2-linear-models-dummies.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Diamonds Price Prediction - Linear Models - Dummies"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### In this notebook we will follow with the implemenation of linear models in the diamonds dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/diego/anaconda3/lib/python3.7/site-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"
     ]
    }
   ],
   "source": [
    "#import general libraries\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Training Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(40339, 9)\n"
     ]
    },
    {
     "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>Unnamed: 0</th>\n",
       "      <th>carat</th>\n",
       "      <th>cut</th>\n",
       "      <th>color</th>\n",
       "      <th>clarity</th>\n",
       "      <th>depth</th>\n",
       "      <th>table</th>\n",
       "      <th>price</th>\n",
       "      <th>volume</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>1.21</td>\n",
       "      <td>Ideal</td>\n",
       "      <td>H</td>\n",
       "      <td>VS2</td>\n",
       "      <td>63.0</td>\n",
       "      <td>57.0</td>\n",
       "      <td>6134</td>\n",
       "      <td>190.734930</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0.28</td>\n",
       "      <td>Very Good</td>\n",
       "      <td>D</td>\n",
       "      <td>VVS2</td>\n",
       "      <td>64.0</td>\n",
       "      <td>56.0</td>\n",
       "      <td>532</td>\n",
       "      <td>45.921708</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>0.42</td>\n",
       "      <td>Premium</td>\n",
       "      <td>F</td>\n",
       "      <td>VS1</td>\n",
       "      <td>61.2</td>\n",
       "      <td>58.0</td>\n",
       "      <td>1103</td>\n",
       "      <td>69.338592</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>0.26</td>\n",
       "      <td>Ideal</td>\n",
       "      <td>H</td>\n",
       "      <td>IF</td>\n",
       "      <td>61.1</td>\n",
       "      <td>57.0</td>\n",
       "      <td>600</td>\n",
       "      <td>43.362176</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>1.10</td>\n",
       "      <td>Good</td>\n",
       "      <td>G</td>\n",
       "      <td>SI1</td>\n",
       "      <td>63.4</td>\n",
       "      <td>57.0</td>\n",
       "      <td>4997</td>\n",
       "      <td>176.802840</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Unnamed: 0  carat        cut color clarity  depth  table  price      volume\n",
       "0           0   1.21      Ideal     H     VS2   63.0   57.0   6134  190.734930\n",
       "1           1   0.28  Very Good     D    VVS2   64.0   56.0    532   45.921708\n",
       "2           2   0.42    Premium     F     VS1   61.2   58.0   1103   69.338592\n",
       "3           3   0.26      Ideal     H      IF   61.1   57.0    600   43.362176\n",
       "4           4   1.10       Good     G     SI1   63.4   57.0   4997  176.802840"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "traindf = pd.read_csv(\"OUTPUT/cleaned_train.csv\")\n",
    "print(traindf.shape)\n",
    "traindf.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### To implement linear models in our data we need int instead of categories"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Dummy Version"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>carat</th>\n",
       "      <th>depth</th>\n",
       "      <th>table</th>\n",
       "      <th>price</th>\n",
       "      <th>volume</th>\n",
       "      <th>cut_Fair</th>\n",
       "      <th>cut_Good</th>\n",
       "      <th>cut_Ideal</th>\n",
       "      <th>cut_Premium</th>\n",
       "      <th>cut_Very Good</th>\n",
       "      <th>...</th>\n",
       "      <th>color_I</th>\n",
       "      <th>color_J</th>\n",
       "      <th>clarity_I1</th>\n",
       "      <th>clarity_IF</th>\n",
       "      <th>clarity_SI1</th>\n",
       "      <th>clarity_SI2</th>\n",
       "      <th>clarity_VS1</th>\n",
       "      <th>clarity_VS2</th>\n",
       "      <th>clarity_VVS1</th>\n",
       "      <th>clarity_VVS2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.21</td>\n",
       "      <td>63.0</td>\n",
       "      <td>57.0</td>\n",
       "      <td>6134</td>\n",
       "      <td>190.734930</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.28</td>\n",
       "      <td>64.0</td>\n",
       "      <td>56.0</td>\n",
       "      <td>532</td>\n",
       "      <td>45.921708</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.42</td>\n",
       "      <td>61.2</td>\n",
       "      <td>58.0</td>\n",
       "      <td>1103</td>\n",
       "      <td>69.338592</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.26</td>\n",
       "      <td>61.1</td>\n",
       "      <td>57.0</td>\n",
       "      <td>600</td>\n",
       "      <td>43.362176</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1.10</td>\n",
       "      <td>63.4</td>\n",
       "      <td>57.0</td>\n",
       "      <td>4997</td>\n",
       "      <td>176.802840</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 25 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   carat  depth  table  price      volume  cut_Fair  cut_Good  cut_Ideal  \\\n",
       "0   1.21   63.0   57.0   6134  190.734930         0         0          1   \n",
       "1   0.28   64.0   56.0    532   45.921708         0         0          0   \n",
       "2   0.42   61.2   58.0   1103   69.338592         0         0          0   \n",
       "3   0.26   61.1   57.0    600   43.362176         0         0          1   \n",
       "4   1.10   63.4   57.0   4997  176.802840         0         1          0   \n",
       "\n",
       "   cut_Premium  cut_Very Good  ...  color_I  color_J  clarity_I1  clarity_IF  \\\n",
       "0            0              0  ...        0        0           0           0   \n",
       "1            0              1  ...        0        0           0           0   \n",
       "2            1              0  ...        0        0           0           0   \n",
       "3            0              0  ...        0        0           0           1   \n",
       "4            0              0  ...        0        0           0           0   \n",
       "\n",
       "   clarity_SI1  clarity_SI2  clarity_VS1  clarity_VS2  clarity_VVS1  \\\n",
       "0            0            0            0            1             0   \n",
       "1            0            0            0            0             0   \n",
       "2            0            0            1            0             0   \n",
       "3            0            0            0            0             0   \n",
       "4            1            0            0            0             0   \n",
       "\n",
       "   clarity_VVS2  \n",
       "0             0  \n",
       "1             1  \n",
       "2             0  \n",
       "3             0  \n",
       "4             0  \n",
       "\n",
       "[5 rows x 25 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_dum = pd.get_dummies(traindf.iloc[:,1:])\n",
    "train_dum.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### But before implementing the linear models, we have to check 3 steps:\n",
    "    - Correlation between variables\n",
    "    - How each independent variable explain price\n",
    "    - If we need to scale our data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7fb774b4c890>"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 792x792 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Generate a mask for the upper triangle\n",
    "#mask = np.triu(np.ones_like(train_dum.corr(), dtype=np.bool))\n",
    "#I let the mask here in case someone prefers to apply it, however I see more clear inputs without the mask\n",
    "\n",
    "plt.figure(figsize=(11, 11))\n",
    "\n",
    "# Generate a custom diverging colormap\n",
    "cmap = sns.diverging_palette(220, 10, as_cmap=True)\n",
    "\n",
    "# Draw the heatmap with the mask and correct aspect ratio\n",
    "sns.heatmap(train_dum.corr(), mask=None, cmap=cmap, vmax=1, center=0,\n",
    "            square=True, linewidths=.5, cbar_kws={\"shrink\": .5})"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### As we have seen in the descriptive analysis, the price is highly correlated with carat and volume and also these to independet variables too. The problem with the corr between the 2 independent variables is that it causes the predictive coefficients to be unstable, because the \"independent variables\" have become \"dependent\". So we are going to do an OLS regression using statsmodels, to check the effect of the independent variables on the dependent one"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_t = train_dum.drop(columns={\"price\"})\n",
    "y_t = train_dum.price"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1149.4188308870382\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<table class=\"simpletable\">\n",
       "<caption>OLS Regression Results</caption>\n",
       "<tr>\n",
       "  <th>Dep. Variable:</th>          <td>price</td>      <th>  R-squared:         </th>  <td>   0.917</td>  \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Model:</th>                   <td>OLS</td>       <th>  Adj. R-squared:    </th>  <td>   0.917</td>  \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Method:</th>             <td>Least Squares</td>  <th>  F-statistic:       </th>  <td>2.111e+04</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Date:</th>             <td>Sun, 10 May 2020</td> <th>  Prob (F-statistic):</th>   <td>  0.00</td>   \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Time:</th>                 <td>18:04:26</td>     <th>  Log-Likelihood:    </th> <td>-3.4151e+05</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>No. Observations:</th>      <td> 40339</td>      <th>  AIC:               </th>  <td>6.831e+05</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Df Residuals:</th>          <td> 40317</td>      <th>  BIC:               </th>  <td>6.832e+05</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Df Model:</th>              <td>    21</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> -820.3748</td> <td>  282.842</td> <td>   -2.900</td> <td> 0.004</td> <td>-1374.751</td> <td> -265.999</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>carat</th>         <td> 8815.3800</td> <td>   51.515</td> <td>  171.122</td> <td> 0.000</td> <td> 8714.409</td> <td> 8916.351</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>depth</th>         <td>  -18.2060</td> <td>    4.718</td> <td>   -3.859</td> <td> 0.000</td> <td>  -27.453</td> <td>   -8.959</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>table</th>         <td>  -23.8873</td> <td>    3.418</td> <td>   -6.989</td> <td> 0.000</td> <td>  -30.586</td> <td>  -17.188</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>volume</th>        <td>    0.7376</td> <td>    0.306</td> <td>    2.409</td> <td> 0.016</td> <td>    0.138</td> <td>    1.338</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>cut_Fair</th>      <td> -772.7999</td> <td>   72.215</td> <td>  -10.701</td> <td> 0.000</td> <td> -914.343</td> <td> -631.257</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>cut_Good</th>      <td> -157.1419</td> <td>   61.084</td> <td>   -2.573</td> <td> 0.010</td> <td> -276.868</td> <td>  -37.415</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>cut_Ideal</th>     <td>   93.1201</td> <td>   50.386</td> <td>    1.848</td> <td> 0.065</td> <td>   -5.637</td> <td>  191.877</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>cut_Premium</th>   <td>   22.9246</td> <td>   55.557</td> <td>    0.413</td> <td> 0.680</td> <td>  -85.968</td> <td>  131.817</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>cut_Very Good</th> <td>   -6.4777</td> <td>   55.945</td> <td>   -0.116</td> <td> 0.908</td> <td> -116.131</td> <td>  103.175</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>color_D</th>       <td>  708.5097</td> <td>   42.978</td> <td>   16.485</td> <td> 0.000</td> <td>  624.271</td> <td>  792.748</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>color_E</th>       <td>  500.8047</td> <td>   42.228</td> <td>   11.859</td> <td> 0.000</td> <td>  418.036</td> <td>  583.573</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>color_F</th>       <td>  401.8368</td> <td>   42.140</td> <td>    9.536</td> <td> 0.000</td> <td>  319.242</td> <td>  484.432</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>color_G</th>       <td>  202.3346</td> <td>   42.155</td> <td>    4.800</td> <td> 0.000</td> <td>  119.710</td> <td>  284.960</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>color_H</th>       <td> -278.0395</td> <td>   42.857</td> <td>   -6.488</td> <td> 0.000</td> <td> -362.040</td> <td> -194.040</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>color_I</th>       <td> -733.2558</td> <td>   44.113</td> <td>  -16.622</td> <td> 0.000</td> <td> -819.719</td> <td> -646.793</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>color_J</th>       <td>-1622.5652</td> <td>   46.947</td> <td>  -34.561</td> <td> 0.000</td> <td>-1714.583</td> <td>-1530.547</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>clarity_I1</th>    <td>-3800.0003</td> <td>   58.190</td> <td>  -65.303</td> <td> 0.000</td> <td>-3914.054</td> <td>-3685.946</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>clarity_IF</th>    <td> 1511.2604</td> <td>   44.686</td> <td>   33.819</td> <td> 0.000</td> <td> 1423.674</td> <td> 1598.846</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>clarity_SI1</th>   <td> -354.7603</td> <td>   38.166</td> <td>   -9.295</td> <td> 0.000</td> <td> -429.566</td> <td> -279.955</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>clarity_SI2</th>   <td>-1298.4922</td> <td>   38.724</td> <td>  -33.532</td> <td> 0.000</td> <td>-1374.392</td> <td>-1222.592</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>clarity_VS1</th>   <td>  603.7396</td> <td>   38.241</td> <td>   15.788</td> <td> 0.000</td> <td>  528.787</td> <td>  678.693</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>clarity_VS2</th>   <td>  292.7714</td> <td>   37.851</td> <td>    7.735</td> <td> 0.000</td> <td>  218.583</td> <td>  366.959</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>clarity_VVS1</th>  <td> 1169.2542</td> <td>   40.944</td> <td>   28.557</td> <td> 0.000</td> <td> 1089.003</td> <td> 1249.506</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>clarity_VVS2</th>  <td> 1055.8525</td> <td>   39.479</td> <td>   26.744</td> <td> 0.000</td> <td>  978.472</td> <td> 1133.233</td>\n",
       "</tr>\n",
       "</table>\n",
       "<table class=\"simpletable\">\n",
       "<tr>\n",
       "  <th>Omnibus:</th>       <td>11462.946</td> <th>  Durbin-Watson:     </th>  <td>   2.005</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Prob(Omnibus):</th>  <td> 0.000</td>   <th>  Jarque-Bera (JB):  </th> <td>104715.125</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Skew:</th>           <td> 1.105</td>   <th>  Prob(JB):          </th>  <td>    0.00</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Kurtosis:</th>       <td>10.577</td>   <th>  Cond. No.          </th>  <td>1.91e+18</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 smallest eigenvalue is 3.15e-28. This might indicate that there are<br/>strong multicollinearity problems or that the design matrix is singular."
      ],
      "text/plain": [
       "<class 'statsmodels.iolib.summary.Summary'>\n",
       "\"\"\"\n",
       "                            OLS Regression Results                            \n",
       "==============================================================================\n",
       "Dep. Variable:                  price   R-squared:                       0.917\n",
       "Model:                            OLS   Adj. R-squared:                  0.917\n",
       "Method:                 Least Squares   F-statistic:                 2.111e+04\n",
       "Date:                Sun, 10 May 2020   Prob (F-statistic):               0.00\n",
       "Time:                        18:04:26   Log-Likelihood:            -3.4151e+05\n",
       "No. Observations:               40339   AIC:                         6.831e+05\n",
       "Df Residuals:                   40317   BIC:                         6.832e+05\n",
       "Df Model:                          21                                         \n",
       "Covariance Type:            nonrobust                                         \n",
       "=================================================================================\n",
       "                    coef    std err          t      P>|t|      [0.025      0.975]\n",
       "---------------------------------------------------------------------------------\n",
       "const          -820.3748    282.842     -2.900      0.004   -1374.751    -265.999\n",
       "carat          8815.3800     51.515    171.122      0.000    8714.409    8916.351\n",
       "depth           -18.2060      4.718     -3.859      0.000     -27.453      -8.959\n",
       "table           -23.8873      3.418     -6.989      0.000     -30.586     -17.188\n",
       "volume            0.7376      0.306      2.409      0.016       0.138       1.338\n",
       "cut_Fair       -772.7999     72.215    -10.701      0.000    -914.343    -631.257\n",
       "cut_Good       -157.1419     61.084     -2.573      0.010    -276.868     -37.415\n",
       "cut_Ideal        93.1201     50.386      1.848      0.065      -5.637     191.877\n",
       "cut_Premium      22.9246     55.557      0.413      0.680     -85.968     131.817\n",
       "cut_Very Good    -6.4777     55.945     -0.116      0.908    -116.131     103.175\n",
       "color_D         708.5097     42.978     16.485      0.000     624.271     792.748\n",
       "color_E         500.8047     42.228     11.859      0.000     418.036     583.573\n",
       "color_F         401.8368     42.140      9.536      0.000     319.242     484.432\n",
       "color_G         202.3346     42.155      4.800      0.000     119.710     284.960\n",
       "color_H        -278.0395     42.857     -6.488      0.000    -362.040    -194.040\n",
       "color_I        -733.2558     44.113    -16.622      0.000    -819.719    -646.793\n",
       "color_J       -1622.5652     46.947    -34.561      0.000   -1714.583   -1530.547\n",
       "clarity_I1    -3800.0003     58.190    -65.303      0.000   -3914.054   -3685.946\n",
       "clarity_IF     1511.2604     44.686     33.819      0.000    1423.674    1598.846\n",
       "clarity_SI1    -354.7603     38.166     -9.295      0.000    -429.566    -279.955\n",
       "clarity_SI2   -1298.4922     38.724    -33.532      0.000   -1374.392   -1222.592\n",
       "clarity_VS1     603.7396     38.241     15.788      0.000     528.787     678.693\n",
       "clarity_VS2     292.7714     37.851      7.735      0.000     218.583     366.959\n",
       "clarity_VVS1   1169.2542     40.944     28.557      0.000    1089.003    1249.506\n",
       "clarity_VVS2   1055.8525     39.479     26.744      0.000     978.472    1133.233\n",
       "==============================================================================\n",
       "Omnibus:                    11462.946   Durbin-Watson:                   2.005\n",
       "Prob(Omnibus):                  0.000   Jarque-Bera (JB):           104715.125\n",
       "Skew:                           1.105   Prob(JB):                         0.00\n",
       "Kurtosis:                      10.577   Cond. No.                     1.91e+18\n",
       "==============================================================================\n",
       "\n",
       "Warnings:\n",
       "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
       "[2] The smallest eigenvalue is 3.15e-28. This might indicate that there are\n",
       "strong multicollinearity problems or that the design matrix is singular.\n",
       "\"\"\""
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import statsmodels.api as sm\n",
    "from statsmodels.regression.linear_model import OLS\n",
    "from statsmodels.tools.eval_measures import rmse\n",
    "\n",
    "X = X_t\n",
    "X = sm.add_constant(X_t)\n",
    "y = y_t\n",
    "model = sm.OLS(y, X).fit()\n",
    "y_pred = model.predict(X)\n",
    "resi = rmse(y, y_pred)\n",
    "print(resi)\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### We can see the columns cut Ideal, Premium and Very Good are not significant and Volume isn't at 1% significance level so we are going to try again without these vars to check if the rmse of our model improves"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_t2 = train_dum.drop(columns={\"price\",\"volume\",\"cut_Very Good\",\"cut_Ideal\",\"cut_Premium\" })"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1150.0282365522778\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<table class=\"simpletable\">\n",
       "<caption>OLS Regression Results</caption>\n",
       "<tr>\n",
       "  <th>Dep. Variable:</th>          <td>price</td>      <th>  R-squared:         </th>  <td>   0.917</td>  \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Model:</th>                   <td>OLS</td>       <th>  Adj. R-squared:    </th>  <td>   0.917</td>  \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Method:</th>             <td>Least Squares</td>  <th>  F-statistic:       </th>  <td>2.460e+04</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Date:</th>             <td>Sun, 10 May 2020</td> <th>  Prob (F-statistic):</th>   <td>  0.00</td>   \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Time:</th>                 <td>18:04:26</td>     <th>  Log-Likelihood:    </th> <td>-3.4153e+05</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>No. Observations:</th>      <td> 40339</td>      <th>  AIC:               </th>  <td>6.831e+05</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Df Residuals:</th>          <td> 40320</td>      <th>  BIC:               </th>  <td>6.833e+05</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Df Model:</th>              <td>    18</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>  -26.3884</td> <td>  300.972</td> <td>   -0.088</td> <td> 0.930</td> <td> -616.300</td> <td>  563.523</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>carat</th>        <td> 8933.4634</td> <td>   13.962</td> <td>  639.841</td> <td> 0.000</td> <td> 8906.098</td> <td> 8960.829</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>depth</th>        <td>  -24.2600</td> <td>    4.624</td> <td>   -5.247</td> <td> 0.000</td> <td>  -33.323</td> <td>  -15.197</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>table</th>        <td>  -34.0668</td> <td>    2.910</td> <td>  -11.708</td> <td> 0.000</td> <td>  -39.770</td> <td>  -28.364</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>cut_Fair</th>     <td> -788.7890</td> <td>   37.187</td> <td>  -21.212</td> <td> 0.000</td> <td> -861.676</td> <td> -715.902</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>cut_Good</th>     <td> -186.0689</td> <td>   20.895</td> <td>   -8.905</td> <td> 0.000</td> <td> -227.023</td> <td> -145.115</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>color_D</th>      <td>  823.1917</td> <td>   45.298</td> <td>   18.173</td> <td> 0.000</td> <td>  734.406</td> <td>  911.977</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>color_E</th>      <td>  613.5122</td> <td>   44.692</td> <td>   13.728</td> <td> 0.000</td> <td>  525.915</td> <td>  701.109</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>color_F</th>      <td>  514.3130</td> <td>   44.630</td> <td>   11.524</td> <td> 0.000</td> <td>  426.838</td> <td>  601.788</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>color_G</th>      <td>  316.1526</td> <td>   44.629</td> <td>    7.084</td> <td> 0.000</td> <td>  228.679</td> <td>  403.626</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>color_H</th>      <td> -165.1011</td> <td>   45.383</td> <td>   -3.638</td> <td> 0.000</td> <td> -254.054</td> <td>  -76.148</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>color_I</th>      <td> -618.5355</td> <td>   46.461</td> <td>  -13.313</td> <td> 0.000</td> <td> -709.599</td> <td> -527.472</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>color_J</th>      <td>-1509.9214</td> <td>   49.317</td> <td>  -30.617</td> <td> 0.000</td> <td>-1606.583</td> <td>-1413.260</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>clarity_I1</th>   <td>-3702.9193</td> <td>   60.034</td> <td>  -61.680</td> <td> 0.000</td> <td>-3820.588</td> <td>-3585.250</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>clarity_IF</th>   <td> 1619.1612</td> <td>   45.807</td> <td>   35.347</td> <td> 0.000</td> <td> 1529.378</td> <td> 1708.945</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>clarity_SI1</th>  <td> -260.0570</td> <td>   40.588</td> <td>   -6.407</td> <td> 0.000</td> <td> -339.609</td> <td> -180.505</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>clarity_SI2</th>  <td>-1204.2441</td> <td>   41.143</td> <td>  -29.270</td> <td> 0.000</td> <td>-1284.885</td> <td>-1123.603</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>clarity_VS1</th>  <td>  702.4763</td> <td>   40.308</td> <td>   17.428</td> <td> 0.000</td> <td>  623.472</td> <td>  781.480</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>clarity_VS2</th>  <td>  391.2267</td> <td>   40.021</td> <td>    9.775</td> <td> 0.000</td> <td>  312.784</td> <td>  469.669</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>clarity_VVS1</th> <td> 1272.0367</td> <td>   42.583</td> <td>   29.872</td> <td> 0.000</td> <td> 1188.573</td> <td> 1355.500</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>clarity_VVS2</th> <td> 1155.9309</td> <td>   41.305</td> <td>   27.985</td> <td> 0.000</td> <td> 1074.972</td> <td> 1236.890</td>\n",
       "</tr>\n",
       "</table>\n",
       "<table class=\"simpletable\">\n",
       "<tr>\n",
       "  <th>Omnibus:</th>       <td>11426.265</td> <th>  Durbin-Watson:     </th>  <td>   2.006</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Prob(Omnibus):</th>  <td> 0.000</td>   <th>  Jarque-Bera (JB):  </th> <td>104410.999</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Skew:</th>           <td> 1.101</td>   <th>  Prob(JB):          </th>  <td>    0.00</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Kurtosis:</th>       <td>10.568</td>   <th>  Cond. No.          </th>  <td>7.00e+17</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 smallest eigenvalue is 5.86e-28. This might indicate that there are<br/>strong multicollinearity problems or that the design matrix is singular."
      ],
      "text/plain": [
       "<class 'statsmodels.iolib.summary.Summary'>\n",
       "\"\"\"\n",
       "                            OLS Regression Results                            \n",
       "==============================================================================\n",
       "Dep. Variable:                  price   R-squared:                       0.917\n",
       "Model:                            OLS   Adj. R-squared:                  0.917\n",
       "Method:                 Least Squares   F-statistic:                 2.460e+04\n",
       "Date:                Sun, 10 May 2020   Prob (F-statistic):               0.00\n",
       "Time:                        18:04:26   Log-Likelihood:            -3.4153e+05\n",
       "No. Observations:               40339   AIC:                         6.831e+05\n",
       "Df Residuals:                   40320   BIC:                         6.833e+05\n",
       "Df Model:                          18                                         \n",
       "Covariance Type:            nonrobust                                         \n",
       "================================================================================\n",
       "                   coef    std err          t      P>|t|      [0.025      0.975]\n",
       "--------------------------------------------------------------------------------\n",
       "const          -26.3884    300.972     -0.088      0.930    -616.300     563.523\n",
       "carat         8933.4634     13.962    639.841      0.000    8906.098    8960.829\n",
       "depth          -24.2600      4.624     -5.247      0.000     -33.323     -15.197\n",
       "table          -34.0668      2.910    -11.708      0.000     -39.770     -28.364\n",
       "cut_Fair      -788.7890     37.187    -21.212      0.000    -861.676    -715.902\n",
       "cut_Good      -186.0689     20.895     -8.905      0.000    -227.023    -145.115\n",
       "color_D        823.1917     45.298     18.173      0.000     734.406     911.977\n",
       "color_E        613.5122     44.692     13.728      0.000     525.915     701.109\n",
       "color_F        514.3130     44.630     11.524      0.000     426.838     601.788\n",
       "color_G        316.1526     44.629      7.084      0.000     228.679     403.626\n",
       "color_H       -165.1011     45.383     -3.638      0.000    -254.054     -76.148\n",
       "color_I       -618.5355     46.461    -13.313      0.000    -709.599    -527.472\n",
       "color_J      -1509.9214     49.317    -30.617      0.000   -1606.583   -1413.260\n",
       "clarity_I1   -3702.9193     60.034    -61.680      0.000   -3820.588   -3585.250\n",
       "clarity_IF    1619.1612     45.807     35.347      0.000    1529.378    1708.945\n",
       "clarity_SI1   -260.0570     40.588     -6.407      0.000    -339.609    -180.505\n",
       "clarity_SI2  -1204.2441     41.143    -29.270      0.000   -1284.885   -1123.603\n",
       "clarity_VS1    702.4763     40.308     17.428      0.000     623.472     781.480\n",
       "clarity_VS2    391.2267     40.021      9.775      0.000     312.784     469.669\n",
       "clarity_VVS1  1272.0367     42.583     29.872      0.000    1188.573    1355.500\n",
       "clarity_VVS2  1155.9309     41.305     27.985      0.000    1074.972    1236.890\n",
       "==============================================================================\n",
       "Omnibus:                    11426.265   Durbin-Watson:                   2.006\n",
       "Prob(Omnibus):                  0.000   Jarque-Bera (JB):           104410.999\n",
       "Skew:                           1.101   Prob(JB):                         0.00\n",
       "Kurtosis:                      10.568   Cond. No.                     7.00e+17\n",
       "==============================================================================\n",
       "\n",
       "Warnings:\n",
       "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
       "[2] The smallest eigenvalue is 5.86e-28. This might indicate that there are\n",
       "strong multicollinearity problems or that the design matrix is singular.\n",
       "\"\"\""
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X = X_t2\n",
    "X = sm.add_constant(X_t2)\n",
    "y = y_t\n",
    "model = sm.OLS(y, X).fit()\n",
    "y_pred = model.predict(X)\n",
    "resi2 = rmse(y, y_pred)\n",
    "print(resi2)\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### As we see, excluding that columns made our rmse a bit higher, so we will not exclude any columns to train our models. But before training our models, lets see if the data needs to be scaled"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x720 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "X_notdum = train_dum.iloc[:,[0,1,2,4]]\n",
    "\n",
    "figure = plt.figure(figsize=(15,10))\n",
    "\n",
    "for n, col in enumerate(X_notdum.columns):\n",
    "    ax = figure.add_subplot(3,4,n+1)\n",
    "    ax.set_title(col)\n",
    "    X_notdum[col].hist(ax=ax, bins=50)\n",
    "    \n",
    "figure.tight_layout() #this feature separate the graphs correctly\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### We can see there is a big difference between the values of the explanatory variables so it is better to scale our data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.preprocessing import StandardScaler\n",
    "sc_X = StandardScaler()\n",
    "\n",
    "Xss = sc_X.fit_transform(X_t)\n",
    "\n",
    "train_SS = pd.DataFrame(Xss, columns=['carat', 'depth', 'table', 'volume', 'cut_Fair', 'cut_Good',\n",
    "       'cut_Ideal', 'cut_Premium', 'cut_Very Good', 'color_D', 'color_E',\n",
    "       'color_F', 'color_G', 'color_H', 'color_I', 'color_J', 'clarity_I1',\n",
    "       'clarity_IF', 'clarity_SI1', 'clarity_SI2', 'clarity_VS1',\n",
    "       'clarity_VS2', 'clarity_VVS1', 'clarity_VVS2'])\n",
    "\n",
    "fig, (ax1, ax2) = plt.subplots(ncols=2, figsize=(15, 5))\n",
    "ax1.set_title('Before Scaling')\n",
    "for e in X_t.columns:\n",
    "    sns.kdeplot(X_t[e], ax=ax1)\n",
    "ax2.set_title('After Standard Scaling')\n",
    "for e in train_SS.columns:\n",
    "    sns.kdeplot(train_SS[e], ax=ax2, legend=None)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.preprocessing import MinMaxScaler\n",
    "mmc_X = MinMaxScaler()\n",
    "\n",
    "Xmm = mmc_X.fit_transform(X_t)\n",
    "\n",
    "train_MM = pd.DataFrame(Xmm, columns=['carat', 'depth', 'table', 'volume', 'cut_Fair', 'cut_Good',\n",
    "       'cut_Ideal', 'cut_Premium', 'cut_Very Good', 'color_D', 'color_E',\n",
    "       'color_F', 'color_G', 'color_H', 'color_I', 'color_J', 'clarity_I1',\n",
    "       'clarity_IF', 'clarity_SI1', 'clarity_SI2', 'clarity_VS1',\n",
    "       'clarity_VS2', 'clarity_VVS1', 'clarity_VVS2'])\n",
    "\n",
    "fig, (ax1, ax2) = plt.subplots(ncols=2, figsize=(15, 5))\n",
    "ax1.set_title('Before Scaling')\n",
    "for e in X_t.columns:\n",
    "    sns.kdeplot(X_t[e], ax=ax1)\n",
    "ax2.set_title('After Min-Max Scaling')\n",
    "for e in train_MM.columns:\n",
    "    sns.kdeplot(train_MM[e], ax=ax2, legend=None)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### After processing the data, the standard scaler looks better to train our models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(32271, 24) (8068, 24) (32271,) (8068,)\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import mean_squared_error, r2_score \n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "X = train_SS\n",
    "y = np.ravel(y_t) #we use np.ravel to change shape (x,1) to (x,)\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)\n",
    "print(X_train.shape, X_test.shape, y_train.shape, y_test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "#1st select the splited X and y. Then after your model achieves the accuracy needed, train with 100% X and y\n",
    "#However if you thing you could be overfitting your data you can use the splited model to test\n",
    "# X, y ----- X_train, y_train\n",
    "X_training = X_train\n",
    "y_training = y_train"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Linear Regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.linear_model import LinearRegression\n",
    "\n",
    "lreg = LinearRegression(fit_intercept=True, normalize=False)\n",
    "\n",
    "lreg.fit(X_training, y_training) \n",
    "  \n",
    "lreg_pred = lreg.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "r2: 0.9148719391376832 rmse: 1154.8497174954416\n"
     ]
    }
   ],
   "source": [
    "lreg_r2 = lreg.score(X_test, y_test)\n",
    "\n",
    "lreg_rmse = mean_squared_error(y_test, lreg_pred, sample_weight=None, multioutput='uniform_average', squared=False)\n",
    "\n",
    "print(\"r2:\",lreg_r2,\"rmse:\",lreg_rmse)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Check for Heteroskedasticity --> there will be if the residuals don't have a constant variance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'Residual plot')"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYsAAAEICAYAAACuxNj9AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+j8jraAAAgAElEQVR4nO3dfZAk9X3f8fd39+6Q9k6YY4+c0cHOIZtYBamKDFcyNopDDObhYheyS7ZlL3BGSs5arPIljh2jXCpS2b6K7cSWoSSBzwb5YMeWsWRFlHQ2QsTEcVIoLAlCIPFwSNzBCfFwh544jAT7zR+/bm3vXPf0PHTPdPd8XlVdO/Obnp5fz+z0d37P5u6IiIh0MzXuDIiISPUpWIiISC4FCxERyaVgISIiuRQsREQkl4KFiIjkUrAQ6YOZzZvZp7s8freZ/asCXucCM3tqwOf+opn9/bB5EElSsJDGMrMnzOwlM/uWmX3VzP7UzDYMc0x3b7v7xUXlcdyKCm7SfAoW0nQ/6e4bgDcBPwi8Z8z5EaklBQuZCO7+VeAOQtAAwMzOM7P/bWZfM7PPmdkFicd+0cy+ZGbfNLMvm9l8Iv3vE/v9uJk9bGZfN7MPAJZ47H1mtpi4v9XM3MzWRPevNrMvRq/xJTP7pV7PJzrOr0TPe97M/ouZpX6fzexHzOzeKI/3mtmPROl7gH8GfCAqfX2g19eXyaNgIRPBzE4DLgMORPe3AJ8Cfhs4Gfg14GNmdoqZrQeuBy5z99cBPwLcn3LMTcBfAf8R2AQ8DpzfR7aeBX4COBG4Gni/mZ3Tx/N/CtgGnANcDrwjJY8nE87zemAW+APgU2Y26+67gf8JvNvdN7j7u/t4bZkwChbSdP/NzL4JPEm4OL83Sr8C2O/u+9192d3vBJaA7dHjy8A/MbPXuvvT7v5QyrG3Aw+5+0fd/TvAHwJf7TVj7v4pd3/cg/8BfJrwS79Xv+vuR939UPTaP5+yz78EHnP3W939FXf/c+Bh4Cf7eB0RBQtpvLdGpYMLgDcSSgAALeBnoiqor5nZ14C3AKe6+4vAzwHvAp42s0+Z2RtTjv16QhACwMOsnE+m7JfKzC4zs3vM7Gj0+tsT+etF8rUORvlJy+PBjrSDwJY+XkdEwUImQ/TL/U+B/xolPQnc6u4nJbb17v470f53uPuPA6cSfon/ccphnwZOj++YmSXvAy8CM4n735vY9wTgY1F+Nrv7ScB+Em0ePUi+1hzwlZR9vkIIjHTsezi6rWmnpScKFjJJ/hD4cTP7p8Ai8JNmdomZTZvZa6KxDaeZ2WYzuzxqu3gZ+BahWqrTp4Czzeyno0brXyEREAjtHD9qZnNm9j2s7om1DjgBeA54xcwuA/rtkvvrZrbRzE4HdgF/kbLPfuAfm9kvmNkaM/s54Czgk9HjzwBv6PN1ZQIpWMjEcPfngFuA/+TuTxIahf8D4YL9JPDrhO/EFPCrhF/lR4F/DiykHO954GeA3wGOAGcC/yvx+J2EC/gDwH2sXKBx928SgsttwAvALwC393lKn4iOez8hcN2UkscjhEb0fxfl8d8DPxHlHeA64G1m9oKZXd/n68sEMS1+JFI/ZubAme5+YNx5kcmgkoWIiORSsBARkVyqhhIRkVwqWYiISK41485AWTZt2uRbt24ddzZERGrlvvvue97dT+lMb2yw2Lp1K0tLS+POhohIrZhZ54h/QNVQIiLSAwULERHJpWAhIiK5FCxERCSXgoWIiORSsBCRWmi3YetWmJoKf9vtcedosjS266yINEe7DTt3wrFj4f7Bg+E+wPz8+PI1SVSyEJHK2717JVDEjh0L6TIaChYiUnmHDvWXLsVTsBCRypub6y9diqdgISKVt2cPzMysTpuZCekyGgoWIlJ58/Owdy+0WmAW/u7dq8btUVJvKBGphfl5BYdxUslCRERyKViIiEguBQsREcmlYCEiIrkULEREJJeChYiI5FKwEBGRXAoWIiKSS8FCRERyKViISCVpsaNq0XQfIlI5WuyoelSyEJHK0WJH1aNgISKVo8WOqkfBQkQqR4sdVY+ChYhUjhY7qh4FCxGpHC12VD3qDSUilaTFjqqlkJKFmd1sZs+a2YOJtJPN7E4zeyz6uzFKNzO73swOmNkDZnZO4jk7ov0fM7MdifRzzezz0XOuNzMrIt8iItKboqqh/hS4tCPtWuAudz8TuCu6D3AZcGa07QRugBBcgPcCPwS8GXhvHGCiff514nmdryUiIiUqJFi4+98BRzuSLwf2Rbf3AW9NpN/iwT3ASWZ2KnAJcKe7H3X3F4A7gUujx05093vc3YFbEscSEZERKLOBe7O7Px3d/iqwObq9BXgysd9TUVq39KdS0kVEZERG0hsqKhF42a9jZjvNbMnMlp577rmyX05EZGKUGSyeiaqQiP4+G6UfBk5P7HdalNYt/bSU9OO4+1533+bu20455ZRCTkJERMoNFrcDcY+mHcAnEulXRb2izgO+HlVX3QFcbGYbo4bti4E7ose+YWbnRb2grkocS0RERqCQcRZm9ufABcAmM3uK0Kvpd4DbzOydwEHgZ6Pd9wPbgQPAMeBqAHc/ama/Bdwb7feb7h43ml9D6HH1WuCvo01EREbEQnNC82zbts2XlpbGnQ0RkVoxs/vcfVtnuqb7EBGRXAoWIiKSS8FCRGTM6rCErCYSFBEZo7osIauShYjIGNVlCVkFCxGRMarLErIKFiIiY1SXJWQVLERExqguS8gqWIiIjFFdlpBVsBARGbP5eXjiCVheDn8HDRRldsFV11kRkQYouwuuShYiUil1GKBWRWV3wVXJQkQqoy4D1Kqo7C64KlmISGXUZYBaFZXdBVfBQkRGqls1U10GqFVR2V1wFSxEZGTiaqaDB8F9pZopDhh1GaBWRWV3wdXiRyIyMlu3hgDRqdUKXUY72ywg/Dqu4riDptLiRyIydnnVTHUZoDaJ1BtKREZmbi69ZJGsZpqfV3CoIpUsRGRk6jIPkhxPwUJERkbVTPWlaigRGSlVM9WTShYiIpJLwUJERHIpWIiISC4FCxERyaVgISIiuRQsREQkl4KFiIjkUrAQEZFcChYiIpJLwUJERHIpWIiISC4FCxERyaVgISIiuUoPFmb2hJl93szuN7OlKO1kM7vTzB6L/m6M0s3MrjezA2b2gJmdkzjOjmj/x8xsR9n5FhGRFaMqWfwLd39TYl3Xa4G73P1M4K7oPsBlwJnRthO4AUJwAd4L/BDwZuC9cYAREZHyjasa6nJgX3R7H/DWRPotHtwDnGRmpwKXAHe6+1F3fwG4E7h01JkWEZlUowgWDnzazO4zs51R2mZ3fzq6/VVgc3R7C/Bk4rlPRWlZ6SIiMgKjWCnvLe5+2Mz+EXCnmT2cfNDd3cy8iBeKgtFOgLnkCvAiIjKU0ksW7n44+vss8HFCm8MzUfUS0d9no90PA6cnnn5alJaV3vlae919m7tvO+WUU4o+FRFpmHYbtm6Fqanwt90ed46qq9RgYWbrzex18W3gYuBB4HYg7tG0A/hEdPt24KqoV9R5wNej6qo7gIvNbGPUsH1xlCYiMpB2G3buhIMHwT383blTASNL2dVQm4GPm1n8Wn/m7n9jZvcCt5nZO4GDwM9G++8HtgMHgGPA1QDuftTMfgu4N9rvN939aMl5F5EG270bjh1bnXbsWEifnx9PnqrM3AtpLqicbdu2+dLS0rizISIVNTUVShSdzGB5efT5qQozuy8xzOG7NIJbRFaZlHr8rD4w6huTTsFCRL5rkurx9+yBmZnVaTMzIV2Op2AhUrC6/DJPy2e3evymmZ+HvXuh1QpVT61WuK/2inRqsxApSLsNu3bBkSOr02dmqncRiksQycAwM3N8oEhaXKzWOUg51GYh0qd+SgjxxbczUMDKL/PO411zTf7xyyqlZJUgprpcEepWHVWXEl5tuHsjt3PPPddFBrW46D4z4x5q7sM2MxPS07Raq/dN2zqPl/Z48vj95qEfZvn5TdtareFfexTKfO+aDljylGuqqqFEUmzdGhp3O7VaoQF09244dCj0nNmzB668Mr0bZmx6Gl59Nf91Z2dhw4Zw7Kmp9Oe0WvDEE72eSbqs88tTl26l3T6/Yd+7plM1lEyEoqoeDh1KT497B3X2Fjr55O7H6yVQQKjGio+d9ZysvGW55hpYsyZc6NesCfezegLNznY/VpndSousNsp6j7q9d6q2ypFW3GjCpmqo5ltcDNUiZuHvwkJvVQ+dz0urmsiqVpqeTk/fsCG76mZ2NmyDVPtk5aFb3pMWFtKPsbCQ/j6kVd+Mohqn6GqjrM8vqxpN1VYryKiGGvtFvaxNwaLZ0r7cWfXwyQtErxeFrP36vbDPzIQL89q1xQWLfi5oWcFtenr1uXYG3fhiGz+/l8DU7bMaNDgP2kZSVJtTXdpoiqRgIY3SS4NyZyBptbJ/4ccXseTj69eH+8mLXD+vOz3d/3PyjtfvBa3b8dzL/0Xd6/GzAr3ZcK+dF6TKfP26ygoWauCWWokHjg3SODuIqSnYuBGOHg319du3w003wbe/nf/cuDE4aw6ifoS5ONOP063Rec2a9LaP6Wl45ZXyG4J7Pf64G6TH/fpVogZuqb3kVBSjsrwcGp3dw+vu2wdr1/b23LgxuIhG4bm5weYy2rmze/ogDcH96PX44556Y9yvXwtpxY0mbKqGqr/OaoQiG4lHsS0srJxHP+0da9asvh9X2wxaZbSwsFKFNT29ki/38uvq+zl+P9VGZRj361cFarOQOun3AlvFrbNhPb4QZbU9xNvatce3laQdp4gLWj8BaJDX7jfAlXF+CgD9UbCQWqlbKSJry7ro5gXCsnrhZHWXzbugDtMQ3usFu+jGdnWHHYyChVRW8mIyOxt6IY37Il/UljaFR2e31G4Bo8hfxFkXz7ir7KDjToq6+BZdJabusINRsJBKakJ1U94WX5z6GRvSmd7LL+K8X/C9duE1W92u4d59Lqmy56satPuqusMORsFCKqmoMQhV3uKLUz8X625BJ00vVS79TB5otvq5eXkv4te6ShbVkBUs1HVWxqqoLppV5t7fxH3u6elp71U8n9EVV6RPOb5r18p8R92mH0/LQ3LBo7SupXl561fR3VfVHbZgaRGkCZtKFuPR2f4Q9+pJ3k42rObV2zdpG3Ra8Hibnc2fC6vo/CZlzTNV5K919YYaPzSCW8rWbsM73tHb6Gbpz9q1YaR28r01yy6FFKFz9PKmTemLO5nBrbdqFb2m0AhuKVTadM67dilQlKHVghNPPP69LTNQrF27urqm3U4PFHE+sgKFpv1ujjXjzoDUT2cJ4uDB/MV/ZHCHDhX/3s7OZl/84fi5ppLtF2nHStO5zne89geoFFJHqoaSvlx0Edx117hzIcPqpQprdhaefz7c7jYZYnK/JE3OV0+qhpJM7XZYytNs9bZ1a1hVbdOmlTQFisHlrUI3Sr38RkyWPLpNVnj0aHp62ZMUymgpWEy4iy4K3S5ffPH4xw4ehBtu6F5dIf3pp/tqlezZszJNeqd+Z8Mtc2lWKU9N/3VlWGefrZLCqB05kr3uRBVNTa00SM/Pw7vedXzA6DZuQeMcmkXBYsK02+EL/4UvjDsnUnXLy6FBOg4YH/pQ6CLbaoX/oVYL9u7Nbqyenw+P97q/VJsauCfIli3wla+MOxdSRevXwz/8Q/qqer00SMcrGB46FKqZ9uxRUKgrNXBPsLjKSYGiGrLq/sfpNa9JDxSQ3yCdXMHQfaWLrMZUNIuCRUO122H9ZVU5VU8/hfn161caxaen4cILy+lV1a0TQ16D9O7d6fNSdRubIfWjYNFA11wTejhl/VKUamu1VmZd+ta34JZbQtryMtx/f6guGpXOkdxpskoeBw9q5HaTKFg0yDXXhJLEDTeMOycyjORAtnYbrrpqpYrnyJH0bs5lOfHE1W0PadN3dCt5pFVLaQqQmkqbXbAJ2yTNOru4WN7Mo9pGv01Nhc+0zKVle50BNznzbLeV9nqZ/Tae9VVLnVYbdV/PwswuNbNHzOyAmV077vxURTyoTppjeRmuvrq8wZAzM2HMRLf1KWLJUkNW28Rtt63uIpvl0CG1b9RZLYKFmU0DHwQuA84Cft7MzhpvrsZv40YNqmuq73ynvGPv3RvGTCQv8LOzsG7d6v06B9BltU3EQe2JJ0Kgy2qAn5vTFCB1VotxFmb2w8D73P2S6P57ANz9P2c9p47jLC644IKe9nvmGXj44XLzIs12wglwxhmwefNK2jPPwJe/DC+/nP74PfeEx7KOd9554RiPPBIqmJLM4Ad+YOX4Wc8fxt133z3cAQSo/ziLLcCTiftPRWmrmNlOM1sys6XnnntuZJkbpcceU6CQ4b38Mjz6aLi4xzZvDhfsN74x3H/44RAg4n3OOKP78e65Jzwn7ffn9HQ4/hlnHD8/1tRU92NLRaQ1ZFRtA94G/Eni/pXAB7o9p4kN3K9/ffmNq9oma+tcDnVh4fjG72QD9Pr16cfppcE8pqVOq42MBu66LH50GDg9cf+0KG1irFtXbj22TKZkW0G7DTfeGC7tSceOhVUQIXslxM7ndJqeXrk9P6+pQOqoLtVQ9wJnmtkZZrYOeDtw+5jzNDIbNypQSDnm5lbGPVxxRfZF/8iREDAG/T/UANH6q0XJwt1fMbN3A3cA08DN7v7QmLM1EjMz8NJL486FNJEZbN++eunTbobpyttqDf5cqYZaBAsAd98P7B93PkapihPOSXO4j2a0v9awaIa6VENNHAUKqZOpqTC+Ih6zEd/WGhbNoWBRMRdd1NxAsX79uHMgZVleDtWlt94Kzz8ftuXlMFBPgaIZFCwq5Oyzmzkie2oKFhbCDKpZdddTU7C4ONp8SbHiaTs0UWAzKVhUxMxM89adWFgI9eKvvhqml4DsdZlvuSX8AlVDaL3FM8xqIaTmUbCoALP69Xiang4lAffwNzkf0OxsSIsDRFLeuszbt48m/zKc5LiJTpoosJlqMTfUIOoyN1Qd2yfMQt10GXXRW7euXs9BqmdmBnbsgH37eutyC+F/Znm53HxJMeo+N1TjxAsV1Y1ZmN66rEZLzT5abbOzq2et7VbCSMpbmlWqT8FiDK65pp6r2c3OhhJFWvUSFNOwqYtKtW3YsPJDYX6+t9KCxlk0g4LFiLXb9QkUyb7yi4uhO2RWiaLdLqZhM60BXKqjc13tvOA+Pa1xFk2hNosROvvsevR4mpnp/wue1dbQaoW+9v1ot0OD6KFD4cKkeYWqKW67SJt8MKa2ivpRm8WYbdlSzUAR91zK6p3Ui3Y7u1F6kDaI+fmVVdf27VNJYxxmZ/Pf93hJ1Xe9K3sfVSs2R23mhqqzLVvgK18Zdy6ONzMD11033JTRcfVTlmEvFnG+duxQCaMsZqtLBvHyqseOhWqkbu/7kSNw/vnhdmcJQ20VzaKSRck2bhx/oJiZCaWHYUsQaXbvzu4+WdTFYn5eJYyyxIEi7tU0OxvuxzPM9hKgd+8OnR5uvbX4/y+pDrVZlOiii4qdvmN6OnyR8+qA16+HTZtCFdDcXLhgl/WlnZrKrq9eXCz2dcfZlpH89b1+PbzmNcNN2V1F8QJb/V4S1C7RLGqzGKF2OyxAX2SgmJqCk07q7Uv5R3+0Uudf9kRuWdVMrVbxr5tsyxj1xSl5AX3xxeYFCgir4A3y21HtEpNBwaJg7TZceWX28pODcoejR/P3m50dbdE/a66nsuuqdYGqBrVLTA4FiwK123DVVYP9OsszN5d/gYwbrEcpb66nsuzZU88R8E0wPa12iUmkYFGQdjusYVxE9UjnRTD+9bZnz0pPlU7xNAww+umhk9VDo1q/YH4+dNlUwBit6enQ2UBrVUweBYsCXHRRCBRFmJkJF8G0X+rz83DzzekzvD7/fLg/SdNDxz1wep2fqCxxwBp3PkbhpJMUICaWuzdyO/fcc71si4vuJ5zgHi7Ng29m4W+rFY45qFYr/fitVkEnXFGLi+4zM8N/Dr1srZb7wkL4a7b6M1tYGE0exrmZjfGDlpEAljzlmqpBeQOKB6O9/PJwx5mdXRkYN6ys0dJNn8k1fu927Sq3l9LCQvdJFPftK++1q0IdCyaXqqEG1G0wWq9mZ7tPztevrC/yJHzB5+dXquKGNTsLF164Uq00PR0CxfnnZ7cHFfH/UHVm6vk0yRQsBjTsr/W1a4vvuTSubqxNc+QI3H336l5tjz56fHvQlVeGC+ikLNhU5jomUn0KFgMa5tf67Cx8+MPlDFobRzfWJnr11ZWeba++GgZYdpYc4mAyCYECsqvgZDIoWAxokH7+ZvnrQgxrHN1YpflarXHnQMZNwWJAcT//frjr4l22qQn7j44XqCqTqjIFFCxyxUuFmq2MXDULE/VBaHvolX6dle+1rx13Dkbr6NFyOjDEjfuqypSYus52EXePjeuqk6Ozjxzpb3lU/TobjRdfLO/YF14IBw6ENoq0NSCKng+sF3Nz8P3fX2y7SfK8RGIqWXSxa9dw3SH162y02u1iqmRmZtK7zn7mM6EdyP34tRs6R9YXKeuc4q6sd99d3Gup9CtZVLLI0G4PPsBrkHWnZXi7dw//q7jXQZJZqwu+4x3FljBarexSQ9wGVuRUMyr9ShaVLDLs2jXY8/SFG59hxr60Wqt7qsVtVf1MyJg2d9cwLrww/OjYsCH98aJeR92spSdpc4A0YRtkbqjFxZU5f3qdK2fDhvR5gmT0subGytsWFlYfJ22uqZmZ/j7bouar6jbf1OxseK1hjy+SRMbcUFpWNdLZmN2LdevCL0n9GquGQT5DOL7aMGtEdr/Vi+027Ngx3PKv09PZz4+XMx20neass+ChhwbPmzSTllXN0evcPnHX2bhRU4GiOuIR7P3qrL4qakLG+fnhJxfsFmjiLrODVkeV2XNMmkfBItLLhWDt2tALRqOjq2t+vv8ePZ3jFIaZkLGzraMsyUn9rruuv/E+sabPRizFKi1YmNn7zOywmd0fbdsTj73HzA6Y2SNmdkki/dIo7YCZXZtIP8PMPhul/4WZZawXN7isC0FyCcky5nOS4qVNqJglrUPCoBMyxtVgnYtPFc1s9aR+8/Phf7NfkzAbsRQorSGjiA14H/BrKelnAZ8DTgDOAB4HpqPtceANwLpon7Oi59wGvD26fSOwkPf6/TZwF9GoKdWxuBgagDsbdNetC+l5HRKSnR167bgwaAN7vE1N5e/TLS/T070fW//bkoWMBu5xVENdDnzE3V929y8DB4A3R9sBd/+Su38b+AhwuZkZ8GPAR6Pn7wPeWnSmNGNrs8TrWywuHj947vnn86sSB5mQcZhqnZkZuOWWcCkftME6qxSzsBCOrf9tGUpaBCliI5QsngAeAG4GNkbpHwCuSOx3E/C2aPuTRPqV0b6bCEEkTj8deDDjNXcCS8DS3NxcSXFXJN2gJYv161eXdtJKRJ2lgm5Lu8YljOlpdY2V/lFGycLMPmNmD6ZslwM3AN8HvAl4Gvj9YV6rF+6+1923ufu2U045peyXkwnQz+C87duPLxXMzOT3VnrxxTBbgEftHN/4RuiWneXYMbjxxtVtI1dcESa3PP98eOWVkP7KK1qDQooz1HQf7n5RL/uZ2R8Dn4zuHiaUDmKnRWlkpB8BTjKzNe7+Ssf+IqXpHLeRbLDurMKJ1+D2xLAlszDO4vzz+5sG5DvfCQFmw4buU310OnIkO38iwyptUJ6ZneruT0e3/y3wQ+7+djM7G/gzQhvF64G7gDMBAx4FLiQEg3uBX3D3h8zsL4GPuftHzOxG4AF37/qbqd9BeSKdui2X2mqt9I7avbv7rK+tVmgn6WdcQzzgbpAlWzU3mQwja1BemRMJ/p6ZvQlwQtvFLwFEF//bgC8ArwC/7O6vRpl8N3AHoWfUze4ejy/9DeAjZvbbwP8jtHOIlKpbg/XBg6G04B5KAt0MMn143K11+/b+psIHjZ+QcpQWLNz9yi6P7QGO67Xu7vuB/SnpXyKURERGZm6u+4W+rPUr1q1bKbXsP+7bkE/jJ6QMGsEtEy+rEbufwX1F+va3Q4P1IFVQyZHdIkXSehYy0XppxN61a/C1TZLiXlG9HittRb7ksV56afV8Zp0ju0WKpJKFTLS0CSSPHQvpEC68WetJwPFzMs3MhEFwadOFXHddaOh2733+qrRBevGxOgeR3nqruspKeRQsZOIkq52yqnmSjcTdGow//OHjR0Z/6EP5swH0U8UVB5fOYw0yylxkUFrPQiZKr2texF1ju3WLHbaLarvdWxWXusLKKI2j66xI5fSybsnMTOiy2i2oDLt8brsd8nLkSCjhLC+n76cGa6kKVUPJRMkbgzA7G6p59u/PDhTDTsSXnMocugcKNVhLVahkIRMlb+zESy+Fv1lBxWz4KqFeSjdxoFCDtVSFShYyUfIaluOeUMOslpenlxHW7oMNyBMpi4KFTJTkuiVZDh0afLW8XvQacDRth1SJgoVMnLjLaVbAmJsrdzGsPXu6T0GezIdIVShYyMTKKz2UOY4hr8d6UaUYkaIoWMjEGtdSurt3p89UOz2tZU+lutQbSiZaPBJ6lLLaIpaXs7vRioybShYiI1ZmTyuRsihYiIxYmT2tRMqiYCEyYuNqKxEZhtosRMZgHG0lIsNQyUJERHIpWMhEyFo6VUR6o2ooabxelk4Vke5UspDGiksTV1zRfelUEcmnkoU0Ui8r4mmiPpHeqWQhjdTLmhEaBCfSOwULaaS8UoMGwYn0R8FCGqlbqUGD4ET6p2AhjZQ1pcbiYvHTjYtMAgULaSRNqSFSLPWGksbSlBoixVHJQkREcilYiIhILgULERHJpWAhE0+TDIrkUwO3TDRNMijSG5UsZKKlTQuiSQZFjjdUsDCznzGzh8xs2cy2dTz2HjM7YGaPmNklifRLo7QDZnZtIv0MM/tslP4XZrYuSj8hun8genzrMHkWScqaFkSTDIqsNmzJ4kHgp4G/Syaa2VnA24GzgUuBD5nZtJlNAx8ELgPOAn4+2hfgd4H3u/v3Ay8A74zS3wm8EKW/P9pPpBBZ04JokkGR1YYKFu7+RXd/JOWhy4GPuPvL7v5l4ADw5mg74O5fcvxXTPEAAAVXSURBVPdvAx8BLjczA34M+Gj0/H3AWxPH2hfd/ihwYbS/yNCypgXRJIMiq5XVZrEFeDJx/6koLSt9Fviau7/Skb7qWNHjX4/2P46Z7TSzJTNbeu655wo6FWkyTQsi0pvc3lBm9hnge1Me2u3unyg+S4Nz973AXoBt27b5mLMjNaFpQUTy5QYLd79ogOMeBk5P3D8tSiMj/QhwkpmtiUoPyf3jYz1lZmuA74n2FxGRESmrGup24O1RT6YzgDOB/wPcC5wZ9XxaR2gEv93dHfhb4G3R83cAn0gca0d0+23Af4/2FxGRERm26+xPmdlTwA8DnzKzOwDc/SHgNuALwN8Av+zur0alhncDdwBfBG6L9gX4DeBXzewAoU3ipij9JmA2Sv9V4LvdbUVEZDSsqT/St23b5ktLS+POhohIrZjZfe6+rTNdI7hFRCRXY0sWZvYccDDloU3A8yPOTtGacA7QjPNowjlAM86jCecA4z+Plruf0pnY2GCRxcyW0opYddKEc4BmnEcTzgGacR5NOAeo7nmoGkpERHIpWIiISK5JDBZ7x52BAjThHKAZ59GEc4BmnEcTzgEqeh4T12YhIiL9m8SShYiI9EnBQkREcjUqWJjZ+8zssJndH23bE48VsnLfuGXltyrM7Akz+3z0/i9FaSeb2Z1m9lj0d2OUbmZ2fXQuD5jZOYnj7Ij2f8zMdmS9XoH5vtnMnjWzBxNpheXbzM6N3pcD0XMLX5Ml4xxq9Z0ws9PN7G/N7AsWVuHcFaXX7bPIOo9afR6ruHtjNuB9wK+lpJ8FfA44ATgDeByYjrbHgTcA66J9zoqecxvw9uj2jcBCBc4vM79V2YAngE0dab8HXBvdvhb43ej2duCvAQPOAz4bpZ8MfCn6uzG6vbHkfP8ocA7wYBn5JkykeV70nL8GLhvROdTqOwGcCpwT3X4d8GiU17p9FlnnUavPI7k1qmTRRZEr941Tan7HnKdeJFc77FwF8RYP7iFMU38qcAlwp7sfdfcXgDsJy/OWxt3/DjhaRr6jx05093s8fLNvoYT/p4xzyFLJ74S7P+3u/ze6/U3ChKNbqN9nkXUeWSr5eSQ1MVi8OyqO3hwXVSl25b5xyspvlTjwaTO7z8x2Rmmb3f3p6PZXgc3R7X4/l1ErKt9botud6aNSy++EmW0FfhD4LDX+LDrOA2r6edQuWJjZZ8zswZTtcuAG4PuANwFPA78/1sxOpre4+znAZcAvm9mPJh+Mfs3Vrr92XfNNTb8TZrYB+Bjwb9z9G8nH6vRZpJxHLT8P6GGlvKrxHlfuM7M/Bj4Z3S1y5b5x6nYeleDuh6O/z5rZxwnF6GfM7FR3fzqqBng22j3rfA4DF3Sk311y1tMUle/D0e3O/Uvn7s/Et+vynTCztYQLbNvd/ypKrt1nkXYedfw8YrUrWXQT/RPFfgqIe4UUuXLfOKXmd8x5+i4zW29mr4tvAxcTPoPkaoedqyBeFfVoOQ/4elTVcAdwsZltjIrpF0dpo1ZIvqPHvmFm50V1zVcxov+nun0novfnJuCL7v4HiYdq9VlknUfdPo9Vymw9H/UG3Ap8HnggevNPTTy2m9Cr4BESvR8IvSkejR7bnUh/Q/RhHQD+Ejhh3OfXLb9V2KL37HPR9lCcP0L96l3AY8BngJOjdAM+GJ3L54FtiWO9I3rvDwBXjyDvf06oFvgOof73nUXmG9hGuDA8DnyAaPaEEZxDrb4TwFsIVUwPAPdH2/YafhZZ51GrzyO5aboPERHJ1ahqKBERKYeChYiI5FKwEBGRXAoWIiKSS8FCRERyKViIiEguBQsREcn1/wHllffUmJMKzQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.scatter(lreg_pred, (lreg_pred - y_test), c='b')\n",
    "\n",
    "plt.hlines(y=0, xmin= -1000, xmax=20000)\n",
    "\n",
    "plt.title('Residual plot')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### We can see there is Heteroskedasticity as the variance of the residuals get broader when x rises, so we will use a polynomial feature to avoid it"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Polynomial Regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import PolynomialFeatures\n",
    "poly_f = PolynomialFeatures(degree = 3) \n",
    "X_poly = poly_f.fit_transform(X_training) \n",
    "  \n",
    "polyreg = LinearRegression()\n",
    "polyreg.fit(X_poly, y_training) \n",
    "\n",
    "poly_pred = polyreg.predict(poly_f.fit_transform(X_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "r2: -170029681230.2417 rmse: 1632117691.5868967\n"
     ]
    }
   ],
   "source": [
    "poly_r2 = polyreg.score(poly_f.fit_transform(X_test), y_test)\n",
    "\n",
    "poly_rmse = mean_squared_error(y_test, poly_pred, sample_weight=None, multioutput='uniform_average', squared=False)\n",
    "\n",
    "print(\"r2:\",poly_r2,\"rmse:\",poly_rmse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'Residual plot')"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXwAAAEVCAYAAADjHF5YAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+j8jraAAAZK0lEQVR4nO3df5DcdZ3n8ecrxOBlOeRXKkRCEneNq9FyYW05gUUoCbeASsCCE3Zwwx5Wds/11i1lLbxc3V65RxXKrXIU7pWzsEXUWUFRJAtxQ4jcgXfgMRFEQhYTUSAQyAiisLnTZfO6P77fSDP0zHRPd89M5vN6VHX198env5/3JD2v/synu79f2SYiIma/OdNdQERETI0EfkREIRL4ERGFSOBHRBQigR8RUYgEfkREIRL4URxJA5JuG2f//5D0wR70c4qknZN87EWSvt1tDRHNEvgxo0n6saT/K+kFSU9Juk7SQd0c0/aQ7X/dqxqnW69eoGL2S+DH/uC9tg8CjgGOBT4xzfVE7JcS+LHfsP0UsJEq+AGQ9A5J/1vSc5K+J+mUpn0XSXpE0vOSfiRpoGn7t5vanSbpHyT9TNLVgJr2/WdJX2paXybJkubW638gaVvdxyOS/rDdn6c+zp/Uj/uJpCsktfydlHSCpHvrGu+VdEK9/TLgJODq+q+gq9vtP8qTwI/9hqTFwBnAjnr9KOBW4L8AhwGXAF+TtEDSrwFXAWfY/pfACcD9LY55BPB14D8CRwA/BE7soKzdwHuAg4E/AD4r6bc7ePw5QAP4bWAV8G9b1HgY1c95FXA48BngVkmH214L3AV82PZBtj/cQd9RmBkf+JL+RtJuSQ+20fadkr4r6UVJ547a9/f1KPCW/lUbffINSc8Dj1MF7J/X2y8ENtjeYHuv7U3AMHBmvX8v8BZJ/8L2LttbWxz7TGCr7Rtt/xNwJfBUu4XZvtX2D135n8BtVCPudn3K9rO2H6v7vqBFm3cD221/0faLtr8M/APw3g76iZj5gQ9cB5zeZtvHgIuAv22x7wrgA70pKabY2fUo/RTgjVQjcYClwHn1C/lzkp4DfgdYZPsfgfcDfwTsknSrpDe2OPZrqV5IAHB1NsHHW7RrSdIZku6R9Gzd/5lN9bWjua9H63pa1fjoqG2PAkd10E/EzA9823cCzzZvk/Qb9Yh9i6S79v0i2/6x7QeoRnajj7MZeH5Kio6+qEfQ1wH/td70OPBF24c03X7N9uV1+422TwMWUY2I/7rFYXcBR+9bkaTmdeAfgflN60c2tT0Q+Fpdz0LbhwAbaHoPoA3NfS0BnmzR5kmqFzdGtX2iXs4pb6MtMz7wxzAI/Hvbb6Oat/2raa4nps6VwGmSfgv4EvBeSb8r6QBJr64/+75Y0kJJq+q5/F8AL9BiIEA1N/5mSe+r34j9E5pCnWre/52Slkh6DS//hNA84EBgBHhR0hlApx/3/DNJh0o6GvgIcEOLNhuAN0j6PUlzJb0fWAHsm558Gvj1DvuNAu13gV9/BvsE4KuS7gc+TzWCiwLYHgG+APwn249TvdH5H6hC93Hgz6ie13OAj1KNjp8FTgb+XYvj/QQ4D7gceAZYDvyvpv2bqEL4AWALL4Ustp+neoH4CvBT4PeA9R3+SDfXx72f6sXn2hY1PkP1xvDH6ho/Drynrh3gvwHnSvqppKs67D8Kov3hAiiSlgG32H6LpIOBh22PGfKSrqvb3zhq+ynAJbbf079qI9ojycBy2zumu5Yow343wrf9c+BHks6Das61/vM+IiLGMeMDX9KXgbuB35S0U9LFwABwsaTvAVup/qxH0tvrc5ecB3xe0tam49wFfBU4tT7O7071zxIRMZ32iymdiIjo3owf4UdERG/Mne4CxnPEEUd42bJl011GRMR+Y8uWLT+xvaDVvhkd+MuWLWN4eHi6y4iI2G9IGv2t7F/pyZSOpNMlPSxph6RLW+z/qKSHJD0gabOk0d8ajIiIPus68CUdAHyO6iyGK4ALJK0Y1ew+oGH7rcCNwKe77TciIjrTixH+ccAO24/Y/iVwPfXHJPexfYftPfXqPcDiHvQbEREd6EXgH8XLz/i3k/HP4ncx8M2xdkpaI2lY0vDIyEgPyouICJjij2VKupDqYg9XjNXG9qDthu3GggUt32iOiIhJ6EXgP8HLT/G6mJdO2/orklYCa4GzbP+iB/1GRMwqQ0OwbBnMmVPdDw319vi9+FjmvcBySa+jCvrzqc4a+CuSjqU6q+Xptnf3oM+IiFllaAjWrIE99budjz5arQMMDPSmj65H+LZfBD5MdXHpbcBXbG+V9ElJZ9XNrgAOoj6lsaROTyEbETGrrV37Utjvs2dPtb1XZvS5dBqNhvPFq4gowZw50CqOJdjb6tI9Y5C0xXajZR+TLS4iInpnyZLOtk9GAj8iYga47DKYP//l2+bPr7b3SgI/ImIGGBiAwUFYurSaxlm6tFrv1Ru2MMNPnhYRUZKBgd4G/GgZ4UdEFCKBHxFRiAR+REQhEvgREYVI4EdEFCKBHxFRiAR+REQhEvgREYVI4EdEFCKBHxFRiAR+REQhEvgREYXoSeBLOl3Sw5J2SLq0xf4DJd1Q7/+OpGW96Dcioh+kV95mg64DX9IBwOeAM4AVwAWSVoxqdjHwU9uvBz4LfKrbfiMi+mGscJ8Nod+LEf5xwA7bj9j+JXA9sGpUm1XAunr5RuBUaTb880VE7D96EfhHAY83re+st7VsU1/0/GfA4a0OJmmNpGFJwyMjIz0oLyIiYAa+aWt70HbDdmPBggXTXU5ExKzRi8B/Aji6aX1xva1lG0lzgdcAz/Sg74iIjg0NwbJlMGdOdT80NN0VTY1eBP69wHJJr5M0DzgfWD+qzXpgdb18LvAt2+5B3xERHRkagjVr4NFHwa7u16x5KfTHSqbZkFhdB349J/9hYCOwDfiK7a2SPinprLrZtcDhknYAHwVe8dHNiIipsHYt7Nnz8m179lTb97FfeZsNNJMH2o1Gw8PDw9NdRkTMInPmtA5wCfbunfp6ek3SFtuNVvtm3Ju2ERH9tGRJZ9tnkwR+RBTlsstg/vyXb5s/v9o+2yXwI6IoAwMwOAhLl1bTOEuXVusDA9NdWf/Nne4CIiKm2sBAGQE/Wkb4ERGFSOBHRBQigR8RUYgEfkREIRL4ERGFSOBHRBQigR8RUYgEfkREIRL4ERGFSOBHRBQigR8RUYiuAl/SYZI2Sdpe3x/aos0xku6WtFXSA5Le302fERExOd2O8C8FNtteDmym9ZWs9gC/b/vNwOnAlZIO6bLfiIjoULeBvwpYVy+vA84e3cD2D2xvr5efBHYDC7rsNyIiOtRt4C+0vatefgpYOF5jSccB84AfjtNmjaRhScMjIyNdlhcREftMeD58SbcDR7bYtbZ5xbYljXmBXEmLgC8Cq22PeeVI24PAIFTXtJ2ovoiIaM+EgW975Vj7JD0taZHtXXWg7x6j3cHArcBa2/dMutqIiJi0bqd01gOr6+XVwM2jG0iaB9wEfMH2jV32FxERk9Rt4F8OnCZpO7CyXkdSQ9I1dZt/A7wTuEjS/fXtmC77jYiIDsmeudPkjUbDw8PD011GRMR+Q9IW241W+/JN24iIQiTwIyIKkcCPiChEAj8iohAJ/IiIQiTwIyIKkcCPiChEAj8iohAJ/IiIQiTwIyIKkcCPiChEAj8iohAJ/IiIQiTwIyIKkcCPiChE14Ev6TBJmyRtr+8PHaftwZJ2Srq6234jIqIzvRjhXwpstr0c2Fyvj+UvgDt70GdERHSoF4G/ClhXL68Dzm7VSNLbgIXAbT3oMyIiOtSLwF9oe1e9/BRVqL+MpDnAXwKX9KC/iIiYhLntNJJ0O3Bki11rm1dsW1Kri+R+CNhge6ekifpaA6wBWLJkSTvlRUREG9oKfNsrx9on6WlJi2zvkrQI2N2i2fHASZI+BBwEzJP0gu1XzPfbHgQGobqIeTv1RUTExNoK/AmsB1YDl9f3N49uYHtg37Kki4BGq7CPiIj+6cUc/uXAaZK2AyvrdSQ1JF3Tg+NHREQPyJ65syaNRsPDw8PTXUZExH5D0hbbjVb78k3biIhCJPAjIgqRwI+IKEQCPyKiEAn8iIhCJPAjIgqRwI+IKEQCPyKiEAn8iIhCJPAjIgqRwI+IKEQCPyKiEAn8iIhCJPAjIgqRwI+IKEQCPyKiEF0FvqTDJG2StL2+P3SMdksk3SZpm6SHJC3rpt+IiOhctyP8S4HNtpcDm+v1Vr4AXGH7TcBxtL7QeURE9FG3gb8KWFcvrwPOHt1A0gpgru1NALZfsL2ny34jIqJD3Qb+Qtu76uWngIUt2rwBeE7S1yXdJ+kKSQeMdUBJayQNSxoeGRnpsryIiNhn7kQNJN0OHNli19rmFduW1OqK6HOBk4BjgceAG4CLgGtb9Wd7EBiE6iLmE9UXERHtmTDwba8ca5+kpyUtsr1L0iJaz83vBO63/Uj9mG8A72CMwI+IiP7odkpnPbC6Xl4N3Nyizb3AIZIW1OvvAh7qst+IiOhQt4F/OXCapO3AynodSQ1J1wDY/mfgEmCzpO8DAv66y34jIqJDE07pjMf2M8CpLbYPAx9sWt8EvLWbviIiojv5pm1ERCES+BERhUjgR0QUIoEfEVGIBH5ERCES+BERhUjgR0QUIoEfEVGIBH5ERCES+BERhUjgR0QUIoEfEVGIBH5ERCES+BERhUjgR0QUouvAl3SYpE2Sttf3h47R7tOStkraJukqSeq274iIaF8vRviXApttLwc21+svI+kE4ESqi6C8BXg7cHIP+o6IiDb1IvBXAevq5XXA2S3aGHg1MA84EHgV8HQP+o6IiDb1IvAX2t5VLz8FLBzdwPbdwB3Arvq20fa2HvQdERFtauuatpJuB45ssWtt84ptS3KLx78eeBOwuN60SdJJtu9q0XYNsAZgyZIl7ZQXERFtaCvwba8ca5+kpyUtsr1L0iJgd4tm5wD32H6hfsw3geOBVwS+7UFgEKDRaLzixSMiIianF1M664HV9fJq4OYWbR4DTpY0V9KrqN6wzZRORMQU6kXgXw6cJmk7sLJeR1JD0jV1mxuBHwLfB74HfM/23/Wg74iIaFNbUzrjsf0McGqL7cPAB+vlfwb+sNu+IiJi8vJN24iIQiTwIyIKkcCPiChEAj8iohAJ/IiIQiTwIyIKkcCPiChEAj8iohAJ/IiIQiTwIyIKkcCPiChEAj8iohAJ/IiIQiTwIyIKkcCPiChEAj8iohBdBb6k8yRtlbRXUmOcdqdLeljSDkmXdtNnRERMTrcj/AeB9wF3jtVA0gHA54AzgBXABZJWdNlvRER0qKtLHNreBiBpvGbHATtsP1K3vR5YBTzUTd8REdGZqZjDPwp4vGl9Z72tJUlrJA1LGh4ZGel7cRERpZhwhC/pduDIFrvW2r651wXZHgQGARqNhnt9/IiIUk0Y+LZXdtnHE8DRTeuL620RETGFpmJK515guaTXSZoHnA+sn4J+IyKiSbcfyzxH0k7geOBWSRvr7a+VtAHA9ovAh4GNwDbgK7a3dld2RER0qttP6dwE3NRi+5PAmU3rG4AN3fQVERHdyTdtIyIKkcCPiChEAj8iohAJ/IiIQiTwIyIKkcCPiChEAj8iohAJ/IiIQiTwIyIKkcCPiChEAj8iohAJ/IiIQiTwIyIKkcCPiChEAj8iohDdXgDlPElbJe2V1BijzdGS7pD0UN32I930GRERk9PtCP9B4H3AneO0eRH4mO0VwDuAP5a0ost+IyKiQ91e8WobgKTx2uwCdtXLz0vaBhwFPNRN3xER0ZkpncOXtAw4FvjOVPYbERFtjPAl3Q4c2WLXWts3t9uRpIOArwF/avvn47RbA6wBWLJkSbuHj4iICUwY+LZXdtuJpFdRhf2Q7a9P0N8gMAjQaDTcbd8REVHp+5SOqgn+a4Fttj/T7/4iIqK1bj+WeY6kncDxwK2SNtbbXytpQ93sROADwLsk3V/fzuyq6oiI6Fi3n9K5CbipxfYngTPr5W8DY3+MJyIipkS+aRsRUYgEfkREIRL4ERGFSOBHRBQigR8RUYgEfkREIRL4ERGFSOBHRBQigR8RUYgEfkREIRL4ERGFSOBHRBQigR8RUYgEfkREIRL4ERGFSOBHRBSi2ytenSdpq6S9khoTtD1A0n2Sbummz4iImJxuR/gPAu8D7myj7UeAbV32FxERk9RV4NveZvvhidpJWgy8G7imm/4iImLypmoO/0rg48DeiRpKWiNpWNLwyMhI/yuLiCjEhIEv6XZJD7a4rWqnA0nvAXbb3tJOe9uDthu2GwsWLGjnIRER0Ya5EzWwvbLLPk4EzpJ0JvBq4GBJX7J9YZfHjYiIDvR9Ssf2J2wvtr0MOB/4VsI+ImLqdfuxzHMk7QSOB26VtLHe/lpJG3pRYKeGhmDZMpgzp7ofGpqOKiIiZh7Znu4axtRoNDw8PNx2+6EhWLMG9ux5adv8+TA4CAMDfSgwImKGkbTFdsvvRc2qb9quXfvysIdqfe3a6aknImImmVWB/9hjnW2PiCjJrAr8JUs62x4RUZJZFfiXXVbN2TebP7/aHhFRulkV+AMD1Ru0S5eCVN3nDduIiMqEX7za3wwMJOAjIlqZVSP8iIgYWwI/IqIQCfyIiEIk8CMiCpHAj4goxIw+l46kEeDRPndzBPCTPvcxGamrfTOxJkhdnZiJNcH+WddS2y0vJjKjA38qSBoe60RD0yl1tW8m1gSpqxMzsSaYfXVlSiciohAJ/IiIQiTwYXC6CxhD6mrfTKwJUlcnZmJNMMvqKn4OPyKiFBnhR0QUIoEfEVGI4gJf0nmStkraK2ncjzVJOkDSfZJumQl1STpa0h2SHqrbfmQm1FW3O13Sw5J2SLp0Cuo6TNImSdvr+0PHaPfpuv5tkq6SpBlQ0xJJt9U1PSRpWb9q6qSuuu3BknZKunq6a5J0jKS76/+/ByS9v4/1jPv8lXSgpBvq/d/p9/9ZB3V9tH4OPSBps6Sl4x2vuMAHHgTeB9zZRtuPANv6W86vtFPXi8DHbK8A3gH8saQV012XpAOAzwFnACuAC6agrkuBzbaXA5vr9dF1nQCcCLwVeAvwduDk6ayp9gXgCttvAo4Ddvexpk7qAvgL2vvdmIqa9gC/b/vNwOnAlZIO6XUhbT5/LwZ+avv1wGeBT/W6jknWdR/QsP1W4Ebg0+Mds7jAt73N9sMTtZO0GHg3cE3/q2qvLtu7bH+3Xn6e6sXoqOmuiyq0dth+xPYvgeuBVf2sqz7+unp5HXB2izYGXg3MAw4EXgU8PZ011b+wc21vArD9gu09fayprbrq2t4GLARu63M9bdVk+we2t9fLT1K9MLb8BmmX2nn+Ntd7I3BqP/9abLcu23c0PX/uARaPd8DiAr8DVwIfB/ZOdyGt1H9SHgt8Z3orAaoXnceb1nfS5xciYKHtXfXyU1RB9TK27wbuAHbVt422+/kX24Q1AW8AnpP09Xq68Ip6JNdPE9YlaQ7wl8Alfa6l7ZqaSTqO6oX7h32opZ3n76/a2H4R+BlweB9q6bSuZhcD3xzvgLPuilcAkm4Hjmyxa63tm9t4/HuA3ba3SDplptTVdJyDgK8Bf2r75zOlrl4br67mFduW9IrPF0t6PfAmXhr1bJJ0ku27pqsmqt+5k6herB8DbgAuAq6dbE09qutDwAbbO3s1cO1BTfuOswj4IrDa9owcgE03SRcCDSaYspyVgW97ZZeHOBE4S9KZVFMCB0v6ku0Lp7kuJL2KKuyHbH+92+NBT+p6Aji6aX1xva0r49Ul6WlJi2zvqgOh1Tz4OcA9tl+oH/NN4Hhg0oHfg5p2AvfbfqR+zDeo3o/pKvB7UNfxwEmSPgQcBMyT9ILtSb8B34OakHQwcCvV4OOeydYygXaev/va7JQ0F3gN8Eyf6umkLiStpHoRPdn2L8Y7YKZ0WrD9CduLbS8Dzge+1W3Y90I9Z3gtsM32Z6a7nib3AsslvU7SPKp/s/V97nM9sLpeXg20+kvkMeBkSXPrF8qT6e+b8O3UdC9wiKR9c9HvAh7qY01t1WV7wPaS+jl/CfCFbsK+FzXVz6Wb6lpu7GMt7Tx/m+s9lyoT+v2t1QnrknQs8HngLNsTv/lvu6gb1ahvJ/ALqjfwNtbbX0v1J+3o9qcAt8yEuoDfoXoj8gHg/vp25nTXVa+fCfyAao517RT8ex1O9emO7cDtwGH19gZwTb18QP3LsI0qVD8z3TXV66fV/4ffB64D5s2EupraXwRcPd01ARcC/9T0XL8fOKZP9bzi+Qt8kipIofpL/6vADuD/AL/e7+d4m3XdXv9e7vv3WT/e8XJqhYiIQmRKJyKiEAn8iIhCJPAjIgqRwI+IKEQCPyJiCkn6G0m7JT3YRtt3SvqupBclnTtq399Lek4dnNwxgR8RMbWuozoZXDseo/qY7N+22HcF8IFOOk7gR0RMIdt3As82b5P0G/WIfYukuyS9sW77Y9sP0OKcXrY3A8930vesPLVCRMR+ZhD4I9vbJf0r4K+ovoXdUwn8iIhpVJ8M8QTgq00nrjuwH30l8CMiptcc4Dnbx0xFRxERMU1cneL8R5LOg+okiZJ+qx995Vw6ERFTSNKXqU7KeATVic/+HPgW8N+BRVRXZrve9iclvZ3qjKGHAv8PeMrVJR+RdBfwRqrTWT8DXGx747h9J/AjIsqQKZ2IiEIk8CMiCpHAj4goRAI/IqIQCfyIiEIk8CMiCpHAj4goxP8Hv772u87b9UkAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.scatter(poly_pred, (poly_pred - y_test), c='b')\n",
    "\n",
    "plt.hlines(y=0, xmin= -1000, xmax=20000)\n",
    "\n",
    "plt.title('Residual plot')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Well, we see that actually we have less Heteroskedasticity, but our model has become less accurate as it has some big outliers"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### After applying a polynomial feature just made our results worse, we will check if we can improve them with regularization (remember our model had Multicollinearity)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7fb768bf41d0>"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "predictors = X_training.columns\n",
    "\n",
    "coef = pd.Series(lreg.coef_,predictors).sort_values()\n",
    "\n",
    "coef.plot(kind='bar', title='Modal Coefficients')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Ridge Regression "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.linear_model import Ridge\n",
    "\n",
    "ridreg = Ridge(fit_intercept=True, normalize=False, alpha=0.01)\n",
    "\n",
    "ridreg.fit(X_training, y_training) \n",
    "\n",
    "ridreg_pred = ridreg.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "r2: 0.9148719487089647 rmse: 1154.8496525732999\n"
     ]
    }
   ],
   "source": [
    "rid_r2 = ridreg.score(X_test, y_test)\n",
    "\n",
    "ridreg_rmse = mean_squared_error(y_test, ridreg_pred, sample_weight=None, multioutput='uniform_average', squared=False)\n",
    "\n",
    "print(\"r2:\",rid_r2,\"rmse:\",ridreg_rmse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7fb76dbee550>"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "predictors = X_training.columns\n",
    "\n",
    "coef = pd.Series(ridreg.coef_,predictors).sort_values()\n",
    "\n",
    "coef.plot(kind='bar', title='Modal Coefficients')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Lasso Regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.linear_model import Lasso\n",
    "\n",
    "lassreg = Lasso(fit_intercept=True, normalize=True, alpha=0.01)\n",
    "\n",
    "lassreg.fit(X_training, y_training) \n",
    "\n",
    "lassreg_pred = lassreg.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "r2: 0.9149012157663253 rmse: 1154.6511166263156\n"
     ]
    }
   ],
   "source": [
    "lass_r2 = lassreg.score(X_test, y_test)\n",
    "\n",
    "lassreg_rmse = mean_squared_error(y_test, lassreg_pred, sample_weight=None, multioutput='uniform_average', squared=False)\n",
    "\n",
    "print(\"r2:\",lass_r2,\"rmse:\",lassreg_rmse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7fb768ce5250>"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "predictors = X_training.columns\n",
    "\n",
    "coef = pd.Series(lassreg.coef_,predictors).sort_values()\n",
    "\n",
    "coef.plot(kind='bar', title='Modal Coefficients')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Elastic Net Regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.linear_model import ElasticNet\n",
    "\n",
    "enetreg = ElasticNet(fit_intercept=True, normalize=False, alpha=0.01, l1_ratio=0.5)\n",
    "\n",
    "enetreg.fit(X_training, y_training) \n",
    "\n",
    "enetreg_pred = enetreg.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "r2: 0.9149108823041231 rmse: 1154.5855352214564\n"
     ]
    }
   ],
   "source": [
    "enet_r2 = enetreg.score(X_test, y_test)\n",
    "\n",
    "enet_rmse = mean_squared_error(y_test, enetreg_pred, sample_weight=None, multioutput='uniform_average', squared=False)\n",
    "\n",
    "print(\"r2:\",enet_r2,\"rmse:\",enet_rmse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7fb72d854a50>"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "predictors = X_training.columns\n",
    "\n",
    "coef = pd.Series(enetreg.coef_,predictors).sort_values()\n",
    "\n",
    "coef.plot(kind='bar', title='Modal Coefficients')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### LARS"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.linear_model import LassoLars\n",
    "\n",
    "LARSreg = LassoLars(fit_intercept=True, normalize=False, alpha=0.01)\n",
    "\n",
    "LARSreg.fit(X_training, y_training) \n",
    "\n",
    "LARSreg_pred = LARSreg.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "r2: 0.9148721307818831 rmse: 1154.8484175693509\n"
     ]
    }
   ],
   "source": [
    "LARS_r2 = LARSreg.score(X_test, y_test)\n",
    "\n",
    "LARS_rmse = mean_squared_error(y_test, LARSreg_pred, sample_weight=None, multioutput='uniform_average', squared=False)\n",
    "\n",
    "print(\"r2:\",LARS_r2,\"rmse:\",LARS_rmse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7fb72d74e450>"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "predictors = X_training.columns\n",
    "\n",
    "coef = pd.Series(LARSreg.coef_,predictors).sort_values()\n",
    "\n",
    "coef.plot(kind='bar', title='Modal Coefficients')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Bayessian Ridge"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.linear_model import BayesianRidge\n",
    "\n",
    "bayreg = BayesianRidge()\n",
    "\n",
    "bayreg.fit(X_training, y_training) \n",
    "\n",
    "bayreg_pred = bayreg.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "r2: 0.9148733674942721 rmse: 1154.8400288915554\n"
     ]
    }
   ],
   "source": [
    "bay_r2 = bayreg.score(X_test, y_test)\n",
    "\n",
    "bay_rmse = mean_squared_error(y_test, bayreg_pred, sample_weight=None, multioutput='uniform_average', squared=False)\n",
    "\n",
    "print(\"r2:\",bay_r2,\"rmse:\",bay_rmse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7fb72d863350>"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "predictors = X_training.columns\n",
    "\n",
    "coef = pd.Series(bayreg.coef_,predictors).sort_values()\n",
    "\n",
    "coef.plot(kind='bar', title='Modal Coefficients')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Robust Regressor: Huber"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.linear_model import HuberRegressor\n",
    "\n",
    "hubreg = HuberRegressor()\n",
    "\n",
    "hubreg.fit(X_training, y_training) \n",
    "\n",
    "hubreg_pred = hubreg.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "r2: 0.9039665574904512 rmse: 1226.5926295918896\n"
     ]
    }
   ],
   "source": [
    "hub_r2 = hubreg.score(X_test, y_test)\n",
    "\n",
    "hub_rmse = mean_squared_error(y_test, hubreg_pred, sample_weight=None, multioutput='uniform_average', squared=False)\n",
    "\n",
    "print(\"r2:\",hub_r2,\"rmse:\",hub_rmse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7fb7688e3b50>"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "predictors = X_training.columns\n",
    "\n",
    "coef = pd.Series(hubreg.coef_,predictors).sort_values()\n",
    "\n",
    "coef.plot(kind='bar', title='Modal Coefficients')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Linear SVR"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.svm import LinearSVR\n",
    "\n",
    "svrReg = LinearSVR()\n",
    "\n",
    "svrReg.fit(X_training, y_training) \n",
    "\n",
    "svrReg_pred = svrReg.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "r2: 0.8715669116060938 rmse: 1418.49420244425\n"
     ]
    }
   ],
   "source": [
    "svr_r2 = svrReg.score(X_test, y_test)\n",
    "\n",
    "svr_rmse = mean_squared_error(y_test, svrReg_pred, sample_weight=None, multioutput='uniform_average', squared=False)\n",
    "\n",
    "print(\"r2:\",svr_r2,\"rmse:\",svr_rmse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7fb72d65da90>"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "predictors = X_training.columns\n",
    "\n",
    "coef = pd.Series(svrReg.coef_,predictors).sort_values()\n",
    "\n",
    "coef.plot(kind='bar', title='Modal Coefficients')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### We can see even following all the steps, the linear models don't look like good predictors for our dataset with dummy encoding"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(13449, 10)\n"
     ]
    },
    {
     "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>Unnamed: 0</th>\n",
       "      <th>carat</th>\n",
       "      <th>cut</th>\n",
       "      <th>color</th>\n",
       "      <th>clarity</th>\n",
       "      <th>depth</th>\n",
       "      <th>table</th>\n",
       "      <th>x</th>\n",
       "      <th>y</th>\n",
       "      <th>z</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0.30</td>\n",
       "      <td>Ideal</td>\n",
       "      <td>H</td>\n",
       "      <td>SI2</td>\n",
       "      <td>60.0</td>\n",
       "      <td>56.0</td>\n",
       "      <td>4.41</td>\n",
       "      <td>4.43</td>\n",
       "      <td>2.65</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0.34</td>\n",
       "      <td>Ideal</td>\n",
       "      <td>D</td>\n",
       "      <td>IF</td>\n",
       "      <td>62.1</td>\n",
       "      <td>57.0</td>\n",
       "      <td>4.52</td>\n",
       "      <td>4.46</td>\n",
       "      <td>2.79</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>1.57</td>\n",
       "      <td>Very Good</td>\n",
       "      <td>I</td>\n",
       "      <td>VS2</td>\n",
       "      <td>60.3</td>\n",
       "      <td>58.0</td>\n",
       "      <td>7.58</td>\n",
       "      <td>7.55</td>\n",
       "      <td>4.56</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>0.31</td>\n",
       "      <td>Ideal</td>\n",
       "      <td>H</td>\n",
       "      <td>VS2</td>\n",
       "      <td>61.8</td>\n",
       "      <td>57.0</td>\n",
       "      <td>4.32</td>\n",
       "      <td>4.36</td>\n",
       "      <td>2.68</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>1.51</td>\n",
       "      <td>Good</td>\n",
       "      <td>I</td>\n",
       "      <td>VVS1</td>\n",
       "      <td>64.0</td>\n",
       "      <td>60.0</td>\n",
       "      <td>7.26</td>\n",
       "      <td>7.21</td>\n",
       "      <td>4.63</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Unnamed: 0  carat        cut color clarity  depth  table     x     y     z\n",
       "0           0   0.30      Ideal     H     SI2   60.0   56.0  4.41  4.43  2.65\n",
       "1           1   0.34      Ideal     D      IF   62.1   57.0  4.52  4.46  2.79\n",
       "2           2   1.57  Very Good     I     VS2   60.3   58.0  7.58  7.55  4.56\n",
       "3           3   0.31      Ideal     H     VS2   61.8   57.0  4.32  4.36  2.68\n",
       "4           4   1.51       Good     I    VVS1   64.0   60.0  7.26  7.21  4.63"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "testdf = pd.read_csv(\"INPUT/diamonds_test.csv\")\n",
    "print(testdf.shape)\n",
    "testdf.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Compute Volume"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "testdf[\"volume\"] = testdf.x * testdf.y * testdf.z\n",
    "index = testdf[\"Unnamed: 0\"]\n",
    "testdf.drop(columns={\"Unnamed: 0\",\"x\",\"y\",\"z\"}, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_dum = pd.get_dummies(testdf)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>carat</th>\n",
       "      <th>depth</th>\n",
       "      <th>table</th>\n",
       "      <th>volume</th>\n",
       "      <th>cut_Fair</th>\n",
       "      <th>cut_Good</th>\n",
       "      <th>cut_Ideal</th>\n",
       "      <th>cut_Premium</th>\n",
       "      <th>cut_Very Good</th>\n",
       "      <th>color_D</th>\n",
       "      <th>...</th>\n",
       "      <th>color_I</th>\n",
       "      <th>color_J</th>\n",
       "      <th>clarity_I1</th>\n",
       "      <th>clarity_IF</th>\n",
       "      <th>clarity_SI1</th>\n",
       "      <th>clarity_SI2</th>\n",
       "      <th>clarity_VS1</th>\n",
       "      <th>clarity_VS2</th>\n",
       "      <th>clarity_VVS1</th>\n",
       "      <th>clarity_VVS2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.30</td>\n",
       "      <td>60.0</td>\n",
       "      <td>56.0</td>\n",
       "      <td>51.771195</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.34</td>\n",
       "      <td>62.1</td>\n",
       "      <td>57.0</td>\n",
       "      <td>56.244168</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.57</td>\n",
       "      <td>60.3</td>\n",
       "      <td>58.0</td>\n",
       "      <td>260.964240</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.31</td>\n",
       "      <td>61.8</td>\n",
       "      <td>57.0</td>\n",
       "      <td>50.478336</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1.51</td>\n",
       "      <td>64.0</td>\n",
       "      <td>60.0</td>\n",
       "      <td>242.355498</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 24 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   carat  depth  table      volume  cut_Fair  cut_Good  cut_Ideal  \\\n",
       "0   0.30   60.0   56.0   51.771195         0         0          1   \n",
       "1   0.34   62.1   57.0   56.244168         0         0          1   \n",
       "2   1.57   60.3   58.0  260.964240         0         0          0   \n",
       "3   0.31   61.8   57.0   50.478336         0         0          1   \n",
       "4   1.51   64.0   60.0  242.355498         0         1          0   \n",
       "\n",
       "   cut_Premium  cut_Very Good  color_D  ...  color_I  color_J  clarity_I1  \\\n",
       "0            0              0        0  ...        0        0           0   \n",
       "1            0              0        1  ...        0        0           0   \n",
       "2            0              1        0  ...        1        0           0   \n",
       "3            0              0        0  ...        0        0           0   \n",
       "4            0              0        0  ...        1        0           0   \n",
       "\n",
       "   clarity_IF  clarity_SI1  clarity_SI2  clarity_VS1  clarity_VS2  \\\n",
       "0           0            0            1            0            0   \n",
       "1           1            0            0            0            0   \n",
       "2           0            0            0            0            1   \n",
       "3           0            0            0            0            1   \n",
       "4           0            0            0            0            0   \n",
       "\n",
       "   clarity_VVS1  clarity_VVS2  \n",
       "0             0             0  \n",
       "1             0             0  \n",
       "2             0             0  \n",
       "3             0             0  \n",
       "4             1             0  \n",
       "\n",
       "[5 rows x 24 columns]"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_dum.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "Xss_test = sc_X.fit_transform(test_dum)\n",
    "\n",
    "test_SS = pd.DataFrame(Xss_test, columns=['carat', 'depth', 'table', 'volume', 'cut_Fair', 'cut_Good',\n",
    "       'cut_Ideal', 'cut_Premium', 'cut_Very Good', 'color_D', 'color_E',\n",
    "       'color_F', 'color_G', 'color_H', 'color_I', 'color_J', 'clarity_I1',\n",
    "       'clarity_IF', 'clarity_SI1', 'clarity_SI2', 'clarity_VS1',\n",
    "       'clarity_VS2', 'clarity_VVS1', 'clarity_VVS2'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Now we train our models again but with 100% of the data and take the best"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Surprisingly, after applying the 100%, the polynomial model has the best rmse by far out of all the models. However the Polynomial in reallity doesn't make good predictions in this case as almost all predictions given are negative (see the modal coefficient graph). Therefore we will use Elastic Net Model (the 2nd best rmse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "#test_pred = enetreg.predict(test_SS)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'test = pd.DataFrame(test_pred, columns={\"price\"})\\ntest[\"id\"] = index\\ntest = test[[\"id\",\"price\"]]'"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\"\"\"test = pd.DataFrame(test_pred, columns={\"price\"})\n",
    "test[\"id\"] = index\n",
    "test = test[[\"id\",\"price\"]]\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "#test.to_csv(\"OUTPUT/elastic_net_dummies.csv\", index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### As conclusion, we have tested the Elastic Net version in new data and the rmse was 1164, pretty close to our tests"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}