "**Simple Linear Regression**\n",
        "\n",
        "With simple linear regression when we have a *single input*, we can use statistics to estimate the coefficients\n",
        "\n",
        "**Ordinary Least Squares**\n",
        "  - When we have *more than one input* we can use Ordinary Least Squares to estimate the values of the coefficients.\n",
        "  - The *Ordinary Least Squares procedure seeks to minimize the sum of the squared residuals.* This means that given a regression line through the data we calculate the distance from each data point to the regression line, square it, and sum all of the squared errors together. This is the quantity that ordinary least squares seeks to minimize.\n",
        "\n",
        "**Regularization**\n",
        "  - There are extensions of the training of the linear model called regularization methods. \n",
        "  - These seek to *both minimize the sum of the squared error of the model on the training data (using ordinary least squares) but also to reduce the complexity of the model (like the number or absolute size of the sum of all coefficients in the model).*\n",
        "  - Two popular examples of regularization procedures for linear regression are:\n",
        "    1. **Lasso Regression:** where Ordinary Least Squares is modified to also minimize the absolute sum of the coefficients (called L1 regularization).\n",
        "    2. **Ridge Regression:** where Ordinary Least Squares is modified to also minimize the squared absolute sum of the coefficients (called L2 regularization).\n",
        "  - *These methods are effective to use when there is collinearity in your input values and ordinary least squares would overfit the training data.*\n",
        "\n",
        "\n",
        "**Data Preparation for Linear Regression Model**\n",
        "1. **Linear Assumption.** \n",
        "  - Linear regression assumes that the relationship between your input and output is linear. It does not support anything else. \n",
        "  - This may be obvious, but it is good to remember when you have a lot of attributes. You may need to transform data to make the relationship linear (e.g. log transform for an exponential relationship).\n",
        "2. **Remove Noise.**\n",
        "  - Linear regression assumes that your input and output variables are not noisy. Consider using data cleaning operations that let you better expose and clarify the signal in your data. \n",
        "  - This is most important for the output variable and you want to remove outliers in the output variable (y) if possible.\n",
        "3. **Remove Collinearity.** \n",
        "  - Linear regression will over-fit your data when you have highly correlated input variables. Consider calculating pairwise correlations for your input data and removing the most correlated.\n",
        "4. **Gaussian Distributions.** \n",
        "- Linear regression will make more reliable predictions if your input and output variables have a Gaussian distribution. \n",
        "- You may get some benefit using transforms (e.g. log or BoxCox) on you variables to make their distribution more Gaussian looking.\n",
        "5. **Rescale Inputs:** \n",
        "  - Linear regression will often make more reliable predictions if you rescale input variables using standardization or normalization."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FUOE9uSRrEca"
      },
      "source": [
        "**Assumptions of Linear Regression**\n",
        "\n",
        "Linear regression is an analysis that assesses whether one or more predictor variables explain the dependent (criterion) variable.  The regression has five key assumptions:\n",
        "1. Linear relationship\n",
        "2. Multivariate normality\n",
        "3. No or little multicollinearity\n",
        "4. No auto-correlation\n",
        "5. Homoscedasticity\n",
        "\n",
        "A note about sample size.  In Linear regression the sample size rule of thumb is that the regression analysis requires at least 20 cases per independent variable in the analysis."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VGe8PLvZqOpy",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "6e549da6-7c08-40df-8d75-5759a479ae7a"
      },
      "source": [
        "# Libraries \n",
        " \n",
        "from sklearn import datasets \n",
        "import pandas as pd\n",
        "import seaborn as sns\n",
        "import numpy as np\n",
        "import statsmodels.stats.api as sms\n",
        "from statsmodels.compat import lzip\n",
        "from statsmodels.stats.outliers_influence import variance_inflation_factor\n",
        "import statsmodels.api as sm\n",
        "import warnings\n",
        "warnings.filterwarnings(\"ignore\")"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/statsmodels/tools/_testing.py:19: FutureWarning: pandas.util.testing is deprecated. Use the functions in the public API at pandas.testing instead.\n",
            "  import pandas.util.testing as tm\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8eoxSpcYp6I4",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        },
        "outputId": "d26013be-da7a-4fcb-9edc-b79aeeb7e58e"
      },
      "source": [
        "# Boston Dataset for Regression\n",
        "boston = datasets.load_boston()\n",
        "boston_pd = pd.DataFrame(boston.data)\n",
        "boston_pd.columns = boston.feature_names\n",
        "boston_pd[\"HOUSEPRICE\"] = boston.target\n",
        "boston_pd.head()"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "execute_result",
          "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>CRIM</th>\n",
              "      <th>ZN</th>\n",
              "      <th>INDUS</th>\n",
              "      <th>CHAS</th>\n",
              "      <th>NOX</th>\n",
              "      <th>RM</th>\n",
              "      <th>AGE</th>\n",
              "      <th>DIS</th>\n",
              "      <th>RAD</th>\n",
              "      <th>TAX</th>\n",
              "      <th>PTRATIO</th>\n",
              "      <th>B</th>\n",
              "      <th>LSTAT</th>\n",
              "      <th>HOUSEPRICE</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0.00632</td>\n",
              "      <td>18.0</td>\n",
              "      <td>2.31</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.538</td>\n",
              "      <td>6.575</td>\n",
              "      <td>65.2</td>\n",
              "      <td>4.0900</td>\n",
              "      <td>1.0</td>\n",
              "      <td>296.0</td>\n",
              "      <td>15.3</td>\n",
              "      <td>396.90</td>\n",
              "      <td>4.98</td>\n",
              "      <td>24.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0.02731</td>\n",
              "      <td>0.0</td>\n",
              "      <td>7.07</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.469</td>\n",
              "      <td>6.421</td>\n",
              "      <td>78.9</td>\n",
              "      <td>4.9671</td>\n",
              "      <td>2.0</td>\n",
              "      <td>242.0</td>\n",
              "      <td>17.8</td>\n",
              "      <td>396.90</td>\n",
              "      <td>9.14</td>\n",
              "      <td>21.6</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>0.02729</td>\n",
              "      <td>0.0</td>\n",
              "      <td>7.07</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.469</td>\n",
              "      <td>7.185</td>\n",
              "      <td>61.1</td>\n",
              "      <td>4.9671</td>\n",
              "      <td>2.0</td>\n",
              "      <td>242.0</td>\n",
              "      <td>17.8</td>\n",
              "      <td>392.83</td>\n",
              "      <td>4.03</td>\n",
              "      <td>34.7</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>0.03237</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.18</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.458</td>\n",
              "      <td>6.998</td>\n",
              "      <td>45.8</td>\n",
              "      <td>6.0622</td>\n",
              "      <td>3.0</td>\n",
              "      <td>222.0</td>\n",
              "      <td>18.7</td>\n",
              "      <td>394.63</td>\n",
              "      <td>2.94</td>\n",
              "      <td>33.4</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>0.06905</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.18</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.458</td>\n",
              "      <td>7.147</td>\n",
              "      <td>54.2</td>\n",
              "      <td>6.0622</td>\n",
              "      <td>3.0</td>\n",
              "      <td>222.0</td>\n",
              "      <td>18.7</td>\n",
              "      <td>396.90</td>\n",
              "      <td>5.33</td>\n",
              "      <td>36.2</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "      CRIM    ZN  INDUS  CHAS    NOX  ...    TAX  PTRATIO       B  LSTAT  HOUSEPRICE\n",
              "0  0.00632  18.0   2.31   0.0  0.538  ...  296.0     15.3  396.90   4.98        24.0\n",
              "1  0.02731   0.0   7.07   0.0  0.469  ...  242.0     17.8  396.90   9.14        21.6\n",
              "2  0.02729   0.0   7.07   0.0  0.469  ...  242.0     17.8  392.83   4.03        34.7\n",
              "3  0.03237   0.0   2.18   0.0  0.458  ...  222.0     18.7  394.63   2.94        33.4\n",
              "4  0.06905   0.0   2.18   0.0  0.458  ...  222.0     18.7  396.90   5.33        36.2\n",
              "\n",
              "[5 rows x 14 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 2
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "11nahMRNstZv",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "baeb118a-b574-4953-8911-1149cb253f64"
      },
      "source": [
        "# Dataset overall Information\n",
        "boston_pd.info()"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "RangeIndex: 506 entries, 0 to 505\n",
            "Data columns (total 14 columns):\n",
            " #   Column      Non-Null Count  Dtype  \n",
            "---  ------      --------------  -----  \n",
            " 0   CRIM        506 non-null    float64\n",
            " 1   ZN          506 non-null    float64\n",
            " 2   INDUS       506 non-null    float64\n",
            " 3   CHAS        506 non-null    float64\n",
            " 4   NOX         506 non-null    float64\n",
            " 5   RM          506 non-null    float64\n",
            " 6   AGE         506 non-null    float64\n",
            " 7   DIS         506 non-null    float64\n",
            " 8   RAD         506 non-null    float64\n",
            " 9   TAX         506 non-null    float64\n",
            " 10  PTRATIO     506 non-null    float64\n",
            " 11  B           506 non-null    float64\n",
            " 12  LSTAT       506 non-null    float64\n",
            " 13  HOUSEPRICE  506 non-null    float64\n",
            "dtypes: float64(14)\n",
            "memory usage: 55.5 KB\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lJ4_Os_JqrRo",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "0034a431-4553-456d-f8c9-a9757d35fa19"
      },
      "source": [
        "# Let set the BASE MODEL on which we will improve\n",
        "\n",
        "# Assigning Independent and Dependent variables\n",
        "dependent_variable = boston_pd[['HOUSEPRICE']]\n",
        "print(\"DEPENDENT VARIABLE : \",dependent_variable.columns)\n",
        "print(\"\\n\")\n",
        "independent_variables = boston_pd[boston_pd.columns[0:12]]\n",
        "print(\"INDEPENDENT VARIABLES : \")\n",
        "print(independent_variables.columns)"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "DEPENDENT VARIABLE :  Index(['HOUSEPRICE'], dtype='object')\n",
            "\n",
            "\n",
            "INDEPENDENT VARIABLES : \n",
            "Index(['CRIM', 'ZN', 'INDUS', 'CHAS', 'NOX', 'RM', 'AGE', 'DIS', 'RAD', 'TAX',\n",
            "       'PTRATIO', 'B'],\n",
            "      dtype='object')\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "SrUy95nr3NOo",
        "outputId": "51914598-0e9b-442b-ddc3-d8d59079aae2"
      },
      "source": [
        "# Add constant to the dataframe\n",
        "# In Statsmodel library, intercept is not considered. So we need to add Manually\n",
        "independent_variables = sm.add_constant(independent_variables)\n",
        "print(\"After Adding constant, Independent Variable names are\")\n",
        "print(independent_variables.columns)"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "After Adding constant, Independent Variable names are\n",
            "Index(['const', 'CRIM', 'ZN', 'INDUS', 'CHAS', 'NOX', 'RM', 'AGE', 'DIS',\n",
            "       'RAD', 'TAX', 'PTRATIO', 'B'],\n",
            "      dtype='object')\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "prTVZUusqccY",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 709
        },
        "outputId": "84174b49-11de-4af1-d88a-3a49a3efb08d"
      },
      "source": [
        "# Base Model\n",
        "\n",
        "import statsmodels.api as sm\n",
        "from statsmodels.api import OLS\n",
        "base_model = OLS(dependent_variable, independent_variables).fit()\n",
        "base_model.summary()"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<table class=\"simpletable\">\n",
              "<caption>OLS Regression Results</caption>\n",
              "<tr>\n",
              "  <th>Dep. Variable:</th>       <td>HOUSEPRICE</td>    <th>  R-squared:         </th> <td>   0.684</td> \n",
              "</tr>\n",
              "<tr>\n",
              "  <th>Model:</th>                   <td>OLS</td>       <th>  Adj. R-squared:    </th> <td>   0.677</td> \n",
              "</tr>\n",
              "<tr>\n",
              "  <th>Method:</th>             <td>Least Squares</td>  <th>  F-statistic:       </th> <td>   89.01</td> \n",
              "</tr>\n",
              "<tr>\n",
              "  <th>Date:</th>             <td>Tue, 07 Dec 2021</td> <th>  Prob (F-statistic):</th> <td>4.90e-115</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>Time:</th>                 <td>06:55:54</td>     <th>  Log-Likelihood:    </th> <td> -1548.6</td> \n",
              "</tr>\n",
              "<tr>\n",
              "  <th>No. Observations:</th>      <td>   506</td>      <th>  AIC:               </th> <td>   3123.</td> \n",
              "</tr>\n",
              "<tr>\n",
              "  <th>Df Residuals:</th>          <td>   493</td>      <th>  BIC:               </th> <td>   3178.</td> \n",
              "</tr>\n",
              "<tr>\n",
              "  <th>Df Model:</th>              <td>    12</td>      <th>                     </th>     <td> </td>    \n",
              "</tr>\n",
              "<tr>\n",
              "  <th>Covariance Type:</th>      <td>nonrobust</td>    <th>                     </th>     <td> </td>    \n",
              "</tr>\n",
              "</table>\n",
              "<table class=\"simpletable\">\n",
              "<tr>\n",
              "     <td></td>        <th>coef</th>     <th>std err</th>      <th>t</th>      <th>P>|t|</th>  <th>[0.025</th>    <th>0.975]</th>  \n",
              "</tr>\n",
              "<tr>\n",
              "  <th>const</th>   <td>   20.6526</td> <td>    5.368</td> <td>    3.848</td> <td> 0.000</td> <td>   10.106</td> <td>   31.199</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>CRIM</th>    <td>   -0.1599</td> <td>    0.036</td> <td>   -4.467</td> <td> 0.000</td> <td>   -0.230</td> <td>   -0.090</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>ZN</th>      <td>    0.0389</td> <td>    0.015</td> <td>    2.573</td> <td> 0.010</td> <td>    0.009</td> <td>    0.069</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>INDUS</th>   <td>   -0.0279</td> <td>    0.068</td> <td>   -0.413</td> <td> 0.680</td> <td>   -0.161</td> <td>    0.105</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>CHAS</th>    <td>    3.2166</td> <td>    0.948</td> <td>    3.393</td> <td> 0.001</td> <td>    1.354</td> <td>    5.079</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>NOX</th>     <td>  -20.4846</td> <td>    4.201</td> <td>   -4.877</td> <td> 0.000</td> <td>  -28.738</td> <td>  -12.231</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>RM</th>      <td>    6.1231</td> <td>    0.389</td> <td>   15.731</td> <td> 0.000</td> <td>    5.358</td> <td>    6.888</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>AGE</th>     <td>   -0.0459</td> <td>    0.014</td> <td>   -3.356</td> <td> 0.001</td> <td>   -0.073</td> <td>   -0.019</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>DIS</th>     <td>   -1.5549</td> <td>    0.220</td> <td>   -7.077</td> <td> 0.000</td> <td>   -1.987</td> <td>   -1.123</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>RAD</th>     <td>    0.2816</td> <td>    0.073</td> <td>    3.852</td> <td> 0.000</td> <td>    0.138</td> <td>    0.425</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>TAX</th>     <td>   -0.0117</td> <td>    0.004</td> <td>   -2.832</td> <td> 0.005</td> <td>   -0.020</td> <td>   -0.004</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>PTRATIO</th> <td>   -1.0142</td> <td>    0.144</td> <td>   -7.040</td> <td> 0.000</td> <td>   -1.297</td> <td>   -0.731</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>B</th>       <td>    0.0136</td> <td>    0.003</td> <td>    4.657</td> <td> 0.000</td> <td>    0.008</td> <td>    0.019</td>\n",
              "</tr>\n",
              "</table>\n",
              "<table class=\"simpletable\">\n",
              "<tr>\n",
              "  <th>Omnibus:</th>       <td>267.269</td> <th>  Durbin-Watson:     </th> <td>   0.934</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>Prob(Omnibus):</th> <td> 0.000</td>  <th>  Jarque-Bera (JB):  </th> <td>2542.250</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>Skew:</th>          <td> 2.108</td>  <th>  Prob(JB):          </th> <td>    0.00</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>Kurtosis:</th>      <td>13.139</td>  <th>  Cond. No.          </th> <td>1.48e+04</td>\n",
              "</tr>\n",
              "</table><br/><br/>Warnings:<br/>[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.<br/>[2] The condition number is large, 1.48e+04. This might indicate that there are<br/>strong multicollinearity or other numerical problems."
            ],
            "text/plain": [
              "<class 'statsmodels.iolib.summary.Summary'>\n",
              "\"\"\"\n",
              "                            OLS Regression Results                            \n",
              "==============================================================================\n",
              "Dep. Variable:             HOUSEPRICE   R-squared:                       0.684\n",
              "Model:                            OLS   Adj. R-squared:                  0.677\n",
              "Method:                 Least Squares   F-statistic:                     89.01\n",
              "Date:                Tue, 07 Dec 2021   Prob (F-statistic):          4.90e-115\n",
              "Time:                        06:55:54   Log-Likelihood:                -1548.6\n",
              "No. Observations:                 506   AIC:                             3123.\n",
              "Df Residuals:                     493   BIC:                             3178.\n",
              "Df Model:                          12                                         \n",
              "Covariance Type:            nonrobust                                         \n",
              "==============================================================================\n",
              "                 coef    std err          t      P>|t|      [0.025      0.975]\n",
              "------------------------------------------------------------------------------\n",
              "const         20.6526      5.368      3.848      0.000      10.106      31.199\n",
              "CRIM          -0.1599      0.036     -4.467      0.000      -0.230      -0.090\n",
              "ZN             0.0389      0.015      2.573      0.010       0.009       0.069\n",
              "INDUS         -0.0279      0.068     -0.413      0.680      -0.161       0.105\n",
              "CHAS           3.2166      0.948      3.393      0.001       1.354       5.079\n",
              "NOX          -20.4846      4.201     -4.877      0.000     -28.738     -12.231\n",
              "RM             6.1231      0.389     15.731      0.000       5.358       6.888\n",
              "AGE           -0.0459      0.014     -3.356      0.001      -0.073      -0.019\n",
              "DIS           -1.5549      0.220     -7.077      0.000      -1.987      -1.123\n",
              "RAD            0.2816      0.073      3.852      0.000       0.138       0.425\n",
              "TAX           -0.0117      0.004     -2.832      0.005      -0.020      -0.004\n",
              "PTRATIO       -1.0142      0.144     -7.040      0.000      -1.297      -0.731\n",
              "B              0.0136      0.003      4.657      0.000       0.008       0.019\n",
              "==============================================================================\n",
              "Omnibus:                      267.269   Durbin-Watson:                   0.934\n",
              "Prob(Omnibus):                  0.000   Jarque-Bera (JB):             2542.250\n",
              "Skew:                           2.108   Prob(JB):                         0.00\n",
              "Kurtosis:                      13.139   Cond. No.                     1.48e+04\n",
              "==============================================================================\n",
              "\n",
              "Warnings:\n",
              "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
              "[2] The condition number is large, 1.48e+04. This might indicate that there are\n",
              "strong multicollinearity or other numerical problems.\n",
              "\"\"\""
            ]
          },
          "metadata": {},
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZPFz19ZaK0Cp"
      },
      "source": [
        "**OUTPUT EXPLAINATION**\n",
        "\n",
        "**Omnibus/Prob(Omnibus)**\n",
        "- **Prob(Omnibus)** is a statistical test measuring the *probability the residuals are normally distributed. 1 would indicate perfectly normal distribution.* \n",
        "- **Omnibus** describes the normalcy of the *distribution of our residuals* using skew and kurtosis as measurements. A *0 would indicate perfect normalcy.* \n",
        "- The **Prob(Omnibus)** indicates the probability that the residuals are normally distributed. We hope to see something close to 1 here. \n",
        "- In this case Omnibus = 267 (way higher than 1) and Prob(Omnibus) = 0 which (normally = 1)is low. So the data is not normal, not ideal. \n",
        "\n",
        "**Skew** \n",
        "- Skew is a measurement of symmetry in our data, with 0 being perfect symmetry. \n",
        "- We want to see something close to zero, indicating the residual distribution is normal. \n",
        "- Note that this value also drives the Omnibus\n",
        "- In this case, Skewness = 2.1, way higher than 0 so error/residual is skewwed\n",
        "\n",
        "**Kurtosis** \n",
        "- Kurtosis measures the *peakiness of our data.*\n",
        "- *Kurtosis of the normal distribution is 3.0*\n",
        "- In this case, Kurtosis = 13.14 which is way too higher\n",
        "\n",
        "**Durbin-Watson**\n",
        "- Durbin-Watson is a *measurement of homoscedasticity, or an even distribution of errors throughout our data.* \n",
        "- Heteroscedasticity would imply an uneven distribution, for example as the data point grows higher the relative error grows higher. *Ideal homoscedasticity will lie between 1 and 2.* \n",
        "- In this case, Durbin-Watson = 0.78 is close, but within limits.\n",
        "\n",
        "**Jarque-Bera (JB)/Prob(JB) -** \n",
        "- Jarque-Bera (JB) and Prob(JB) are *alternate methods of measuring the same value as Omnibus and Prob(Omnibus)* using skewness and kurtosis. We use these values to confirm each other\n",
        "- It is like the Omnibus test in that it tests both skew and kurtosis. \n",
        "- It is also performed for the distribution analysis of the regression errors.\n",
        "- A *large value of JB test indicates that the errors are not normally distributed.*\n",
        "- In this case, JB = 2542 which is way too higher so error are not normally distributed\n",
        "\n",
        "**Condition Number -**\n",
        "- Condition number is a measurement of the sensitivity of our model as compared to the size of changes in the data it is analyzing. \n",
        "- *Multicollinearity is strongly implied by a high condition number.* \n",
        "- Multicollinearity a term to describe two or more independent variables that are strongly related to each other and are falsely affecting our predicted variable by redundancy.\n",
        "- When we have multicollinearity, we can expect much higher fluctuations to small changes in the data hence, we hope to see a relatively small number, something below 30. \n",
        "- In this case, Condition Number = well above 30, so multicollinearity present\n",
        "\n",
        "**R Square and Adjusted R Square -**\n",
        "- R Squared is the measurement of how much of the independent variable is explained by changed in our dependent variable\n",
        "- In percentage terms, 0.68 would mean our model explains 68% of the dependent variable\n",
        "- Adjusted R-squared is important for analyzing multiple dependent variables’ efficacy on the model. \n",
        "- R-squared value will never go down with additional variables, only equal or higher. \n",
        "- Therefore, your model could look more accurate with multiple variables even if they are poorly contributing. \n",
        "- The adjusted R-squared penalizes the R-squared formula based on the number of variables, therefore a lower adjusted score may be telling you some variables are not contributing to your model’s R-squared properly.\n",
        "- Both measures model performance and Possible values range from 0.0 to 1.0. \n",
        "- Higher the Adjusted R Square, better is the model\n",
        "\n",
        "**F Statistics and Prob(F Statistics) -**\n",
        "- This test for overall significance has the following two hypothesis:\n",
        "- **Null hypothesis :** Model with no independent variables fits the data as well as your model.\n",
        "- **Alternative hypothesis :** Model fits the data better than the intercept-only model.\n",
        "- In this case, P(F-Statistics) = less than 0.05 suggests Independent variables are important\n",
        "\n",
        "**P Value -** \n",
        "- If the p-value for a variable is less than your significance level (0.05), sample data provide enough evidence to reject the null hypothesis for the entire population.\n",
        "- Null hypothesis that the variable has no correlation with the dependent variable. \n",
        "- If No correlation, no association between the changes in the independent & dependent variable.\n",
        "- In this case, Remove all variables with P values greater than 0.05\n",
        "- P>|t| uses the t statistic to produce the p value, a measurement of how likely your coefficient is measured through our model by chance. \n",
        "- The p value of 0.68 for Wealth is saying there is a 68% chance the Indus variable has no affect on the dependent variable, Houseprice and our results are produced by chance. \n",
        "- A common alpha is 0.05, which few of our variables pass in this instance.\n",
        "\n",
        "**NOTE :** Log-Likelihood, AIC and BIC are for LOGISTIC REGRESSION MODEL"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "AgkzKtyUUSlj",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "f3048d58-61af-4c16-9334-f18a3b72b8b2"
      },
      "source": [
        "# Model Properties\n",
        "\n",
        "# Normality of the residuals\n",
        "# Jarque-Bera test:\n",
        "name = ['Jarque-Bera', 'Chi^2 two-tail prob.', 'Skew', 'Kurtosis']\n",
        "test = sms.jarque_bera(base_model.resid)\n",
        "base_model_properties_1 = lzip(name, test)\n",
        "print(\"BASE MODEL PARAMETERS\")\n",
        "print(\"\\n\")\n",
        "print(\"Normality of the residuals\")\n",
        "print(\"Jarque-Bera test : \")\n",
        "print(\"JB, ideal is close to 0\")\n",
        "print(\"Probability(JB), less than 0.05\")\n",
        "print(\"Skewness, ideal is -1 to 1\")\n",
        "print(\"Kurtosis, ideal value is 3\")\n",
        "print(pd.DataFrame(base_model_properties_1,columns = ['Test','Values']).round(2))\n",
        "\n",
        "# Omni test:\n",
        "print(\"\\n\")\n",
        "print(\"Normality of the residuals (also)\")\n",
        "print(\"Omni Test : \")\n",
        "print(\"Omni, ideal is close to 0\")\n",
        "print(\"Probability(Omni), less than 0.05\")\n",
        "name = ['Chi^2', 'Two-tail probability']\n",
        "test = sms.omni_normtest(base_model.resid)\n",
        "print(pd.DataFrame(lzip(name, test),columns = ['Test','Values']).round(2))\n",
        "\n",
        "# Multicollinearity\n",
        "print(\"\\n\")\n",
        "print(\"Multicollinearity\")\n",
        "print(\"Ideal Condition Number will be less than 30\")\n",
        "print(\"Condition number:\",np.linalg.cond(base_model.model.exog).round(2))\n",
        "\n",
        "# Adjusted R Square\n",
        "print(\"\\n\")\n",
        "print(\"Model Performance\")\n",
        "print(\"Adjusted R Square, ideal is close to 1\")\n",
        "print(\"Adjusted R Square : \",base_model.rsquared_adj.round(2))\n",
        "\n",
        "# F Statistics\n",
        "print(\"\\n\")\n",
        "print(\"Significance of Independent Variables\")\n",
        "print(\"Probability(F Stat), less than 0.05 suggests independent variables are important\")\n",
        "print(\"F Statistics : \",base_model.fvalue.round(2))\n",
        "print(\"Probability of F Statistics : \",base_model.f_pvalue.round(2))\n",
        "\n",
        "# Durbin Watson Test\n",
        "print(\"\\n\")\n",
        "print(\"Homoscedasticity of Error\")\n",
        "print(\"Durbin Watson, ideal is between 1 to 2\")\n",
        "print(\"Durbin Watson Value : 0.78\")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "BASE MODEL PARAMETERS\n",
            "\n",
            "\n",
            "Normality of the residuals\n",
            "Jarque-Bera test : \n",
            "JB, ideal is close to 0\n",
            "Probability(JB), less than 0.05\n",
            "Skewness, ideal is -1 to 1\n",
            "Kurtosis, ideal value is 3\n",
            "                   Test   Values\n",
            "0           Jarque-Bera  2542.25\n",
            "1  Chi^2 two-tail prob.     0.00\n",
            "2                  Skew     2.11\n",
            "3              Kurtosis    13.14\n",
            "\n",
            "\n",
            "Normality of the residuals (also)\n",
            "Omni Test : \n",
            "Omni, ideal is close to 0\n",
            "Probability(Omni), less than 0.05\n",
            "                   Test  Values\n",
            "0                 Chi^2  267.27\n",
            "1  Two-tail probability    0.00\n",
            "\n",
            "\n",
            "Multicollinearity\n",
            "Ideal Condition Number will be less than 30\n",
            "Condition number: 14793.16\n",
            "\n",
            "\n",
            "Model Performance\n",
            "Adjusted R Square, ideal is close to 1\n",
            "Adjusted R Square :  0.68\n",
            "\n",
            "\n",
            "Significance of Independent Variables\n",
            "Probability(F Stat), less than 0.05 suggests independent variables are important\n",
            "F Statistics :  89.01\n",
            "Probability of F Statistics :  0.0\n",
            "\n",
            "\n",
            "Homoscedasticity of Error\n",
            "Durbin Watson, ideal is between 1 to 2\n",
            "Durbin Watson Value : 0.78\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_7l05MKW_mR6",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "a652f80b-b8e0-46b6-fad1-2e89f1a37d48"
      },
      "source": [
        "# Let test the Base Model\n",
        "\n",
        "# Dividing Data into training and test set\n",
        "from sklearn.model_selection import train_test_split \n",
        "data_train, data_test, house_price_train, house_price_test = train_test_split(independent_variables,dependent_variable,train_size = 0.7,random_state = 21)\n",
        "print(\"Training Data Shape :\", data_train.shape)\n",
        "print(\"Training House Price Shape :\", house_price_train.shape)\n",
        "print(\"Test Data Shape :\",data_test.shape)\n",
        "print(\"Test House Price Shape :\",house_price_test.shape)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Training Data Shape : (354, 13)\n",
            "Training House Price Shape : (354, 1)\n",
            "Test Data Shape : (152, 13)\n",
            "Test House Price Shape : (152, 1)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zpi0PRoS_0LZ",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 709
        },
        "outputId": "228a3fd5-ef28-47fc-ba32-1eeab0ba7313"
      },
      "source": [
        "# Base Model\n",
        "Model_1 = OLS(house_price_train,data_train).fit()\n",
        "Model_1.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<table class=\"simpletable\">\n",
              "<caption>OLS Regression Results</caption>\n",
              "<tr>\n",
              "  <th>Dep. Variable:</th>       <td>HOUSEPRICE</td>    <th>  R-squared:         </th> <td>   0.691</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>Model:</th>                   <td>OLS</td>       <th>  Adj. R-squared:    </th> <td>   0.680</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>Method:</th>             <td>Least Squares</td>  <th>  F-statistic:       </th> <td>   63.54</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>Date:</th>             <td>Tue, 26 Jan 2021</td> <th>  Prob (F-statistic):</th> <td>2.32e-79</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>Time:</th>                 <td>13:35:59</td>     <th>  Log-Likelihood:    </th> <td> -1069.2</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>No. Observations:</th>      <td>   354</td>      <th>  AIC:               </th> <td>   2164.</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>Df Residuals:</th>          <td>   341</td>      <th>  BIC:               </th> <td>   2215.</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>Df Model:</th>              <td>    12</td>      <th>                     </th>     <td> </td>   \n",
              "</tr>\n",
              "<tr>\n",
              "  <th>Covariance Type:</th>      <td>nonrobust</td>    <th>                     </th>     <td> </td>   \n",
              "</tr>\n",
              "</table>\n",
              "<table class=\"simpletable\">\n",
              "<tr>\n",
              "     <td></td>        <th>coef</th>     <th>std err</th>      <th>t</th>      <th>P>|t|</th>  <th>[0.025</th>    <th>0.975]</th>  \n",
              "</tr>\n",
              "<tr>\n",
              "  <th>const</th>   <td>   25.7285</td> <td>    6.186</td> <td>    4.159</td> <td> 0.000</td> <td>   13.562</td> <td>   37.895</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>CRIM</th>    <td>   -0.2415</td> <td>    0.055</td> <td>   -4.364</td> <td> 0.000</td> <td>   -0.350</td> <td>   -0.133</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>ZN</th>      <td>    0.0495</td> <td>    0.018</td> <td>    2.676</td> <td> 0.008</td> <td>    0.013</td> <td>    0.086</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>INDUS</th>   <td>   -0.0334</td> <td>    0.075</td> <td>   -0.442</td> <td> 0.658</td> <td>   -0.182</td> <td>    0.115</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>CHAS</th>    <td>    2.6172</td> <td>    1.052</td> <td>    2.488</td> <td> 0.013</td> <td>    0.548</td> <td>    4.686</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>NOX</th>     <td>  -20.1834</td> <td>    4.897</td> <td>   -4.121</td> <td> 0.000</td> <td>  -29.816</td> <td>  -10.550</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>RM</th>      <td>    5.3255</td> <td>    0.443</td> <td>   12.028</td> <td> 0.000</td> <td>    4.455</td> <td>    6.196</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>AGE</th>     <td>   -0.0414</td> <td>    0.016</td> <td>   -2.615</td> <td> 0.009</td> <td>   -0.073</td> <td>   -0.010</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>DIS</th>     <td>   -1.5677</td> <td>    0.252</td> <td>   -6.222</td> <td> 0.000</td> <td>   -2.063</td> <td>   -1.072</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>RAD</th>     <td>    0.3606</td> <td>    0.085</td> <td>    4.229</td> <td> 0.000</td> <td>    0.193</td> <td>    0.528</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>TAX</th>     <td>   -0.0131</td> <td>    0.005</td> <td>   -2.837</td> <td> 0.005</td> <td>   -0.022</td> <td>   -0.004</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>PTRATIO</th> <td>   -1.1124</td> <td>    0.173</td> <td>   -6.418</td> <td> 0.000</td> <td>   -1.453</td> <td>   -0.772</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>B</th>       <td>    0.0175</td> <td>    0.004</td> <td>    4.936</td> <td> 0.000</td> <td>    0.011</td> <td>    0.025</td>\n",
              "</tr>\n",
              "</table>\n",
              "<table class=\"simpletable\">\n",
              "<tr>\n",
              "  <th>Omnibus:</th>       <td>186.735</td> <th>  Durbin-Watson:     </th> <td>   2.016</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>Prob(Omnibus):</th> <td> 0.000</td>  <th>  Jarque-Bera (JB):  </th> <td>1626.044</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>Skew:</th>          <td> 2.039</td>  <th>  Prob(JB):          </th> <td>    0.00</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>Kurtosis:</th>      <td>12.675</td>  <th>  Cond. No.          </th> <td>1.49e+04</td>\n",
              "</tr>\n",
              "</table><br/><br/>Warnings:<br/>[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.<br/>[2] The condition number is large, 1.49e+04. This might indicate that there are<br/>strong multicollinearity or other numerical problems."
            ],
            "text/plain": [
              "<class 'statsmodels.iolib.summary.Summary'>\n",
              "\"\"\"\n",
              "                            OLS Regression Results                            \n",
              "==============================================================================\n",
              "Dep. Variable:             HOUSEPRICE   R-squared:                       0.691\n",
              "Model:                            OLS   Adj. R-squared:                  0.680\n",
              "Method:                 Least Squares   F-statistic:                     63.54\n",
              "Date:                Tue, 26 Jan 2021   Prob (F-statistic):           2.32e-79\n",
              "Time:                        13:35:59   Log-Likelihood:                -1069.2\n",
              "No. Observations:                 354   AIC:                             2164.\n",
              "Df Residuals:                     341   BIC:                             2215.\n",
              "Df Model:                          12                                         \n",
              "Covariance Type:            nonrobust                                         \n",
              "==============================================================================\n",
              "                 coef    std err          t      P>|t|      [0.025      0.975]\n",
              "------------------------------------------------------------------------------\n",
              "const         25.7285      6.186      4.159      0.000      13.562      37.895\n",
              "CRIM          -0.2415      0.055     -4.364      0.000      -0.350      -0.133\n",
              "ZN             0.0495      0.018      2.676      0.008       0.013       0.086\n",
              "INDUS         -0.0334      0.075     -0.442      0.658      -0.182       0.115\n",
              "CHAS           2.6172      1.052      2.488      0.013       0.548       4.686\n",
              "NOX          -20.1834      4.897     -4.121      0.000     -29.816     -10.550\n",
              "RM             5.3255      0.443     12.028      0.000       4.455       6.196\n",
              "AGE           -0.0414      0.016     -2.615      0.009      -0.073      -0.010\n",
              "DIS           -1.5677      0.252     -6.222      0.000      -2.063      -1.072\n",
              "RAD            0.3606      0.085      4.229      0.000       0.193       0.528\n",
              "TAX           -0.0131      0.005     -2.837      0.005      -0.022      -0.004\n",
              "PTRATIO       -1.1124      0.173     -6.418      0.000      -1.453      -0.772\n",
              "B              0.0175      0.004      4.936      0.000       0.011       0.025\n",
              "==============================================================================\n",
              "Omnibus:                      186.735   Durbin-Watson:                   2.016\n",
              "Prob(Omnibus):                  0.000   Jarque-Bera (JB):             1626.044\n",
              "Skew:                           2.039   Prob(JB):                         0.00\n",
              "Kurtosis:                      12.675   Cond. No.                     1.49e+04\n",
              "==============================================================================\n",
              "\n",
              "Warnings:\n",
              "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
              "[2] The condition number is large, 1.49e+04. This might indicate that there are\n",
              "strong multicollinearity or other numerical problems.\n",
              "\"\"\""
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vhTxPiBCmOtv",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "c050694d-6958-46bb-bf00-ef82b64ead76"
      },
      "source": [
        "# Model Properties\n",
        "\n",
        "# Normality of the residuals\n",
        "# Jarque-Bera test:\n",
        "name = ['Jarque-Bera', 'Chi^2 two-tail prob.', 'Skew', 'Kurtosis']\n",
        "test = sms.jarque_bera(Model_1.resid)\n",
        "model_properties_1 = lzip(name, test)\n",
        "print(\"MODEL 1 PARAMETERS\")\n",
        "print(\"\\n\")\n",
        "print(\"Normality of the residuals\")\n",
        "print(\"Jarque-Bera test : \")\n",
        "print(\"JB, ideal is close to 0\")\n",
        "print(\"Probability(JB), less than 0.05\")\n",
        "print(\"Skewness, ideal is -1 to 1\")\n",
        "print(\"Kurtosis, ideal value is 3\")\n",
        "print(pd.DataFrame(model_properties_1,columns = ['Test','Values']).round(2))\n",
        "\n",
        "# Omni test:\n",
        "print(\"Omni Test : \")\n",
        "print(\"Omni, ideal is close to 0\")\n",
        "print(\"Probability(Omni), less than 0.05\")\n",
        "name = ['Chi^2', 'Two-tail probability']\n",
        "test = sms.omni_normtest(Model_1.resid)\n",
        "print(pd.DataFrame(lzip(name, test),columns = ['Test','Values']).round(2))\n",
        "\n",
        "# Multicollinearity\n",
        "print(\"\\n\")\n",
        "print(\"Multicollinearity\")\n",
        "print(\"Ideal Condition Number will be less than 30\")\n",
        "print(\"Condition number:\",np.linalg.cond(Model_1.model.exog).round(2))\n",
        "\n",
        "# Adjusted R Square\n",
        "print(\"\\n\")\n",
        "print(\"Model Performance\")\n",
        "print(\"Adjusted R Square, ideal is close to 1\")\n",
        "print(\"Adjusted R Square : \",Model_1.rsquared_adj.round(2))\n",
        "\n",
        "# F Statistics\n",
        "print(\"\\n\")\n",
        "print(\"Significance of Independent Variables\")\n",
        "print(\"Probability(F Stat), less than 0.05 suggests independent variables are important\")\n",
        "print(\"F Statistics : \",base_model.fvalue.round(2))\n",
        "print(\"Probability of F Statistics : \",Model_1.f_pvalue.round(2))\n",
        "\n",
        "# Durbin Watson Test\n",
        "print(\"\\n\")\n",
        "print(\"Homoscedasticity of Error\")\n",
        "print(\"Durbin Watson, ideal is between 1 to 2\")\n",
        "print(\"Durbin Watson Value : 1.99\")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "MODEL 1 PARAMETERS\n",
            "\n",
            "\n",
            "Normality of the residuals\n",
            "Jarque-Bera test : \n",
            "JB, ideal is close to 0\n",
            "Probability(JB), less than 0.05\n",
            "Skewness, ideal is -1 to 1\n",
            "Kurtosis, ideal value is 3\n",
            "                   Test   Values\n",
            "0           Jarque-Bera  1626.04\n",
            "1  Chi^2 two-tail prob.     0.00\n",
            "2                  Skew     2.04\n",
            "3              Kurtosis    12.68\n",
            "Omni Test : \n",
            "Omni, ideal is close to 0\n",
            "Probability(Omni), less than 0.05\n",
            "                   Test  Values\n",
            "0                 Chi^2  186.74\n",
            "1  Two-tail probability    0.00\n",
            "\n",
            "\n",
            "Multicollinearity\n",
            "Ideal Condition Number will be less than 30\n",
            "Condition number: 14949.83\n",
            "\n",
            "\n",
            "Model Performance\n",
            "Adjusted R Square, ideal is close to 1\n",
            "Adjusted R Square :  0.68\n",
            "\n",
            "\n",
            "Significance of Independent Variables\n",
            "Probability(F Stat), less than 0.05 suggests independent variables are important\n",
            "F Statistics :  89.01\n",
            "Probability of F Statistics :  0.0\n",
            "\n",
            "\n",
            "Homoscedasticity of Error\n",
            "Durbin Watson, ideal is between 1 to 2\n",
            "Durbin Watson Value : 1.99\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xnIM_5IKBoOm",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "f946a06b-3fcf-4ba4-eb0a-8a9a50402276"
      },
      "source": [
        "# Base Model Test and Metrics\n",
        "\n",
        "from sklearn import metrics\n",
        "Prediction_1 = Model_1.predict(data_test)\n",
        "print(\"Mean Absolute Error :\",metrics.mean_absolute_error(house_price_test,Prediction_1).round(2))\n",
        "print(\"Mean Squared Error :\",metrics.mean_squared_error(house_price_test,Prediction_1).round(2))\n",
        "print(\"Root Mean Squared Error :\",metrics.mean_squared_error(house_price_test,Prediction_1).round(2))\n",
        "\n",
        "def mape(actual, pred): \n",
        "    actual, pred = np.array(actual), np.array(pred)\n",
        "    return np.mean(np.abs((actual - pred) / actual)) * 100\n",
        "\n",
        "print(\"Mean Absolute Percentage Error :\",mape(house_price_test,Prediction_1).round(2))\n",
        "\n",
        "# MAE is Absolute difference between the Model Predictions & True(Actual) values\n",
        "# MSE is Square difference between Model Predictions & True values\n",
        "# MSE affected by Outliers, because we Square the distance\n",
        "# RMSE, square root(MSE) therefore cancelling the affect the square \n",
        "# MAE value can range from 0 to infinity, difficult to interpret\n",
        "# MAPE is equivalent to MAE but provide the error in a percentage(0-100)\n",
        "\n",
        "# MAPE = 52,average difference between predicted & actual is 52% which is very high"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Mean Absolute Error : 4.01\n",
            "Mean Squared Error : 33.66\n",
            "Root Mean Squared Error : 33.66\n",
            "Mean Absolute Percentage Error : 52.39\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "R6DPYZNwJvSO",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 668
        },
        "outputId": "39ec8dc8-7826-4a8b-ca2f-34afd8dbb015"
      },
      "source": [
        "# Model 2, first Data Cleaning\n",
        "\n",
        "# Correlation \n",
        "# Increasing the size of the plot\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "fig, ax = plt.subplots(figsize=(11,11)) # Sample figsize in inches, change the number accordingly\n",
        "sns.heatmap(independent_variables.corr(method = 'spearman'), annot= True, cmap = 'Blues',ax = ax)\n",
        "\n",
        "# Inference : \n",
        "# Lots of High Correlation Values"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7f3e43dbe080>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 13
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 792x792 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "n-jBkgOTjtla",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 452
        },
        "outputId": "ce491975-6677-4a5c-b230-541f4c1e1ae1"
      },
      "source": [
        "# Filtering Highly Positive or Negative Correlated Values\n",
        "\n",
        "# Create correlation matrix\n",
        "corr_matrix = independent_variables.corr().round(2)\n",
        "corr_matrix"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "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>const</th>\n",
              "      <th>CRIM</th>\n",
              "      <th>ZN</th>\n",
              "      <th>INDUS</th>\n",
              "      <th>CHAS</th>\n",
              "      <th>NOX</th>\n",
              "      <th>RM</th>\n",
              "      <th>AGE</th>\n",
              "      <th>DIS</th>\n",
              "      <th>RAD</th>\n",
              "      <th>TAX</th>\n",
              "      <th>PTRATIO</th>\n",
              "      <th>B</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>const</th>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>CRIM</th>\n",
              "      <td>NaN</td>\n",
              "      <td>1.00</td>\n",
              "      <td>-0.20</td>\n",
              "      <td>0.41</td>\n",
              "      <td>-0.06</td>\n",
              "      <td>0.42</td>\n",
              "      <td>-0.22</td>\n",
              "      <td>0.35</td>\n",
              "      <td>-0.38</td>\n",
              "      <td>0.63</td>\n",
              "      <td>0.58</td>\n",
              "      <td>0.29</td>\n",
              "      <td>-0.39</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>ZN</th>\n",
              "      <td>NaN</td>\n",
              "      <td>-0.20</td>\n",
              "      <td>1.00</td>\n",
              "      <td>-0.53</td>\n",
              "      <td>-0.04</td>\n",
              "      <td>-0.52</td>\n",
              "      <td>0.31</td>\n",
              "      <td>-0.57</td>\n",
              "      <td>0.66</td>\n",
              "      <td>-0.31</td>\n",
              "      <td>-0.31</td>\n",
              "      <td>-0.39</td>\n",
              "      <td>0.18</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>INDUS</th>\n",
              "      <td>NaN</td>\n",
              "      <td>0.41</td>\n",
              "      <td>-0.53</td>\n",
              "      <td>1.00</td>\n",
              "      <td>0.06</td>\n",
              "      <td>0.76</td>\n",
              "      <td>-0.39</td>\n",
              "      <td>0.64</td>\n",
              "      <td>-0.71</td>\n",
              "      <td>0.60</td>\n",
              "      <td>0.72</td>\n",
              "      <td>0.38</td>\n",
              "      <td>-0.36</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>CHAS</th>\n",
              "      <td>NaN</td>\n",
              "      <td>-0.06</td>\n",
              "      <td>-0.04</td>\n",
              "      <td>0.06</td>\n",
              "      <td>1.00</td>\n",
              "      <td>0.09</td>\n",
              "      <td>0.09</td>\n",
              "      <td>0.09</td>\n",
              "      <td>-0.10</td>\n",
              "      <td>-0.01</td>\n",
              "      <td>-0.04</td>\n",
              "      <td>-0.12</td>\n",
              "      <td>0.05</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>NOX</th>\n",
              "      <td>NaN</td>\n",
              "      <td>0.42</td>\n",
              "      <td>-0.52</td>\n",
              "      <td>0.76</td>\n",
              "      <td>0.09</td>\n",
              "      <td>1.00</td>\n",
              "      <td>-0.30</td>\n",
              "      <td>0.73</td>\n",
              "      <td>-0.77</td>\n",
              "      <td>0.61</td>\n",
              "      <td>0.67</td>\n",
              "      <td>0.19</td>\n",
              "      <td>-0.38</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>RM</th>\n",
              "      <td>NaN</td>\n",
              "      <td>-0.22</td>\n",
              "      <td>0.31</td>\n",
              "      <td>-0.39</td>\n",
              "      <td>0.09</td>\n",
              "      <td>-0.30</td>\n",
              "      <td>1.00</td>\n",
              "      <td>-0.24</td>\n",
              "      <td>0.21</td>\n",
              "      <td>-0.21</td>\n",
              "      <td>-0.29</td>\n",
              "      <td>-0.36</td>\n",
              "      <td>0.13</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>AGE</th>\n",
              "      <td>NaN</td>\n",
              "      <td>0.35</td>\n",
              "      <td>-0.57</td>\n",
              "      <td>0.64</td>\n",
              "      <td>0.09</td>\n",
              "      <td>0.73</td>\n",
              "      <td>-0.24</td>\n",
              "      <td>1.00</td>\n",
              "      <td>-0.75</td>\n",
              "      <td>0.46</td>\n",
              "      <td>0.51</td>\n",
              "      <td>0.26</td>\n",
              "      <td>-0.27</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>DIS</th>\n",
              "      <td>NaN</td>\n",
              "      <td>-0.38</td>\n",
              "      <td>0.66</td>\n",
              "      <td>-0.71</td>\n",
              "      <td>-0.10</td>\n",
              "      <td>-0.77</td>\n",
              "      <td>0.21</td>\n",
              "      <td>-0.75</td>\n",
              "      <td>1.00</td>\n",
              "      <td>-0.49</td>\n",
              "      <td>-0.53</td>\n",
              "      <td>-0.23</td>\n",
              "      <td>0.29</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>RAD</th>\n",
              "      <td>NaN</td>\n",
              "      <td>0.63</td>\n",
              "      <td>-0.31</td>\n",
              "      <td>0.60</td>\n",
              "      <td>-0.01</td>\n",
              "      <td>0.61</td>\n",
              "      <td>-0.21</td>\n",
              "      <td>0.46</td>\n",
              "      <td>-0.49</td>\n",
              "      <td>1.00</td>\n",
              "      <td>0.91</td>\n",
              "      <td>0.46</td>\n",
              "      <td>-0.44</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>TAX</th>\n",
              "      <td>NaN</td>\n",
              "      <td>0.58</td>\n",
              "      <td>-0.31</td>\n",
              "      <td>0.72</td>\n",
              "      <td>-0.04</td>\n",
              "      <td>0.67</td>\n",
              "      <td>-0.29</td>\n",
              "      <td>0.51</td>\n",
              "      <td>-0.53</td>\n",
              "      <td>0.91</td>\n",
              "      <td>1.00</td>\n",
              "      <td>0.46</td>\n",
              "      <td>-0.44</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>PTRATIO</th>\n",
              "      <td>NaN</td>\n",
              "      <td>0.29</td>\n",
              "      <td>-0.39</td>\n",
              "      <td>0.38</td>\n",
              "      <td>-0.12</td>\n",
              "      <td>0.19</td>\n",
              "      <td>-0.36</td>\n",
              "      <td>0.26</td>\n",
              "      <td>-0.23</td>\n",
              "      <td>0.46</td>\n",
              "      <td>0.46</td>\n",
              "      <td>1.00</td>\n",
              "      <td>-0.18</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>B</th>\n",
              "      <td>NaN</td>\n",
              "      <td>-0.39</td>\n",
              "      <td>0.18</td>\n",
              "      <td>-0.36</td>\n",
              "      <td>0.05</td>\n",
              "      <td>-0.38</td>\n",
              "      <td>0.13</td>\n",
              "      <td>-0.27</td>\n",
              "      <td>0.29</td>\n",
              "      <td>-0.44</td>\n",
              "      <td>-0.44</td>\n",
              "      <td>-0.18</td>\n",
              "      <td>1.00</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "         const  CRIM    ZN  INDUS  CHAS  ...   DIS   RAD   TAX  PTRATIO     B\n",
              "const      NaN   NaN   NaN    NaN   NaN  ...   NaN   NaN   NaN      NaN   NaN\n",
              "CRIM       NaN  1.00 -0.20   0.41 -0.06  ... -0.38  0.63  0.58     0.29 -0.39\n",
              "ZN         NaN -0.20  1.00  -0.53 -0.04  ...  0.66 -0.31 -0.31    -0.39  0.18\n",
              "INDUS      NaN  0.41 -0.53   1.00  0.06  ... -0.71  0.60  0.72     0.38 -0.36\n",
              "CHAS       NaN -0.06 -0.04   0.06  1.00  ... -0.10 -0.01 -0.04    -0.12  0.05\n",
              "NOX        NaN  0.42 -0.52   0.76  0.09  ... -0.77  0.61  0.67     0.19 -0.38\n",
              "RM         NaN -0.22  0.31  -0.39  0.09  ...  0.21 -0.21 -0.29    -0.36  0.13\n",
              "AGE        NaN  0.35 -0.57   0.64  0.09  ... -0.75  0.46  0.51     0.26 -0.27\n",
              "DIS        NaN -0.38  0.66  -0.71 -0.10  ...  1.00 -0.49 -0.53    -0.23  0.29\n",
              "RAD        NaN  0.63 -0.31   0.60 -0.01  ... -0.49  1.00  0.91     0.46 -0.44\n",
              "TAX        NaN  0.58 -0.31   0.72 -0.04  ... -0.53  0.91  1.00     0.46 -0.44\n",
              "PTRATIO    NaN  0.29 -0.39   0.38 -0.12  ... -0.23  0.46  0.46     1.00 -0.18\n",
              "B          NaN -0.39  0.18  -0.36  0.05  ...  0.29 -0.44 -0.44    -0.18  1.00\n",
              "\n",
              "[13 rows x 13 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ElQoLcvkk3TU",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "edd006bf-a6b0-48b4-c6dc-f3536421bf5a"
      },
      "source": [
        "# Find index of feature columns with correlation greater than 0.75\n",
        "Pos_corr_var=np.where(corr_matrix > 0.75)\n",
        "Pos_corr_var=[(corr_matrix.columns[x],corr_matrix.columns[y]) for x,y in zip(*Pos_corr_var) if x!=y and x<y]\n",
        "print(\"Highly Positive Related (> 0.75) : \")\n",
        "print(Pos_corr_var)\n",
        "\n",
        "# Find index of feature columns with correlation less than - 0.75\n",
        "Neg_corr_var=np.where(corr_matrix < -0.75)\n",
        "Neg_corr_var=[(corr_matrix.columns[x],corr_matrix.columns[y]) for x,y in zip(*Neg_corr_var) if x!=y and x<y]\n",
        "print(\"Highly Negatively Related (< - 0.75) : \")\n",
        "print(Neg_corr_var)\n",
        "\n",
        "# Inference : \n",
        "# Remove 'INDUS' / 'NOX', 'RAD'/'TAX , 'NOX'/'DIS'\n",
        "# Removing NOX\n",
        "independent_variables = independent_variables.drop(['NOX'], axis = 1)\n",
        "\n",
        "# Now out of 'RAD' & 'TAX' which one to remove ? VIF"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Highly Positive Related (> 0.75) : \n",
            "[('INDUS', 'NOX'), ('RAD', 'TAX')]\n",
            "Highly Negatively Related (< - 0.75) : \n",
            "[('NOX', 'DIS')]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "b7LwqXFxk_L1",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "3c89c9a4-3b3d-4453-8458-19b184b1d9ae"
      },
      "source": [
        "# Function for VIF\n",
        " \n",
        "def calc_vif(X):\n",
        "  # Calculating VIF\n",
        "  vif = pd.DataFrame()\n",
        "  vif[\"variables\"] = X.columns\n",
        "  vif[\"VIF\"] = [variance_inflation_factor(X.values, i) for i in range(X.shape[1])]\n",
        "  return(vif)\n",
        " \n",
        "print(pd.DataFrame(calc_vif(independent_variables).sort_values(\"VIF\")))\n",
        " \n",
        "# VIF = 1, no correlation between the independent variable and the other variables\n",
        "# VIF exceeding 20 indicates high multicollinearity between this independent variable and the others\n",
        " \n",
        "# Inference : \n",
        "# Remove 'RM','TAX' and 'PTRATIO'\n",
        "# independent_variables = independent_variables.drop(['RM', 'TAX', 'PTRATIO'], axis = 1)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "   variables         VIF\n",
            "4       CHAS    1.069265\n",
            "11         B    1.306190\n",
            "5         RM    1.347445\n",
            "10   PTRATIO    1.602691\n",
            "1       CRIM    1.746058\n",
            "2         ZN    2.289884\n",
            "6        AGE    2.480534\n",
            "7        DIS    3.640058\n",
            "3      INDUS    3.677764\n",
            "8        RAD    7.317727\n",
            "9        TAX    8.957485\n",
            "0      const  344.083623\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "eQEr2okU-cFt",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 517
        },
        "outputId": "8b775754-42af-4594-ec5b-9e98ef73e7ea"
      },
      "source": [
        "# Outlier in the data (Column Wise Checking)\n",
        "\n",
        "# CRIM Column\n",
        "print(\"Quantile Distribution\")\n",
        "print(pd.DataFrame(independent_variables['CRIM'].quantile([0, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9,0.95 ,1])))\n",
        "sns.boxplot(data = independent_variables, orient = 'h', order = ['CRIM'])\n",
        "\n",
        "# Inference : \n",
        "# Capping to 95% Datapoints\n",
        "independent_variables['CRIM'] = np.where(independent_variables['CRIM'].round(1) > 15.0, 15.0, independent_variables['CRIM'])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Quantile Distribution\n",
            "           CRIM\n",
            "0.00   0.006320\n",
            "0.10   0.038195\n",
            "0.20   0.064170\n",
            "0.30   0.099245\n",
            "0.40   0.150380\n",
            "0.50   0.256510\n",
            "0.60   0.550070\n",
            "0.70   1.728440\n",
            "0.80   5.581070\n",
            "0.90  10.753000\n",
            "0.95  15.789150\n",
            "1.00  88.976200\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3ODwjtS9E3e9",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 535
        },
        "outputId": "af4c5a14-866f-4dbf-9b01-7db763eba5ef"
      },
      "source": [
        "# ZN Column\n",
        "\n",
        "print(\"Quantile Distribution\")\n",
        "print(pd.DataFrame(independent_variables['ZN'].quantile([0, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9,0.95 ,1])))\n",
        "sns.boxplot(data = independent_variables, orient = 'h', order = ['ZN'])\n",
        "\n",
        "# Inference : \n",
        "# Most Values (70%) data points are zero. So lets not do any capping !!"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Quantile Distribution\n",
            "         ZN\n",
            "0.00    0.0\n",
            "0.10    0.0\n",
            "0.20    0.0\n",
            "0.30    0.0\n",
            "0.40    0.0\n",
            "0.50    0.0\n",
            "0.60    0.0\n",
            "0.70    0.0\n",
            "0.80   20.0\n",
            "0.90   42.5\n",
            "0.95   80.0\n",
            "1.00  100.0\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7f3e41bbf6a0>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 18
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xFuNEV0GFNj7",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 535
        },
        "outputId": "60e35231-0d4c-4e3b-8751-d60b5f5ab272"
      },
      "source": [
        "# CRIM Column\n",
        "\n",
        "print(\"Quantile Distribution\")\n",
        "print(pd.DataFrame(independent_variables['INDUS'].quantile([0, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9,0.95 ,1])))\n",
        "sns.boxplot(data = independent_variables, orient = 'h', order = ['INDUS'])\n",
        "\n",
        "# Inference : NO OUTLIERS"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Quantile Distribution\n",
            "      INDUS\n",
            "0.00   0.46\n",
            "0.10   2.91\n",
            "0.20   4.39\n",
            "0.30   5.96\n",
            "0.40   7.38\n",
            "0.50   9.69\n",
            "0.60  12.83\n",
            "0.70  18.10\n",
            "0.80  18.10\n",
            "0.90  19.58\n",
            "0.95  21.89\n",
            "1.00  27.74\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7f3e41b99320>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 19
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SUOBOA8LGEFC",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 517
        },
        "outputId": "1a3b36d4-5eac-4e5c-a961-4b0579900d9e"
      },
      "source": [
        "# CHAS Column\n",
        "\n",
        "print(\"Quantile Distribution\")\n",
        "print(pd.DataFrame(independent_variables['CHAS'].quantile([0, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9,0.95 ,1])))\n",
        "sns.boxplot(data = independent_variables, orient = 'h', order = ['CHAS'])\n",
        "\n",
        "# Inference : \n",
        "# 90% of Data is 0. So lets drop the variable\n",
        "independent_variables = independent_variables.drop(['CHAS'], axis = 1)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Quantile Distribution\n",
            "      CHAS\n",
            "0.00   0.0\n",
            "0.10   0.0\n",
            "0.20   0.0\n",
            "0.30   0.0\n",
            "0.40   0.0\n",
            "0.50   0.0\n",
            "0.60   0.0\n",
            "0.70   0.0\n",
            "0.80   0.0\n",
            "0.90   0.0\n",
            "0.95   1.0\n",
            "1.00   1.0\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cFG10BSrGdbk",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 535
        },
        "outputId": "a57365ae-f977-43d4-945f-ce10120001e0"
      },
      "source": [
        "# AGE Column \n",
        "\n",
        "print(\"Quantile Distribution\")\n",
        "print(pd.DataFrame(independent_variables['AGE'].quantile([0, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9,0.95 ,1])))\n",
        "sns.boxplot(data = independent_variables, orient = 'h', order = ['AGE'])\n",
        "\n",
        "# Inference : NO OUTLIERS"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Quantile Distribution\n",
            "         AGE\n",
            "0.00    2.90\n",
            "0.10   26.95\n",
            "0.20   37.80\n",
            "0.30   52.40\n",
            "0.40   65.40\n",
            "0.50   77.50\n",
            "0.60   85.90\n",
            "0.70   91.80\n",
            "0.80   95.60\n",
            "0.90   98.80\n",
            "0.95  100.00\n",
            "1.00  100.00\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7f3e41ae6828>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 21
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXgAAAD4CAYAAADmWv3KAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAKj0lEQVR4nO3dX4ymZ1nH8d/VXYFdCJa1ZKPThgWnsakSpTZYtVGpHkAltgdNtGrEhIQTXdfGxGA8oHrWhIjLxpA0oFZiQC1EKhIilBpDjNXd2kCllb7876SlWyqFuNVSuD143zXDdqe7O52XZ/aazyeZzDzPvNO57rlnvztzz7StMUYA6OeCqQcAYDkEHqApgQdoSuABmhJ4gKZ2Tz3AehdddNE4cODA1GMAnFeOHTv22Bjjpafe31aBP3DgQI4ePTr1GADnlar6wunuO6IBaErgAZoSeICmBB6gKYEHaErgAZoSeICmBB6gKYEHaErgAZoSeICmBB6gKYEHaErgAZoSeICmBB6gKYEHaErgAZoSeICmttX/kxW2wpEjRzKbzaYeg1Osra0lSVZWViaeZOusrq7m4MGDU4+xIYGnndlslnvvuz/f3Ltv6lFYZ9eJJ5Ikj/xvj+zsOvH41COcUY+PNJzim3v35cnLrp16DNbZ88CHkqTNvpxcz3bmDB6gKYEHaErgAZoSeICmBB6gKYEHaErgAZoSeICmBB6gKYEHaErgAZoSeICmBB6gKYEHaErgAZoSeICmBB6gKYEHaErgAZoSeICmBB6gKYEHaErgAZoSeICmBB6gKYEHaErgAZoSeICmBB6gKYEHaErgAZoSeICmBB6gKYEHaErgAZoSeICmBB6gKYEHaErgAZoSeICmBB6gqRaBP3LkSI4cOTL1GACbsqyG7d7yf+IEZrPZ1CMAbNqyGtbiK3gAnkngAZoSeICmBB6gKYEHaErgAZoSeICmBB6gKYEHaErgAZoSeICmBB6gKYEHaErgAZoSeICmBB6gKYEHaErgAZoSeICmBB6gKYEHaErgAZoSeICmBB6gKYEHaErgAZoSeICmBB6gKYEHaErgAZoSeICmBB6gKYEHaErgAZoSeICmBB6gKYEHaErgAZoSeICmzirwVXV9VY2qumzdvVdX1T9W1YNVdU9V/X1VvXLxupuraq2q7l33dOGyFgHAM+0+y8fdmOTji+dvqar9Sf46yS+PMf45Sarq6iTfn+STi7d52xjjrVs8LwBn6YyBr6oXJbk6yWuS/F2StyT5zSS3nYx7kowxPr6sIc9kbW0tTz75ZA4dOjTVCGwjs9ksFzw1ph6D5i74n69lNvv6lnRnNptlz549WzDVtzubI5rrknx4jPHpJF+pqh9N8oNJ7jnD29207njmro0eVFVvqqqjVXX0+PHjZz85AM/qbI5obkxyePHyexfX36aq7k7y4iT/MMY4+dfZWR3RjDFuTXJrklx55ZWb+rJrZWUlSXL48OEzPJKd4NChQzn22S9PPQbNfesFL87qK/ZvSXeWdfrwrIGvqn1JrknyyqoaSXYlGUluS3JFkg8kyRjjx6rqhiSvX8qUAJyzMx3R3JDk3WOMl40xDowxLknyuSQfSfLrVfUT6x67d1lDAnDuznREc2OSW065977F/V9McktVrSR5NMljSf5w3eNuqqpfXXd9/Rjj889tXADO1rMGfozxmtPce/u6y5/e4O1uTnLzcxkMgOfGv8kK0JTAAzQl8ABNCTxAUwIP0JTAAzQl8ABNCTxAUwIP0JTAAzQl8ABNCTxAUwIP0JTAAzQl8ABNCTxAUwIP0JTAAzQl8ABNCTxAUwIP0JTAAzQl8ABNCTxAUwIP0JTAAzQl8ABNCTxAUwIP0JTAAzQl8ABNCTxAUwIP0JTAAzQl8ABNCTxAUwIP0JTAAzS1e+oBtsLq6urUIwBs2rIa1iLwBw8enHoEgE1bVsMc0QA0JfAATQk8QFMCD9CUwAM0JfAATQk8QFMCD9CUwAM0JfAATQk8QFMCD9CUwAM0JfAATQk8QFMCD9CUwAM0JfAATQk8QFMCD9CUwAM0JfAATQk8QFMCD9CUwAM0JfAATQk8QFMCD9CUwAM0JfAATQk8QFMCD9CUwAM0JfAATQk8QFMCD9CUwAM0JfAATQk8QFMCD9DU7qkHgGXYdeLx7HngQ1OPwTq7TnwlSdrsy64TjyfZP/UYz0rgaWd1dXXqETiNtbWnkyQrK9s7imdv/7b/XBN42jl48ODUI8C24AweoCmBB2hK4AGaEniApgQeoCmBB2hK4AGaEniApgQeoCmBB2hK4AGaEniApgQeoCmBB2hK4AGaEniApgQeoCmBB2hK4AGaEniApmqMMfUM/6+qjif5wmledVGSx77D42wXO3XtO3XdibVb+7l72Rjjpafe3FaB30hVHR1jXDn1HFPYqWvfqetOrN3at44jGoCmBB6gqfMl8LdOPcCEdurad+q6E2vfqbZ87efFGTwA5+58+QoegHMk8ABNbfvAV9Vrq+o/q2pWVW+eep5lqapLququqvpUVf1HVR1a3N9XVR+pqgcXz18y9azLUlW7qurfq+qDi+uXV9Xdi73/q6p63tQzLkNVXVhVt1fVA1V1f1X9+E7Y96q6afG5fl9VvaeqXtB1z6vqT6vq0aq6b9290+5xzb198TH4RFVdsdn3u60DX1W7kvxJktcluTzJjVV1+bRTLc3TSX5njHF5kquS/MZirW9OcucY49Ikdy6uuzqU5P5117ckedsYYzXJfyV54yRTLd/hJB8eY1yW5Icz/xi03veqWknyW0muHGP8UJJdSX4pfff8z5O89pR7G+3x65Jcunh6U5J3bPadbuvAJ3l1ktkY47NjjKeSvDfJdRPPtBRjjIfHGPcsXv565n/IVzJf722Lh92W5PppJlyuqro4yc8neefiupJck+T2xUNarr2qvjvJTyV5V5KMMZ4aY3w1O2PfdyfZU1W7k+xN8nCa7vkY45+SPH7K7Y32+LokfzHm/iXJhVX1vZt5v9s98CtJvrTu+qHFvdaq6kCSVyW5O8n+McbDi1c9kmT/RGMt2x8n+d0k31pcf0+Sr44xnl5cd937lyc5nuTPFsdT76yqF6b5vo8x1pK8NckXMw/7E0mOZWfs+Ukb7fGWdW+7B37HqaoXJXlfkt8eY3xt/evG/Hda2/1ea1W9PsmjY4xjU88ygd1JrkjyjjHGq5L8d045jum474vz5usy/wvu+5K8MM88wtgxlrXH2z3wa0kuWXd98eJeS1X1XZnH/S/HGO9f3P7yyW/PFs8fnWq+JfrJJL9QVZ/P/BjumszPpS9cfPue9N37h5I8NMa4e3F9e+bB777vP5fkc2OM42OMbyR5f+afBzthz0/aaI+3rHvbPfD/luTSxU/Wn5f5D2HumHimpVicOb8ryf1jjD9a96o7krxh8fIbknzgOz3bso0xfm+McfEY40Dme/yxMcavJLkryQ2Lh3Vd+yNJvlRVP7C49bNJPpX++/7FJFdV1d7F5/7Jdbff83U22uM7kvza4rdprkryxLqjnHMzxtjWT0muTfLpJJ9J8vtTz7PEdV6d+bdon0hy7+Lp2szPou9M8mCSjybZN/WsS/44/EySDy5efkWSf00yS/I3SZ4/9XxLWvOPJDm62Pu/TfKSnbDvSf4gyQNJ7kvy7iTP77rnSd6T+c8avpH5d21v3GiPk1Tmvz34mSSfzPw3jTb1fv2nCgCa2u5HNABsksADNCXwAE0JPEBTAg/QlMADNCXwAE39H036AthGBWWDAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "i-PA04qwGcxi",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 517
        },
        "outputId": "a4840362-6f8e-4364-98a4-cc5237599d01"
      },
      "source": [
        "# DIS Column\n",
        "\n",
        "print(\"Quantile Distribution\")\n",
        "print(pd.DataFrame(independent_variables['DIS'].quantile([0, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9,0.95 ,1])))\n",
        "sns.boxplot(data = independent_variables, orient = 'h', order = ['DIS'])\n",
        "\n",
        "# Inference : Capping to 95% Datapoints\n",
        "independent_variables['DIS'] = np.where(independent_variables['DIS'].round(1) > 7.0, 7.0, independent_variables['DIS'])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Quantile Distribution\n",
            "           DIS\n",
            "0.00   1.12960\n",
            "0.10   1.62830\n",
            "0.20   1.95120\n",
            "0.30   2.25965\n",
            "0.40   2.64030\n",
            "0.50   3.20745\n",
            "0.60   3.87500\n",
            "0.70   4.54040\n",
            "0.80   5.61500\n",
            "0.90   6.81660\n",
            "0.95   7.82780\n",
            "1.00  12.12650\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1ePvy_enGbtT",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 535
        },
        "outputId": "2787654f-92b0-4d5e-8f3d-fc9c5884f431"
      },
      "source": [
        "# RAD Column \n",
        "\n",
        "print(\"Quantile Distribution\")\n",
        "print(pd.DataFrame(independent_variables['RAD'].quantile([0, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9,0.95 ,1])))\n",
        "sns.boxplot(data = independent_variables, orient = 'h', order = ['RAD'])\n",
        "\n",
        "# Inference : NO OUTLIERS"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Quantile Distribution\n",
            "       RAD\n",
            "0.00   1.0\n",
            "0.10   3.0\n",
            "0.20   4.0\n",
            "0.30   4.0\n",
            "0.40   5.0\n",
            "0.50   5.0\n",
            "0.60   5.0\n",
            "0.70   8.0\n",
            "0.80  24.0\n",
            "0.90  24.0\n",
            "0.95  24.0\n",
            "1.00  24.0\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7f3e41a2a4e0>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 23
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KHHA9PwnGa9j",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "8f2ceb2c-9c78-4524-b800-6782a099ddad"
      },
      "source": [
        "# B Column\n",
        "\n",
        "print(\"Quantile Distribution\")\n",
        "print(pd.DataFrame(independent_variables['B'].quantile([0, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9,0.95 ,1])))\n",
        "sns.boxplot(data = independent_variables, orient = 'h', order = ['B'])\n",
        "\n",
        "# Inference : Capping to first 10% Datapoints\n",
        "independent_variables['B'] = np.where(independent_variables['B'].round(1) < 290.0, 290.0, independent_variables['B'])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Quantile Distribution\n",
            "            B\n",
            "0.00    0.320\n",
            "0.10  290.270\n",
            "0.20  364.310\n",
            "0.30  378.665\n",
            "0.40  387.970\n",
            "0.50  391.440\n",
            "0.60  393.530\n",
            "0.70  395.465\n",
            "0.80  396.900\n",
            "0.90  396.900\n",
            "0.95  396.900\n",
            "1.00  396.900\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "p19JMawzBblD",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "cc3fd34f-7919-4515-c893-96abbec23d5f"
      },
      "source": [
        "# Skewness in the Data\n",
        "\n",
        "# If Skewness = 0, data is perfectly symmetrical\n",
        "# If Skewness is less than -1 or greater +1, distribution is highly skewed\n",
        "# If Skewness is between -1 and -0.5 or between 0.5 and 1, distribution is moderately skewed\n",
        "# If Skewness is between -0.5 and 0.5 then it is approximately symmetric  \n",
        "\n",
        "pd.DataFrame(independent_variables.skew().round(2),columns = [\"Skewness\"])\n",
        "\n",
        "# Inference : Positive Skew - 'CRIM','ZN' & Negative Skew - 'B'"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "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>Skewness</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>const</th>\n",
              "      <td>0.00</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>CRIM</th>\n",
              "      <td>1.71</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>ZN</th>\n",
              "      <td>2.23</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>INDUS</th>\n",
              "      <td>0.30</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>RM</th>\n",
              "      <td>0.40</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>AGE</th>\n",
              "      <td>-0.60</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>DIS</th>\n",
              "      <td>0.55</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>RAD</th>\n",
              "      <td>1.00</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>TAX</th>\n",
              "      <td>0.67</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>PTRATIO</th>\n",
              "      <td>-0.80</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>B</th>\n",
              "      <td>-1.76</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "         Skewness\n",
              "const        0.00\n",
              "CRIM         1.71\n",
              "ZN           2.23\n",
              "INDUS        0.30\n",
              "RM           0.40\n",
              "AGE         -0.60\n",
              "DIS          0.55\n",
              "RAD          1.00\n",
              "TAX          0.67\n",
              "PTRATIO     -0.80\n",
              "B           -1.76"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 25
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4T8AURorE2N9",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "23a9df45-1da4-437e-f0af-ca418498803f"
      },
      "source": [
        "# Remove Skewness one by one (Column Wise)\n",
        "\n",
        "# CRIM Variable\n",
        "# Log Transformation\n",
        "CRIM_Log = np.log(independent_variables['CRIM'])\n",
        "\n",
        "# Square Root Transformation\n",
        "CRIM_sqrt = np.sqrt(independent_variables['CRIM'])\n",
        "\n",
        "# Cube Root Transformation\n",
        "CRIM_cbrt = np.cbrt(independent_variables['CRIM'])\n",
        "\n",
        "# Box Cox Transformation\n",
        "from scipy import stats\n",
        "CRIM_BoxCox = stats.boxcox(independent_variables['CRIM'])[0]\n",
        "Variable = pd.Series(CRIM_BoxCox)\n",
        "\n",
        "print(\"Comparing all Skewness values\")\n",
        "print(pd.DataFrame([independent_variables['CRIM'].skew(),CRIM_Log.skew(),CRIM_sqrt.skew(),CRIM_cbrt.skew(),Variable.skew()],\n",
        "             index=['Normal','Log','Square Root','Cube Root','Box Cox'],\n",
        "             columns = [\"Skewness\"]))\n",
        "\n",
        "# Inference : Best transformations is Log Transformation (Near to 0)\n",
        "\n",
        "independent_variables['CRIM_Log'] = CRIM_Log"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Comparing all Skewness values\n",
            "             Skewness\n",
            "Normal       1.705324\n",
            "Log          0.321681\n",
            "Square Root  1.168043\n",
            "Cube Root    0.935891\n",
            "Box Cox      0.083939\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6Cudu17XLlWV",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "60232920-de4e-4c8d-d10a-4cf4f15f74ea"
      },
      "source": [
        "# ZN Variable\n",
        "\n",
        "# Log Transformation\n",
        "ZN_Log = np.log(independent_variables['ZN'])\n",
        "\n",
        "# Square Root Transformation\n",
        "ZN_sqrt = np.sqrt(independent_variables['ZN'])\n",
        "\n",
        "# Cube Root Transformation\n",
        "ZN_cbrt = np.cbrt(independent_variables['ZN'])\n",
        "\n",
        "print(\"Comparing all Skewness values\")\n",
        "print(pd.DataFrame([independent_variables['ZN'].skew(),ZN_Log.skew(),ZN_sqrt.skew(),ZN_cbrt.skew()],\n",
        "             index=['Normal','Log','Square Root','Cube Root'],\n",
        "             columns = [\"Skewness\"]))\n",
        "\n",
        "# Inference : Best transformations is Cube Transformation (Near to 0)\n",
        "\n",
        "independent_variables['ZN_Cbrt'] = ZN_cbrt"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Comparing all Skewness values\n",
            "             Skewness\n",
            "Normal       2.225666\n",
            "Log               NaN\n",
            "Square Root  1.476293\n",
            "Cube Root    1.262563\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/pandas/core/series.py:726: RuntimeWarning: divide by zero encountered in log\n",
            "  result = getattr(ufunc, method)(*inputs, **kwargs)\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tmRPbMMLMGGB",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "40d9cc51-dfd4-41ce-e1d3-e03b346ff648"
      },
      "source": [
        "# B Variable\n",
        "\n",
        "# Log Transformation\n",
        "B_Log = np.log(independent_variables['B'])\n",
        "\n",
        "# Square Root Transformation\n",
        "B_sqrt = np.sqrt(independent_variables['B'])\n",
        "\n",
        "# Cube Root Transformation\n",
        "B_cbrt = np.cbrt(independent_variables['B'])\n",
        "\n",
        "# Box Cox Transformation\n",
        "from scipy import stats\n",
        "B_BoxCox = stats.boxcox(independent_variables['B'])[0]\n",
        "Variable = pd.Series(B_BoxCox)\n",
        "\n",
        "print(\"Comparing all Skewness values\")\n",
        "print(pd.DataFrame([independent_variables['B'].skew(),B_Log.skew(),B_sqrt.skew(),B_cbrt.skew(),Variable.skew()],\n",
        "             index=['Normal','Log','Square Root','Cube Root','Box Cox'],\n",
        "             columns = [\"Skewness\"]))\n",
        "\n",
        "# Inference : Best transformations is Box Cox Transformation (Approx to zero)\n",
        "# But avoiding Box Cox because it give values e raise to, second best is Square Root Transformation\n",
        "independent_variables['B_Sqrt'] = B_sqrt"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Comparing all Skewness values\n",
            "             Skewness\n",
            "Normal      -1.760681\n",
            "Log         -1.837693\n",
            "Square Root -1.799513\n",
            "Cube Root   -1.812319\n",
            "Box Cox     -0.867610\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "B0LnS5sbOB8p",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "36ce65b7-92c4-40b3-af95-e542d6927e53"
      },
      "source": [
        "# Removing variables (raw variables) which are transformed and keep one set also (no removing)\n",
        "independent_variables_v2 = independent_variables\n",
        "\n",
        "# Removing variables\n",
        "independent_variables = independent_variables.drop(['CRIM', 'ZN', 'B'], axis = 1)\n",
        "\n",
        "# Round to 2 decimal point\n",
        "independent_variables = independent_variables.round(2)\n",
        "independent_variables.head()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "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>const</th>\n",
              "      <th>INDUS</th>\n",
              "      <th>RM</th>\n",
              "      <th>AGE</th>\n",
              "      <th>DIS</th>\n",
              "      <th>RAD</th>\n",
              "      <th>TAX</th>\n",
              "      <th>PTRATIO</th>\n",
              "      <th>CRIM_Log</th>\n",
              "      <th>ZN_Cbrt</th>\n",
              "      <th>B_Sqrt</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>1.0</td>\n",
              "      <td>2.31</td>\n",
              "      <td>6.58</td>\n",
              "      <td>65.2</td>\n",
              "      <td>4.09</td>\n",
              "      <td>1.0</td>\n",
              "      <td>296.0</td>\n",
              "      <td>15.3</td>\n",
              "      <td>-5.06</td>\n",
              "      <td>2.62</td>\n",
              "      <td>19.92</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1.0</td>\n",
              "      <td>7.07</td>\n",
              "      <td>6.42</td>\n",
              "      <td>78.9</td>\n",
              "      <td>4.97</td>\n",
              "      <td>2.0</td>\n",
              "      <td>242.0</td>\n",
              "      <td>17.8</td>\n",
              "      <td>-3.60</td>\n",
              "      <td>0.00</td>\n",
              "      <td>19.92</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>1.0</td>\n",
              "      <td>7.07</td>\n",
              "      <td>7.18</td>\n",
              "      <td>61.1</td>\n",
              "      <td>4.97</td>\n",
              "      <td>2.0</td>\n",
              "      <td>242.0</td>\n",
              "      <td>17.8</td>\n",
              "      <td>-3.60</td>\n",
              "      <td>0.00</td>\n",
              "      <td>19.82</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>1.0</td>\n",
              "      <td>2.18</td>\n",
              "      <td>7.00</td>\n",
              "      <td>45.8</td>\n",
              "      <td>6.06</td>\n",
              "      <td>3.0</td>\n",
              "      <td>222.0</td>\n",
              "      <td>18.7</td>\n",
              "      <td>-3.43</td>\n",
              "      <td>0.00</td>\n",
              "      <td>19.87</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>1.0</td>\n",
              "      <td>2.18</td>\n",
              "      <td>7.15</td>\n",
              "      <td>54.2</td>\n",
              "      <td>6.06</td>\n",
              "      <td>3.0</td>\n",
              "      <td>222.0</td>\n",
              "      <td>18.7</td>\n",
              "      <td>-2.67</td>\n",
              "      <td>0.00</td>\n",
              "      <td>19.92</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "   const  INDUS    RM   AGE   DIS  ...    TAX  PTRATIO  CRIM_Log  ZN_Cbrt  B_Sqrt\n",
              "0    1.0   2.31  6.58  65.2  4.09  ...  296.0     15.3     -5.06     2.62   19.92\n",
              "1    1.0   7.07  6.42  78.9  4.97  ...  242.0     17.8     -3.60     0.00   19.92\n",
              "2    1.0   7.07  7.18  61.1  4.97  ...  242.0     17.8     -3.60     0.00   19.82\n",
              "3    1.0   2.18  7.00  45.8  6.06  ...  222.0     18.7     -3.43     0.00   19.87\n",
              "4    1.0   2.18  7.15  54.2  6.06  ...  222.0     18.7     -2.67     0.00   19.92\n",
              "\n",
              "[5 rows x 11 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 29
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2TQWyaWWQxhg",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "2d3ad0e7-987e-4658-88a9-65dc1582a1a9"
      },
      "source": [
        "# Now Train the model\n",
        "\n",
        "# Splitting Data into 70:30\n",
        "data_train, data_test, house_price_train, house_price_test = train_test_split(independent_variables,dependent_variable,train_size = 0.7,random_state = 21)\n",
        "\n",
        "# Fitting the model\n",
        "Model_2 = OLS(house_price_train,data_train).fit()\n",
        "Model_2.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<table class=\"simpletable\">\n",
              "<caption>OLS Regression Results</caption>\n",
              "<tr>\n",
              "  <th>Dep. Variable:</th>       <td>HOUSEPRICE</td>    <th>  R-squared:         </th> <td>   0.639</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>Model:</th>                   <td>OLS</td>       <th>  Adj. R-squared:    </th> <td>   0.628</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>Method:</th>             <td>Least Squares</td>  <th>  F-statistic:       </th> <td>   60.65</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>Date:</th>             <td>Tue, 26 Jan 2021</td> <th>  Prob (F-statistic):</th> <td>9.29e-70</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>Time:</th>                 <td>13:36:02</td>     <th>  Log-Likelihood:    </th> <td> -1096.8</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>No. Observations:</th>      <td>   354</td>      <th>  AIC:               </th> <td>   2216.</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>Df Residuals:</th>          <td>   343</td>      <th>  BIC:               </th> <td>   2258.</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>Df Model:</th>              <td>    10</td>      <th>                     </th>     <td> </td>   \n",
              "</tr>\n",
              "<tr>\n",
              "  <th>Covariance Type:</th>      <td>nonrobust</td>    <th>                     </th>     <td> </td>   \n",
              "</tr>\n",
              "</table>\n",
              "<table class=\"simpletable\">\n",
              "<tr>\n",
              "      <td></td>        <th>coef</th>     <th>std err</th>      <th>t</th>      <th>P>|t|</th>  <th>[0.025</th>    <th>0.975]</th>  \n",
              "</tr>\n",
              "<tr>\n",
              "  <th>const</th>    <td>   -7.1318</td> <td>    8.474</td> <td>   -0.842</td> <td> 0.401</td> <td>  -23.799</td> <td>    9.535</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>INDUS</th>    <td>   -0.0841</td> <td>    0.080</td> <td>   -1.058</td> <td> 0.291</td> <td>   -0.240</td> <td>    0.072</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>RM</th>       <td>    5.9865</td> <td>    0.461</td> <td>   12.981</td> <td> 0.000</td> <td>    5.079</td> <td>    6.894</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>AGE</th>      <td>   -0.0709</td> <td>    0.017</td> <td>   -4.126</td> <td> 0.000</td> <td>   -0.105</td> <td>   -0.037</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>DIS</th>      <td>   -1.5377</td> <td>    0.322</td> <td>   -4.768</td> <td> 0.000</td> <td>   -2.172</td> <td>   -0.903</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>RAD</th>      <td>    0.2449</td> <td>    0.105</td> <td>    2.343</td> <td> 0.020</td> <td>    0.039</td> <td>    0.450</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>TAX</th>      <td>   -0.0149</td> <td>    0.005</td> <td>   -3.054</td> <td> 0.002</td> <td>   -0.025</td> <td>   -0.005</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>PTRATIO</th>  <td>   -1.0319</td> <td>    0.183</td> <td>   -5.641</td> <td> 0.000</td> <td>   -1.392</td> <td>   -0.672</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>CRIM_Log</th> <td>   -0.4998</td> <td>    0.381</td> <td>   -1.313</td> <td> 0.190</td> <td>   -1.248</td> <td>    0.249</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>ZN_Cbrt</th>  <td>    0.2892</td> <td>    0.308</td> <td>    0.940</td> <td> 0.348</td> <td>   -0.316</td> <td>    0.895</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>B_Sqrt</th>   <td>    1.3286</td> <td>    0.372</td> <td>    3.572</td> <td> 0.000</td> <td>    0.597</td> <td>    2.060</td>\n",
              "</tr>\n",
              "</table>\n",
              "<table class=\"simpletable\">\n",
              "<tr>\n",
              "  <th>Omnibus:</th>       <td>202.916</td> <th>  Durbin-Watson:     </th> <td>   2.107</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>Prob(Omnibus):</th> <td> 0.000</td>  <th>  Jarque-Bera (JB):  </th> <td>2161.246</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>Skew:</th>          <td> 2.191</td>  <th>  Prob(JB):          </th> <td>    0.00</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>Kurtosis:</th>      <td>14.283</td>  <th>  Cond. No.          </th> <td>1.32e+04</td>\n",
              "</tr>\n",
              "</table><br/><br/>Warnings:<br/>[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.<br/>[2] The condition number is large, 1.32e+04. This might indicate that there are<br/>strong multicollinearity or other numerical problems."
            ],
            "text/plain": [
              "<class 'statsmodels.iolib.summary.Summary'>\n",
              "\"\"\"\n",
              "                            OLS Regression Results                            \n",
              "==============================================================================\n",
              "Dep. Variable:             HOUSEPRICE   R-squared:                       0.639\n",
              "Model:                            OLS   Adj. R-squared:                  0.628\n",
              "Method:                 Least Squares   F-statistic:                     60.65\n",
              "Date:                Tue, 26 Jan 2021   Prob (F-statistic):           9.29e-70\n",
              "Time:                        13:36:02   Log-Likelihood:                -1096.8\n",
              "No. Observations:                 354   AIC:                             2216.\n",
              "Df Residuals:                     343   BIC:                             2258.\n",
              "Df Model:                          10                                         \n",
              "Covariance Type:            nonrobust                                         \n",
              "==============================================================================\n",
              "                 coef    std err          t      P>|t|      [0.025      0.975]\n",
              "------------------------------------------------------------------------------\n",
              "const         -7.1318      8.474     -0.842      0.401     -23.799       9.535\n",
              "INDUS         -0.0841      0.080     -1.058      0.291      -0.240       0.072\n",
              "RM             5.9865      0.461     12.981      0.000       5.079       6.894\n",
              "AGE           -0.0709      0.017     -4.126      0.000      -0.105      -0.037\n",
              "DIS           -1.5377      0.322     -4.768      0.000      -2.172      -0.903\n",
              "RAD            0.2449      0.105      2.343      0.020       0.039       0.450\n",
              "TAX           -0.0149      0.005     -3.054      0.002      -0.025      -0.005\n",
              "PTRATIO       -1.0319      0.183     -5.641      0.000      -1.392      -0.672\n",
              "CRIM_Log      -0.4998      0.381     -1.313      0.190      -1.248       0.249\n",
              "ZN_Cbrt        0.2892      0.308      0.940      0.348      -0.316       0.895\n",
              "B_Sqrt         1.3286      0.372      3.572      0.000       0.597       2.060\n",
              "==============================================================================\n",
              "Omnibus:                      202.916   Durbin-Watson:                   2.107\n",
              "Prob(Omnibus):                  0.000   Jarque-Bera (JB):             2161.246\n",
              "Skew:                           2.191   Prob(JB):                         0.00\n",
              "Kurtosis:                      14.283   Cond. No.                     1.32e+04\n",
              "==============================================================================\n",
              "\n",
              "Warnings:\n",
              "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
              "[2] The condition number is large, 1.32e+04. This might indicate that there are\n",
              "strong multicollinearity or other numerical problems.\n",
              "\"\"\""
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 30
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TrEUVnQklh1A",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "ea5c2f45-a2a6-498e-e157-3391f6131fd4"
      },
      "source": [
        "# Model Properties\n",
        "\n",
        "# Normality of the residuals\n",
        "# Jarque-Bera test:\n",
        "name = ['Jarque-Bera', 'Chi^2 two-tail prob.', 'Skew', 'Kurtosis']\n",
        "test = sms.jarque_bera(Model_2.resid)\n",
        "base_model_properties_2 = lzip(name, test)\n",
        "print(\"MODEL 2 PARAMETERS\")\n",
        "print(\"\\n\")\n",
        "print(\"Normality of the residuals\")\n",
        "print(\"Jarque-Bera test : \")\n",
        "print(\"JB, ideal is close to 0\")\n",
        "print(\"Probability(JB), less than 0.05\")\n",
        "print(\"Skewness, ideal is -1 to 1\")\n",
        "print(\"Kurtosis, ideal value is 3\")\n",
        "print(pd.DataFrame(base_model_properties_2,columns = ['Test','Values']).round(2))\n",
        "\n",
        "# Omni test:\n",
        "print(\"Omni Test : \")\n",
        "print(\"Omni, ideal is close to 0\")\n",
        "print(\"Probability(Omni), less than 0.05\")\n",
        "name = ['Chi^2', 'Two-tail probability']\n",
        "test = sms.omni_normtest(Model_2.resid)\n",
        "print(pd.DataFrame(lzip(name, test),columns = ['Test','Values']).round(2))\n",
        "\n",
        "# Multicollinearity\n",
        "print(\"\\n\")\n",
        "print(\"Multicollinearity\")\n",
        "print(\"Ideal Condition Number will be less than 30\")\n",
        "print(\"Condition number:\",np.linalg.cond(Model_2.model.exog).round(2))\n",
        "\n",
        "# Adjusted R Square\n",
        "print(\"\\n\")\n",
        "print(\"Model Performance\")\n",
        "print(\"Adjusted R Square, ideal is close to 1\")\n",
        "print(\"Adjusted R Square : \",Model_2.rsquared_adj.round(2))\n",
        "\n",
        "# F Statistics\n",
        "print(\"\\n\")\n",
        "print(\"Significance of Independent Variables\")\n",
        "print(\"Probability(F Stat), less than 0.05 suggests independent variables are important\")\n",
        "print(\"F Statistics : \",Model_2.fvalue.round(2))\n",
        "print(\"Probability of F Statistics : \",Model_2.f_pvalue.round(2))\n",
        "\n",
        "# Durbin Watson Test\n",
        "print(\"\\n\")\n",
        "print(\"Homoscedasticity of Error\")\n",
        "print(\"Durbin Watson, ideal is between 1 to 2\")\n",
        "print(\"Durbin Watson Value : 2.1\")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "MODEL 2 PARAMETERS\n",
            "\n",
            "\n",
            "Normality of the residuals\n",
            "Jarque-Bera test : \n",
            "JB, ideal is close to 0\n",
            "Probability(JB), less than 0.05\n",
            "Skewness, ideal is -1 to 1\n",
            "Kurtosis, ideal value is 3\n",
            "                   Test   Values\n",
            "0           Jarque-Bera  2161.25\n",
            "1  Chi^2 two-tail prob.     0.00\n",
            "2                  Skew     2.19\n",
            "3              Kurtosis    14.28\n",
            "Omni Test : \n",
            "Omni, ideal is close to 0\n",
            "Probability(Omni), less than 0.05\n",
            "                   Test  Values\n",
            "0                 Chi^2  202.92\n",
            "1  Two-tail probability    0.00\n",
            "\n",
            "\n",
            "Multicollinearity\n",
            "Ideal Condition Number will be less than 30\n",
            "Condition number: 13155.98\n",
            "\n",
            "\n",
            "Model Performance\n",
            "Adjusted R Square, ideal is close to 1\n",
            "Adjusted R Square :  0.63\n",
            "\n",
            "\n",
            "Significance of Independent Variables\n",
            "Probability(F Stat), less than 0.05 suggests independent variables are important\n",
            "F Statistics :  60.65\n",
            "Probability of F Statistics :  0.0\n",
            "\n",
            "\n",
            "Homoscedasticity of Error\n",
            "Durbin Watson, ideal is between 1 to 2\n",
            "Durbin Watson Value : 2.1\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sQ1Om5Y1RHIO",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "05310157-631e-498f-f12b-cb97128b64a4"
      },
      "source": [
        "# Prediction of Model 2 \n",
        "\n",
        "Prediction_2 = Model_2.predict(data_test)\n",
        "print(\"Mean Absolute Error :\",metrics.mean_absolute_error(house_price_test,Prediction_2).round(2))\n",
        "print(\"Mean Squared Error :\",metrics.mean_squared_error(house_price_test,Prediction_2).round(2))\n",
        "print(\"Root Mean Squared Error :\",metrics.mean_squared_error(house_price_test,Prediction_2).round(2))\n",
        "print(\"Mean Absolute Percentage Error :\",mape(house_price_test,Prediction_2).round(2))\n",
        "\n",
        "# Inference : \n",
        "# Model 1 MAPE = 49.7% decreased to 48.3% in Model 2. Good Progress !!"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Mean Absolute Error : 3.9\n",
            "Mean Squared Error : 34.54\n",
            "Root Mean Squared Error : 34.54\n",
            "Mean Absolute Percentage Error : 50.5\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "W1vLWsZVU6s2",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "d369d1f6-3031-495d-b9c1-e1064e98a47b"
      },
      "source": [
        "# Model 3\n",
        "\n",
        "# Now lets drop variable based on P Values\n",
        "independent_variables = independent_variables.drop(['RAD', 'CRIM_Log'], axis = 1)\n",
        "\n",
        "# Splitting Data into 70:30\n",
        "data_train, data_test, house_price_train, house_price_test = train_test_split(independent_variables,dependent_variable,train_size = 0.7, random_state = 21)\n",
        "\n",
        "# Fitting the model\n",
        "Model_3 = OLS(house_price_train,data_train).fit()\n",
        "Model_3.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<table class=\"simpletable\">\n",
              "<caption>OLS Regression Results</caption>\n",
              "<tr>\n",
              "  <th>Dep. Variable:</th>       <td>HOUSEPRICE</td>    <th>  R-squared:         </th> <td>   0.633</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>Model:</th>                   <td>OLS</td>       <th>  Adj. R-squared:    </th> <td>   0.624</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>Method:</th>             <td>Least Squares</td>  <th>  F-statistic:       </th> <td>   74.38</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>Date:</th>             <td>Tue, 26 Jan 2021</td> <th>  Prob (F-statistic):</th> <td>1.83e-70</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>Time:</th>                 <td>13:36:02</td>     <th>  Log-Likelihood:    </th> <td> -1099.6</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>No. Observations:</th>      <td>   354</td>      <th>  AIC:               </th> <td>   2217.</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>Df Residuals:</th>          <td>   345</td>      <th>  BIC:               </th> <td>   2252.</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>Df Model:</th>              <td>     8</td>      <th>                     </th>     <td> </td>   \n",
              "</tr>\n",
              "<tr>\n",
              "  <th>Covariance Type:</th>      <td>nonrobust</td>    <th>                     </th>     <td> </td>   \n",
              "</tr>\n",
              "</table>\n",
              "<table class=\"simpletable\">\n",
              "<tr>\n",
              "     <td></td>        <th>coef</th>     <th>std err</th>      <th>t</th>      <th>P>|t|</th>  <th>[0.025</th>    <th>0.975]</th>  \n",
              "</tr>\n",
              "<tr>\n",
              "  <th>const</th>   <td>  -11.3930</td> <td>    8.273</td> <td>   -1.377</td> <td> 0.169</td> <td>  -27.664</td> <td>    4.878</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>INDUS</th>   <td>   -0.1324</td> <td>    0.077</td> <td>   -1.715</td> <td> 0.087</td> <td>   -0.284</td> <td>    0.019</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>RM</th>      <td>    6.1019</td> <td>    0.461</td> <td>   13.241</td> <td> 0.000</td> <td>    5.196</td> <td>    7.008</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>AGE</th>     <td>   -0.0769</td> <td>    0.017</td> <td>   -4.602</td> <td> 0.000</td> <td>   -0.110</td> <td>   -0.044</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>DIS</th>     <td>   -1.5708</td> <td>    0.318</td> <td>   -4.943</td> <td> 0.000</td> <td>   -2.196</td> <td>   -0.946</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>TAX</th>     <td>   -0.0068</td> <td>    0.003</td> <td>   -2.518</td> <td> 0.012</td> <td>   -0.012</td> <td>   -0.001</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>PTRATIO</th> <td>   -0.9154</td> <td>    0.177</td> <td>   -5.184</td> <td> 0.000</td> <td>   -1.263</td> <td>   -0.568</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>ZN_Cbrt</th> <td>    0.4088</td> <td>    0.302</td> <td>    1.356</td> <td> 0.176</td> <td>   -0.184</td> <td>    1.002</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>B_Sqrt</th>  <td>    1.4215</td> <td>    0.360</td> <td>    3.947</td> <td> 0.000</td> <td>    0.713</td> <td>    2.130</td>\n",
              "</tr>\n",
              "</table>\n",
              "<table class=\"simpletable\">\n",
              "<tr>\n",
              "  <th>Omnibus:</th>       <td>210.014</td> <th>  Durbin-Watson:     </th> <td>   2.095</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>Prob(Omnibus):</th> <td> 0.000</td>  <th>  Jarque-Bera (JB):  </th> <td>2347.177</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>Skew:</th>          <td> 2.277</td>  <th>  Prob(JB):          </th> <td>    0.00</td>\n",
              "</tr>\n",
              "<tr>\n",
              "  <th>Kurtosis:</th>      <td>14.764</td>  <th>  Cond. No.          </th> <td>1.28e+04</td>\n",
              "</tr>\n",
              "</table><br/><br/>Warnings:<br/>[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.<br/>[2] The condition number is large, 1.28e+04. This might indicate that there are<br/>strong multicollinearity or other numerical problems."
            ],
            "text/plain": [
              "<class 'statsmodels.iolib.summary.Summary'>\n",
              "\"\"\"\n",
              "                            OLS Regression Results                            \n",
              "==============================================================================\n",
              "Dep. Variable:             HOUSEPRICE   R-squared:                       0.633\n",
              "Model:                            OLS   Adj. R-squared:                  0.624\n",
              "Method:                 Least Squares   F-statistic:                     74.38\n",
              "Date:                Tue, 26 Jan 2021   Prob (F-statistic):           1.83e-70\n",
              "Time:                        13:36:02   Log-Likelihood:                -1099.6\n",
              "No. Observations:                 354   AIC:                             2217.\n",
              "Df Residuals:                     345   BIC:                             2252.\n",
              "Df Model:                           8                                         \n",
              "Covariance Type:            nonrobust                                         \n",
              "==============================================================================\n",
              "                 coef    std err          t      P>|t|      [0.025      0.975]\n",
              "------------------------------------------------------------------------------\n",
              "const        -11.3930      8.273     -1.377      0.169     -27.664       4.878\n",
              "INDUS         -0.1324      0.077     -1.715      0.087      -0.284       0.019\n",
              "RM             6.1019      0.461     13.241      0.000       5.196       7.008\n",
              "AGE           -0.0769      0.017     -4.602      0.000      -0.110      -0.044\n",
              "DIS           -1.5708      0.318     -4.943      0.000      -2.196      -0.946\n",
              "TAX           -0.0068      0.003     -2.518      0.012      -0.012      -0.001\n",
              "PTRATIO       -0.9154      0.177     -5.184      0.000      -1.263      -0.568\n",
              "ZN_Cbrt        0.4088      0.302      1.356      0.176      -0.184       1.002\n",
              "B_Sqrt         1.4215      0.360      3.947      0.000       0.713       2.130\n",
              "==============================================================================\n",
              "Omnibus:                      210.014   Durbin-Watson:                   2.095\n",
              "Prob(Omnibus):                  0.000   Jarque-Bera (JB):             2347.177\n",
              "Skew:                           2.277   Prob(JB):                         0.00\n",
              "Kurtosis:                      14.764   Cond. No.                     1.28e+04\n",
              "==============================================================================\n",
              "\n",
              "Warnings:\n",
              "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
              "[2] The condition number is large, 1.28e+04. This might indicate that there are\n",
              "strong multicollinearity or other numerical problems.\n",
              "\"\"\""
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 33
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_9ECtz0gjVhi",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "b5ad4f4b-c3a7-4757-f230-03b7ef884c70"
      },
      "source": [
        "# Model 3 Properties\n",
        "\n",
        "# Normality of the residuals\n",
        "# Jarque-Bera test:\n",
        "name = ['Jarque-Bera', 'Chi^2 two-tail prob.', 'Skew', 'Kurtosis']\n",
        "test = sms.jarque_bera(Model_3.resid)\n",
        "base_model_properties_3 = lzip(name, test)\n",
        "print(\"MODEL 3 PARAMETERS\")\n",
        "print(\"\\n\")\n",
        "print(\"Normality of the residuals\")\n",
        "print(\"Jarque-Bera test : \")\n",
        "print(\"JB, ideal is close to 0\")\n",
        "print(\"Probability(JB), less than 0.05\")\n",
        "print(\"Skewness, ideal is -1 to 1\")\n",
        "print(\"Kurtosis, ideal value is 3\")\n",
        "print(pd.DataFrame(base_model_properties_3,columns = ['Test','Values']).round(2))\n",
        "\n",
        "# Omni test:\n",
        "print(\"Omni Test : \")\n",
        "print(\"Omni, ideal is close to 0\")\n",
        "print(\"Probability(Omni), less than 0.05\")\n",
        "name = ['Chi^2', 'Two-tail probability']\n",
        "test = sms.omni_normtest(Model_3.resid)\n",
        "print(pd.DataFrame(lzip(name, test),columns = ['Test','Values']).round(2))\n",
        "\n",
        "# Multicollinearity\n",
        "print(\"\\n\")\n",
        "print(\"Multicollinearity\")\n",
        "print(\"Ideal Condition Number will be less than 30\")\n",
        "print(\"Condition number:\",np.linalg.cond(Model_3.model.exog).round(2))\n",
        "\n",
        "# Adjusted R Square\n",
        "print(\"\\n\")\n",
        "print(\"Model Performance\")\n",
        "print(\"Adjusted R Square, ideal is close to 1\")\n",
        "print(\"Adjusted R Square : \",Model_3.rsquared_adj.round(2))\n",
        "\n",
        "# F Statistics\n",
        "print(\"\\n\")\n",
        "print(\"Significance of Independent Variables\")\n",
        "print(\"Probability(F Stat), less than 0.05 suggests independent variables are important\")\n",
        "print(\"F Statistics : \",base_model.fvalue.round(2))\n",
        "print(\"Probability of F Statistics : \",Model_3.f_pvalue.round(2))\n",
        "\n",
        "# Durbin Watson Test\n",
        "print(\"\\n\")\n",
        "print(\"Homoscedasticity of Error\")\n",
        "print(\"Durbin Watson, ideal is between 1 to 2\")\n",
        "print(\"Durbin Watson Value : 2.07\")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "MODEL 3 PARAMETERS\n",
            "\n",
            "\n",
            "Normality of the residuals\n",
            "Jarque-Bera test : \n",
            "JB, ideal is close to 0\n",
            "Probability(JB), less than 0.05\n",
            "Skewness, ideal is -1 to 1\n",
            "Kurtosis, ideal value is 3\n",
            "                   Test   Values\n",
            "0           Jarque-Bera  2347.18\n",
            "1  Chi^2 two-tail prob.     0.00\n",
            "2                  Skew     2.28\n",
            "3              Kurtosis    14.76\n",
            "Omni Test : \n",
            "Omni, ideal is close to 0\n",
            "Probability(Omni), less than 0.05\n",
            "                   Test  Values\n",
            "0                 Chi^2  210.01\n",
            "1  Two-tail probability    0.00\n",
            "\n",
            "\n",
            "Multicollinearity\n",
            "Ideal Condition Number will be less than 30\n",
            "Condition number: 12774.1\n",
            "\n",
            "\n",
            "Model Performance\n",
            "Adjusted R Square, ideal is close to 1\n",
            "Adjusted R Square :  0.62\n",
            "\n",
            "\n",
            "Significance of Independent Variables\n",
            "Probability(F Stat), less than 0.05 suggests independent variables are important\n",
            "F Statistics :  89.01\n",
            "Probability of F Statistics :  0.0\n",
            "\n",
            "\n",
            "Homoscedasticity of Error\n",
            "Durbin Watson, ideal is between 1 to 2\n",
            "Durbin Watson Value : 2.07\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YBiI1s4aWJaY",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "d5c4bc3b-a5fb-4ec0-f217-aefc5ea90d67"
      },
      "source": [
        "# Prediction of Model 3\n",
        "\n",
        "Prediction_3 = Model_3.predict(data_test)\n",
        "print(\"Mean Absolute Error :\",metrics.mean_absolute_error(house_price_test,Prediction_3).round(2))\n",
        "print(\"Mean Squared Error :\",metrics.mean_squared_error(house_price_test,Prediction_3).round(2))\n",
        "print(\"Root Mean Squared Error :\",metrics.mean_squared_error(house_price_test,Prediction_3).round(2))\n",
        "print(\"Mean Absolute Percentage Error :\",mape(house_price_test,Prediction_3).round(2))\n",
        "\n",
        "# Inference : \n",
        "# Model 1 MAPE = 49.7% decreased to 48.3% in Model 2 and decreased to 47.9% in Model 3.\n",
        "# Model 3 is performed well out of 3 models (Model 3 should be selected)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Mean Absolute Error : 3.83\n",
            "Mean Squared Error : 34.57\n",
            "Root Mean Squared Error : 34.57\n",
            "Mean Absolute Percentage Error : 50.33\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Hb228jU0nmJE",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "39606c31-4bc0-4493-968c-4efcb3adae6f"
      },
      "source": [
        "# Overall Comparision\n",
        "print(\"ALL MODEL COMPARISION\")\n",
        "\n",
        "print(\"\\n\")\n",
        "# Normality of the residuals\n",
        "# Jarque-Bera test:\n",
        "print(\"Normality of the residuals\")\n",
        "print(\"Jarque-Bera test : \")\n",
        "print(\"JB, ideal is close to 0\")\n",
        "print(\"Probability(JB), less than 0.05\")\n",
        "print(\"Base Model\")\n",
        "print(base_model_properties_1[0])\n",
        "print(base_model_properties_1[1])\n",
        "print(\"Model 1\")\n",
        "print(model_properties_1[0])\n",
        "print(model_properties_1[1])\n",
        "print(\"Model 2\")\n",
        "print(base_model_properties_2[0])\n",
        "print(base_model_properties_2[1])\n",
        "print(\"Model 3\")\n",
        "print(base_model_properties_3[0])\n",
        "print(base_model_properties_3[1])\n",
        "\n",
        "# Skewness and Kurtosis\n",
        "print(\"\\n\")\n",
        "print(\"Skewness and Kurtosis of Residual\")\n",
        "print(\"Skewness, ideal is -1 to 1\")\n",
        "print(\"Kurtosis, ideal value is 3\")\n",
        "print(\"Base Model\")\n",
        "print(base_model_properties_1[2])\n",
        "print(base_model_properties_1[3])\n",
        "print(\"Model 1\")\n",
        "print(model_properties_1[2])\n",
        "print(model_properties_1[3])\n",
        "print(\"Model 2\")\n",
        "print(base_model_properties_2[2])\n",
        "print(base_model_properties_2[3])\n",
        "print(\"Model 3\")\n",
        "print(base_model_properties_3[2])\n",
        "print(base_model_properties_3[3])\n",
        "\n",
        "# Omni test:\n",
        "print(\"\\n\")\n",
        "print(\"Normality of the Residual also\")\n",
        "print(\"Omni Test : \")\n",
        "print(\"Omni, ideal is close to 0\")\n",
        "print(\"Probability(Omni), less than 0.05\")\n",
        "print(\"Base Model\")\n",
        "name = ['Chi^2', 'Two-tail probability']\n",
        "test = sms.omni_normtest(base_model.resid)\n",
        "print(pd.DataFrame(lzip(name, test),columns = ['Test','Values']).round(2))\n",
        "print(\"Model 1\")\n",
        "name = ['Chi^2', 'Two-tail probability']\n",
        "test = sms.omni_normtest(Model_1.resid)\n",
        "print(pd.DataFrame(lzip(name, test),columns = ['Test','Values']).round(2))\n",
        "print(\"Model 2\")\n",
        "name = ['Chi^2', 'Two-tail probability']\n",
        "test = sms.omni_normtest(Model_2.resid)\n",
        "print(pd.DataFrame(lzip(name, test),columns = ['Test','Values']).round(2))\n",
        "print(\"Model 3\")\n",
        "name = ['Chi^2', 'Two-tail probability']\n",
        "test = sms.omni_normtest(Model_3.resid)\n",
        "print(pd.DataFrame(lzip(name, test),columns = ['Test','Values']).round(2))\n",
        "\n",
        "# Multicollinearity\n",
        "print(\"\\n\")\n",
        "print(\"Multicollinearity\")\n",
        "print(\"Ideal Condition Number will be less than 30\")\n",
        "print(\"Base Model, Condition number:\",np.linalg.cond(base_model.model.exog).round(2))\n",
        "print(\"Model 1, Condition number:\",np.linalg.cond(Model_1.model.exog).round(2))\n",
        "print(\"Model 2, Condition number:\",np.linalg.cond(Model_2.model.exog).round(2))\n",
        "print(\"Model 3, Condition number:\",np.linalg.cond(Model_3.model.exog).round(2))\n",
        "\n",
        "# Adjusted R Square\n",
        "print(\"\\n\")\n",
        "print(\"Model Performance\")\n",
        "print(\"Adjusted R Square, ideal is close to 1\")\n",
        "print(\"Base Model, Adjusted R Square : \",base_model.rsquared_adj.round(2))\n",
        "print(\"Model 1, Adjusted R Square : \",Model_1.rsquared_adj.round(2))\n",
        "print(\"Model 2, Adjusted R Square : \",Model_2.rsquared_adj.round(2))\n",
        "print(\"Model 3, Adjusted R Square : \",Model_3.rsquared_adj.round(2))\n",
        "\n",
        "# F Statistics\n",
        "print(\"\\n\")\n",
        "print(\"Significance of Independent Variables\")\n",
        "print(\"Probability(F Stat), less than 0.05 suggests independent variables are important\")\n",
        "print(\"Base Model, Probability of F Statistics : \",base_model.f_pvalue.round(2))\n",
        "print(\"Model 1, Probability of F Statistics : \",Model_1.f_pvalue.round(2))\n",
        "print(\"Model 2, Probability of F Statistics : \",Model_2.f_pvalue.round(2))\n",
        "print(\"Model 3, Probability of F Statistics : \",Model_3.f_pvalue.round(2))\n",
        "\n",
        "# Durbin Watson Test\n",
        "print(\"\\n\")\n",
        "print(\"Homoscedasticity of Error\")\n",
        "print(\"Durbin Watson, ideal is between 1 to 2\")\n",
        "print(\"Base Model, Durbin Watson Value : 0.78\")\n",
        "print(\"Model 1, Durbin Watson Value : 0.78\")\n",
        "print(\"Model 2, Durbin Watson Value : 2.1\")\n",
        "print(\"Model 3, Durbin Watson Value : 2.07\")\n",
        "\n",
        "# MAPE\n",
        "print(\"\\n\")\n",
        "print(\"Accuracy of the Model\")\n",
        "print(\"MAPE should be close to 0\")\n",
        "print(\"Base Model, will not have MAPE value because we train the data on full dataset\")\n",
        "print(\"Model 1, Mean Absolute Percentage Error :\",mape(house_price_test,Prediction_1).round(2))\n",
        "print(\"Model 2, Mean Absolute Percentage Error :\",mape(house_price_test,Prediction_2).round(2))\n",
        "print(\"Model 3, Mean Absolute Percentage Error :\",mape(house_price_test,Prediction_3).round(2))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "ALL MODEL COMPARISION\n",
            "\n",
            "\n",
            "Normality of the residuals\n",
            "Jarque-Bera test : \n",
            "JB, ideal is close to 0\n",
            "Probability(JB), less than 0.05\n",
            "Base Model\n",
            "('Jarque-Bera', 2542.250488008214)\n",
            "('Chi^2 two-tail prob.', 0.0)\n",
            "Model 1\n",
            "('Jarque-Bera', 1626.0437912680602)\n",
            "('Chi^2 two-tail prob.', 0.0)\n",
            "Model 2\n",
            "('Jarque-Bera', 2161.2461350595318)\n",
            "('Chi^2 two-tail prob.', 0.0)\n",
            "Model 3\n",
            "('Jarque-Bera', 2347.1769653084443)\n",
            "('Chi^2 two-tail prob.', 0.0)\n",
            "\n",
            "\n",
            "Skewness and Kurtosis of Residual\n",
            "Skewness, ideal is -1 to 1\n",
            "Kurtosis, ideal value is 3\n",
            "Base Model\n",
            "('Skew', 2.1082399036809814)\n",
            "('Kurtosis', 13.139149315831563)\n",
            "Model 1\n",
            "('Skew', 2.0386930060436312)\n",
            "('Kurtosis', 12.675493803602887)\n",
            "Model 2\n",
            "('Skew', 2.191454152691215)\n",
            "('Kurtosis', 14.283407131073531)\n",
            "Model 3\n",
            "('Skew', 2.2769330038791407)\n",
            "('Kurtosis', 14.764053132284399)\n",
            "\n",
            "\n",
            "Normality of the Residual also\n",
            "Omni Test : \n",
            "Omni, ideal is close to 0\n",
            "Probability(Omni), less than 0.05\n",
            "Base Model\n",
            "                   Test  Values\n",
            "0                 Chi^2  267.27\n",
            "1  Two-tail probability    0.00\n",
            "Model 1\n",
            "                   Test  Values\n",
            "0                 Chi^2  186.74\n",
            "1  Two-tail probability    0.00\n",
            "Model 2\n",
            "                   Test  Values\n",
            "0                 Chi^2  202.92\n",
            "1  Two-tail probability    0.00\n",
            "Model 3\n",
            "                   Test  Values\n",
            "0                 Chi^2  210.01\n",
            "1  Two-tail probability    0.00\n",
            "\n",
            "\n",
            "Multicollinearity\n",
            "Ideal Condition Number will be less than 30\n",
            "Base Model, Condition number: 14793.16\n",
            "Model 1, Condition number: 14949.83\n",
            "Model 2, Condition number: 13155.98\n",
            "Model 3, Condition number: 12774.1\n",
            "\n",
            "\n",
            "Model Performance\n",
            "Adjusted R Square, ideal is close to 1\n",
            "Base Model, Adjusted R Square :  0.68\n",
            "Model 1, Adjusted R Square :  0.68\n",
            "Model 2, Adjusted R Square :  0.63\n",
            "Model 3, Adjusted R Square :  0.62\n",
            "\n",
            "\n",
            "Significance of Independent Variables\n",
            "Probability(F Stat), less than 0.05 suggests independent variables are important\n",
            "Base Model, Probability of F Statistics :  0.0\n",
            "Model 1, Probability of F Statistics :  0.0\n",
            "Model 2, Probability of F Statistics :  0.0\n",
            "Model 3, Probability of F Statistics :  0.0\n",
            "\n",
            "\n",
            "Homoscedasticity of Error\n",
            "Durbin Watson, ideal is between 1 to 2\n",
            "Base Model, Durbin Watson Value : 0.78\n",
            "Model 1, Durbin Watson Value : 0.78\n",
            "Model 2, Durbin Watson Value : 2.1\n",
            "Model 3, Durbin Watson Value : 2.07\n",
            "\n",
            "\n",
            "Accuracy of the Model\n",
            "MAPE should be close to 0\n",
            "Base Model, will not have MAPE value because we train the data on full dataset\n",
            "Model 1, Mean Absolute Percentage Error : 52.39\n",
            "Model 2, Mean Absolute Percentage Error : 50.5\n",
            "Model 3, Mean Absolute Percentage Error : 50.33\n"
          ],
          "name": "stdout"
        }
      ]
    }
  ]
}