1.Data Preparation.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# I. Data Preparation (equal bin)\n",
    "\n",
    "The goal of this dataset is to clean the data set we'll use for data visualizations and training the model.\n",
    "\n",
    "We'll proceed to:\n",
    "- Data Exploration and data split\n",
    "- Feature Engineering (categorical encoding using equal bin discretiser and top/bottom encoding)\n",
    "- Save data for feature predictions\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Part I : Data Exploratoin"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "####  1. Importing libraries and data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# to handle datasets\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "# for plotting\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "\n",
    "# to divide train and test set\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "# feature scaling\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "\n",
    "# for tree binarisation\n",
    "from sklearn.tree import DecisionTreeRegressor\n",
    "from sklearn.model_selection import cross_val_score\n",
    "\n",
    "\n",
    "# to build the models\n",
    "from sklearn.linear_model import LinearRegression, Lasso\n",
    "from sklearn.ensemble import RandomForestRegressor\n",
    "from sklearn.svm import SVR\n",
    "import xgboost as xgb\n",
    "\n",
    "# to evaluate the models\n",
    "from sklearn.metrics import mean_squared_error\n",
    "from math import sqrt\n",
    "\n",
    "pd.pandas.set_option('display.max_columns', None)\n",
    "\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')\n",
    "\n",
    "\n",
    "#Scaling\n",
    "from sklearn.preprocessing import RobustScaler\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 2. Import and merge datasets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(216, 26)"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#import first data set\n",
    "autos=pd.read_csv('autos.csv',sep=',')\n",
    "autos.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>symboling</th>\n",
       "      <th>normalized-losses</th>\n",
       "      <th>make-id</th>\n",
       "      <th>fuel-type</th>\n",
       "      <th>aspiration</th>\n",
       "      <th>num-of-doors</th>\n",
       "      <th>body-style</th>\n",
       "      <th>drive-wheels</th>\n",
       "      <th>engine-location</th>\n",
       "      <th>wheel-base</th>\n",
       "      <th>length</th>\n",
       "      <th>width</th>\n",
       "      <th>height</th>\n",
       "      <th>curb-weight</th>\n",
       "      <th>engine-type</th>\n",
       "      <th>num-of-cylinders</th>\n",
       "      <th>engine-size</th>\n",
       "      <th>fuel-system</th>\n",
       "      <th>bore</th>\n",
       "      <th>stroke</th>\n",
       "      <th>compression-ratio</th>\n",
       "      <th>horsepower</th>\n",
       "      <th>peak-rpm</th>\n",
       "      <th>city-mpg</th>\n",
       "      <th>highway-mpg</th>\n",
       "      <th>price</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>3</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1</td>\n",
       "      <td>gas</td>\n",
       "      <td>std</td>\n",
       "      <td>two</td>\n",
       "      <td>convertible</td>\n",
       "      <td>rwd</td>\n",
       "      <td>front</td>\n",
       "      <td>88.6</td>\n",
       "      <td>168.8</td>\n",
       "      <td>64.1</td>\n",
       "      <td>48.8</td>\n",
       "      <td>2548</td>\n",
       "      <td>dohc</td>\n",
       "      <td>four</td>\n",
       "      <td>130</td>\n",
       "      <td>mpfi</td>\n",
       "      <td>3.47</td>\n",
       "      <td>2.68</td>\n",
       "      <td>9.0</td>\n",
       "      <td>111.0</td>\n",
       "      <td>5000.0</td>\n",
       "      <td>21</td>\n",
       "      <td>27</td>\n",
       "      <td>13495.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>3</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1</td>\n",
       "      <td>gas</td>\n",
       "      <td>std</td>\n",
       "      <td>two</td>\n",
       "      <td>convertible</td>\n",
       "      <td>rwd</td>\n",
       "      <td>front</td>\n",
       "      <td>88.6</td>\n",
       "      <td>168.8</td>\n",
       "      <td>64.1</td>\n",
       "      <td>48.8</td>\n",
       "      <td>2548</td>\n",
       "      <td>dohc</td>\n",
       "      <td>four</td>\n",
       "      <td>130</td>\n",
       "      <td>mpfi</td>\n",
       "      <td>3.47</td>\n",
       "      <td>2.68</td>\n",
       "      <td>9.0</td>\n",
       "      <td>111.0</td>\n",
       "      <td>5000.0</td>\n",
       "      <td>21</td>\n",
       "      <td>27</td>\n",
       "      <td>16500.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   symboling  normalized-losses  make-id fuel-type aspiration num-of-doors  \\\n",
       "0          3                NaN        1       gas        std          two   \n",
       "1          3                NaN        1       gas        std          two   \n",
       "\n",
       "    body-style drive-wheels engine-location  wheel-base  length  width  \\\n",
       "0  convertible          rwd           front        88.6   168.8   64.1   \n",
       "1  convertible          rwd           front        88.6   168.8   64.1   \n",
       "\n",
       "   height  curb-weight engine-type num-of-cylinders  engine-size fuel-system  \\\n",
       "0    48.8         2548        dohc             four          130        mpfi   \n",
       "1    48.8         2548        dohc             four          130        mpfi   \n",
       "\n",
       "   bore  stroke  compression-ratio  horsepower  peak-rpm  city-mpg  \\\n",
       "0  3.47    2.68                9.0       111.0    5000.0        21   \n",
       "1  3.47    2.68                9.0       111.0    5000.0        21   \n",
       "\n",
       "   highway-mpg    price  \n",
       "0           27  13495.0  \n",
       "1           27  16500.0  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#View data\n",
    "autos.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(22, 2)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#import second data set\n",
    "makes=pd.read_csv('makes.csv',sep=',')\n",
    "makes.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(216, 28)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Merge the two datasets\n",
    "df= pd.concat([autos,makes],axis=1)\n",
    "df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['symboling', 'normalized-losses', 'make-id', 'fuel-type', 'aspiration',\n",
       "       'num-of-doors', 'body-style', 'drive-wheels', 'engine-location',\n",
       "       'wheel-base', 'length', 'width', 'height', 'curb-weight', 'engine-type',\n",
       "       'num-of-cylinders', 'engine-size', 'fuel-system', 'bore', 'stroke',\n",
       "       'compression-ratio', 'horsepower', 'peak-rpm', 'city-mpg',\n",
       "       'highway-mpg', 'price', 'make-id', 'make'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#column list\n",
    "df.columns"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "I can foresee problems when dealing with problems with a - sign, so I shall replace it by a lower case _."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "old_names = ['symboling', 'normalized-losses', 'make-id', 'fuel-type', 'aspiration',\n",
    "       'num-of-doors', 'body-style', 'drive-wheels', 'engine-location',\n",
    "       'wheel-base', 'length', 'width', 'height', 'curb-weight', 'engine-type',\n",
    "       'num-of-cylinders', 'engine-size', 'fuel-system', 'bore', 'stroke',\n",
    "       'compression-ratio', 'horsepower', 'peak-rpm', 'city-mpg',\n",
    "       'highway-mpg', 'price', 'make-id', 'make'] \n",
    "new_names = ['symboling', 'normalized_losses', 'make_id', 'fuel_type', 'aspiration',\n",
    "       'num_of_doors', 'body_style', 'drive_wheels', 'engine_location',\n",
    "       'wheel_base', 'length', 'width', 'height', 'curb_weight', 'engine_type',\n",
    "       'num_of_cylinders', 'engine_size', 'fuel_system', 'bore', 'stroke',\n",
    "       'compression_ratio', 'horsepower', 'peak_rpm', 'city_mpg',\n",
    "       'highway_mpg', 'price', 'make_id', 'make']\n",
    "df.rename(columns=dict(zip(old_names, new_names)), inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>symboling</th>\n",
       "      <th>normalized_losses</th>\n",
       "      <th>make_id</th>\n",
       "      <th>fuel_type</th>\n",
       "      <th>aspiration</th>\n",
       "      <th>num_of_doors</th>\n",
       "      <th>body_style</th>\n",
       "      <th>drive_wheels</th>\n",
       "      <th>engine_location</th>\n",
       "      <th>wheel_base</th>\n",
       "      <th>length</th>\n",
       "      <th>width</th>\n",
       "      <th>height</th>\n",
       "      <th>curb_weight</th>\n",
       "      <th>engine_type</th>\n",
       "      <th>num_of_cylinders</th>\n",
       "      <th>engine_size</th>\n",
       "      <th>fuel_system</th>\n",
       "      <th>bore</th>\n",
       "      <th>stroke</th>\n",
       "      <th>compression_ratio</th>\n",
       "      <th>horsepower</th>\n",
       "      <th>peak_rpm</th>\n",
       "      <th>city_mpg</th>\n",
       "      <th>highway_mpg</th>\n",
       "      <th>price</th>\n",
       "      <th>make_id</th>\n",
       "      <th>make</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>3</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1</td>\n",
       "      <td>gas</td>\n",
       "      <td>std</td>\n",
       "      <td>two</td>\n",
       "      <td>convertible</td>\n",
       "      <td>rwd</td>\n",
       "      <td>front</td>\n",
       "      <td>88.6</td>\n",
       "      <td>168.8</td>\n",
       "      <td>64.1</td>\n",
       "      <td>48.8</td>\n",
       "      <td>2548</td>\n",
       "      <td>dohc</td>\n",
       "      <td>four</td>\n",
       "      <td>130</td>\n",
       "      <td>mpfi</td>\n",
       "      <td>3.47</td>\n",
       "      <td>2.68</td>\n",
       "      <td>9.0</td>\n",
       "      <td>111.0</td>\n",
       "      <td>5000.0</td>\n",
       "      <td>21</td>\n",
       "      <td>27</td>\n",
       "      <td>13495.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>alfa-romero</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>3</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1</td>\n",
       "      <td>gas</td>\n",
       "      <td>std</td>\n",
       "      <td>two</td>\n",
       "      <td>convertible</td>\n",
       "      <td>rwd</td>\n",
       "      <td>front</td>\n",
       "      <td>88.6</td>\n",
       "      <td>168.8</td>\n",
       "      <td>64.1</td>\n",
       "      <td>48.8</td>\n",
       "      <td>2548</td>\n",
       "      <td>dohc</td>\n",
       "      <td>four</td>\n",
       "      <td>130</td>\n",
       "      <td>mpfi</td>\n",
       "      <td>3.47</td>\n",
       "      <td>2.68</td>\n",
       "      <td>9.0</td>\n",
       "      <td>111.0</td>\n",
       "      <td>5000.0</td>\n",
       "      <td>21</td>\n",
       "      <td>27</td>\n",
       "      <td>16500.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>audi</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1</td>\n",
       "      <td>gas</td>\n",
       "      <td>std</td>\n",
       "      <td>two</td>\n",
       "      <td>hatchback</td>\n",
       "      <td>rwd</td>\n",
       "      <td>front</td>\n",
       "      <td>94.5</td>\n",
       "      <td>171.2</td>\n",
       "      <td>65.5</td>\n",
       "      <td>52.4</td>\n",
       "      <td>2823</td>\n",
       "      <td>ohcv</td>\n",
       "      <td>six</td>\n",
       "      <td>152</td>\n",
       "      <td>mpfi</td>\n",
       "      <td>2.68</td>\n",
       "      <td>3.47</td>\n",
       "      <td>9.0</td>\n",
       "      <td>154.0</td>\n",
       "      <td>5000.0</td>\n",
       "      <td>19</td>\n",
       "      <td>26</td>\n",
       "      <td>16500.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>bmw</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2</td>\n",
       "      <td>164.0</td>\n",
       "      <td>2</td>\n",
       "      <td>gas</td>\n",
       "      <td>std</td>\n",
       "      <td>four</td>\n",
       "      <td>sedan</td>\n",
       "      <td>fwd</td>\n",
       "      <td>front</td>\n",
       "      <td>99.8</td>\n",
       "      <td>176.6</td>\n",
       "      <td>66.2</td>\n",
       "      <td>54.3</td>\n",
       "      <td>2337</td>\n",
       "      <td>ohc</td>\n",
       "      <td>four</td>\n",
       "      <td>109</td>\n",
       "      <td>mpfi</td>\n",
       "      <td>3.19</td>\n",
       "      <td>3.40</td>\n",
       "      <td>10.0</td>\n",
       "      <td>102.0</td>\n",
       "      <td>5500.0</td>\n",
       "      <td>24</td>\n",
       "      <td>30</td>\n",
       "      <td>13950.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>chevrolet</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2</td>\n",
       "      <td>164.0</td>\n",
       "      <td>2</td>\n",
       "      <td>gas</td>\n",
       "      <td>std</td>\n",
       "      <td>four</td>\n",
       "      <td>sedan</td>\n",
       "      <td>4wd</td>\n",
       "      <td>front</td>\n",
       "      <td>99.4</td>\n",
       "      <td>176.6</td>\n",
       "      <td>66.4</td>\n",
       "      <td>54.3</td>\n",
       "      <td>2824</td>\n",
       "      <td>ohc</td>\n",
       "      <td>five</td>\n",
       "      <td>136</td>\n",
       "      <td>mpfi</td>\n",
       "      <td>3.19</td>\n",
       "      <td>3.40</td>\n",
       "      <td>8.0</td>\n",
       "      <td>115.0</td>\n",
       "      <td>5500.0</td>\n",
       "      <td>18</td>\n",
       "      <td>22</td>\n",
       "      <td>17450.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>dodge</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   symboling  normalized_losses  make_id fuel_type aspiration num_of_doors  \\\n",
       "0          3                NaN        1       gas        std          two   \n",
       "1          3                NaN        1       gas        std          two   \n",
       "2          1                NaN        1       gas        std          two   \n",
       "3          2              164.0        2       gas        std         four   \n",
       "4          2              164.0        2       gas        std         four   \n",
       "\n",
       "    body_style drive_wheels engine_location  wheel_base  length  width  \\\n",
       "0  convertible          rwd           front        88.6   168.8   64.1   \n",
       "1  convertible          rwd           front        88.6   168.8   64.1   \n",
       "2    hatchback          rwd           front        94.5   171.2   65.5   \n",
       "3        sedan          fwd           front        99.8   176.6   66.2   \n",
       "4        sedan          4wd           front        99.4   176.6   66.4   \n",
       "\n",
       "   height  curb_weight engine_type num_of_cylinders  engine_size fuel_system  \\\n",
       "0    48.8         2548        dohc             four          130        mpfi   \n",
       "1    48.8         2548        dohc             four          130        mpfi   \n",
       "2    52.4         2823        ohcv              six          152        mpfi   \n",
       "3    54.3         2337         ohc             four          109        mpfi   \n",
       "4    54.3         2824         ohc             five          136        mpfi   \n",
       "\n",
       "   bore  stroke  compression_ratio  horsepower  peak_rpm  city_mpg  \\\n",
       "0  3.47    2.68                9.0       111.0    5000.0        21   \n",
       "1  3.47    2.68                9.0       111.0    5000.0        21   \n",
       "2  2.68    3.47                9.0       154.0    5000.0        19   \n",
       "3  3.19    3.40               10.0       102.0    5500.0        24   \n",
       "4  3.19    3.40                8.0       115.0    5500.0        18   \n",
       "\n",
       "   highway_mpg    price  make_id           make  \n",
       "0           27  13495.0      1.0  alfa-romero    \n",
       "1           27  16500.0      2.0         audi    \n",
       "2           26  16500.0      3.0          bmw    \n",
       "3           30  13950.0      4.0    chevrolet    \n",
       "4           22  17450.0      5.0        dodge    "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['symboling', 'normalized_losses', 'make_id', 'fuel_type', 'aspiration',\n",
       "       'num_of_doors', 'body_style', 'drive_wheels', 'engine_location',\n",
       "       'wheel_base', 'length', 'width', 'height', 'curb_weight', 'engine_type',\n",
       "       'num_of_cylinders', 'engine_size', 'fuel_system', 'bore', 'stroke',\n",
       "       'compression_ratio', 'horsepower', 'peak_rpm', 'city_mpg',\n",
       "       'highway_mpg', 'price', 'make_id', 'make'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.columns"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 3. Drop duplicate rows and columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "#drop duplicate rows (using subset, I drop raws where values of columns mentioned match)\n",
    "df.drop_duplicates(subset=['symboling', 'normalized_losses', 'make_id', 'fuel_type', 'aspiration',\n",
    "       'num_of_doors', 'body_style', 'drive_wheels', 'engine_location',\n",
    "       'wheel_base', 'length', 'width', 'height', 'curb_weight', 'engine_type',\n",
    "       'num_of_cylinders', 'engine_size', 'fuel_system', 'bore', 'stroke',\n",
    "       'compression_ratio', 'horsepower', 'peak_rpm', 'city_mpg',\n",
    "       'highway_mpg', 'price', 'make_id', 'make'], inplace = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(205, 28)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#see shape after duplicate row removal\n",
    "df.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "I had 11 duplicates that have been removed"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(205, 27)"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# remove duplicate columns\n",
    "_, i = np.unique(df.columns, return_index=True)\n",
    "df=df.iloc[:, i]\n",
    "df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "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>aspiration</th>\n",
       "      <th>body_style</th>\n",
       "      <th>bore</th>\n",
       "      <th>city_mpg</th>\n",
       "      <th>compression_ratio</th>\n",
       "      <th>curb_weight</th>\n",
       "      <th>drive_wheels</th>\n",
       "      <th>engine_location</th>\n",
       "      <th>engine_size</th>\n",
       "      <th>engine_type</th>\n",
       "      <th>fuel_system</th>\n",
       "      <th>fuel_type</th>\n",
       "      <th>height</th>\n",
       "      <th>highway_mpg</th>\n",
       "      <th>horsepower</th>\n",
       "      <th>length</th>\n",
       "      <th>make</th>\n",
       "      <th>make_id</th>\n",
       "      <th>normalized_losses</th>\n",
       "      <th>num_of_cylinders</th>\n",
       "      <th>num_of_doors</th>\n",
       "      <th>peak_rpm</th>\n",
       "      <th>price</th>\n",
       "      <th>stroke</th>\n",
       "      <th>symboling</th>\n",
       "      <th>wheel_base</th>\n",
       "      <th>width</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>std</td>\n",
       "      <td>convertible</td>\n",
       "      <td>3.47</td>\n",
       "      <td>21</td>\n",
       "      <td>9.0</td>\n",
       "      <td>2548</td>\n",
       "      <td>rwd</td>\n",
       "      <td>front</td>\n",
       "      <td>130</td>\n",
       "      <td>dohc</td>\n",
       "      <td>mpfi</td>\n",
       "      <td>gas</td>\n",
       "      <td>48.8</td>\n",
       "      <td>27</td>\n",
       "      <td>111.0</td>\n",
       "      <td>168.8</td>\n",
       "      <td>alfa-romero</td>\n",
       "      <td>1</td>\n",
       "      <td>NaN</td>\n",
       "      <td>four</td>\n",
       "      <td>two</td>\n",
       "      <td>5000.0</td>\n",
       "      <td>13495.0</td>\n",
       "      <td>2.68</td>\n",
       "      <td>3</td>\n",
       "      <td>88.6</td>\n",
       "      <td>64.1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>std</td>\n",
       "      <td>convertible</td>\n",
       "      <td>3.47</td>\n",
       "      <td>21</td>\n",
       "      <td>9.0</td>\n",
       "      <td>2548</td>\n",
       "      <td>rwd</td>\n",
       "      <td>front</td>\n",
       "      <td>130</td>\n",
       "      <td>dohc</td>\n",
       "      <td>mpfi</td>\n",
       "      <td>gas</td>\n",
       "      <td>48.8</td>\n",
       "      <td>27</td>\n",
       "      <td>111.0</td>\n",
       "      <td>168.8</td>\n",
       "      <td>audi</td>\n",
       "      <td>1</td>\n",
       "      <td>NaN</td>\n",
       "      <td>four</td>\n",
       "      <td>two</td>\n",
       "      <td>5000.0</td>\n",
       "      <td>16500.0</td>\n",
       "      <td>2.68</td>\n",
       "      <td>3</td>\n",
       "      <td>88.6</td>\n",
       "      <td>64.1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>std</td>\n",
       "      <td>hatchback</td>\n",
       "      <td>2.68</td>\n",
       "      <td>19</td>\n",
       "      <td>9.0</td>\n",
       "      <td>2823</td>\n",
       "      <td>rwd</td>\n",
       "      <td>front</td>\n",
       "      <td>152</td>\n",
       "      <td>ohcv</td>\n",
       "      <td>mpfi</td>\n",
       "      <td>gas</td>\n",
       "      <td>52.4</td>\n",
       "      <td>26</td>\n",
       "      <td>154.0</td>\n",
       "      <td>171.2</td>\n",
       "      <td>bmw</td>\n",
       "      <td>1</td>\n",
       "      <td>NaN</td>\n",
       "      <td>six</td>\n",
       "      <td>two</td>\n",
       "      <td>5000.0</td>\n",
       "      <td>16500.0</td>\n",
       "      <td>3.47</td>\n",
       "      <td>1</td>\n",
       "      <td>94.5</td>\n",
       "      <td>65.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>std</td>\n",
       "      <td>sedan</td>\n",
       "      <td>3.19</td>\n",
       "      <td>24</td>\n",
       "      <td>10.0</td>\n",
       "      <td>2337</td>\n",
       "      <td>fwd</td>\n",
       "      <td>front</td>\n",
       "      <td>109</td>\n",
       "      <td>ohc</td>\n",
       "      <td>mpfi</td>\n",
       "      <td>gas</td>\n",
       "      <td>54.3</td>\n",
       "      <td>30</td>\n",
       "      <td>102.0</td>\n",
       "      <td>176.6</td>\n",
       "      <td>chevrolet</td>\n",
       "      <td>2</td>\n",
       "      <td>164.0</td>\n",
       "      <td>four</td>\n",
       "      <td>four</td>\n",
       "      <td>5500.0</td>\n",
       "      <td>13950.0</td>\n",
       "      <td>3.40</td>\n",
       "      <td>2</td>\n",
       "      <td>99.8</td>\n",
       "      <td>66.2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>std</td>\n",
       "      <td>sedan</td>\n",
       "      <td>3.19</td>\n",
       "      <td>18</td>\n",
       "      <td>8.0</td>\n",
       "      <td>2824</td>\n",
       "      <td>4wd</td>\n",
       "      <td>front</td>\n",
       "      <td>136</td>\n",
       "      <td>ohc</td>\n",
       "      <td>mpfi</td>\n",
       "      <td>gas</td>\n",
       "      <td>54.3</td>\n",
       "      <td>22</td>\n",
       "      <td>115.0</td>\n",
       "      <td>176.6</td>\n",
       "      <td>dodge</td>\n",
       "      <td>2</td>\n",
       "      <td>164.0</td>\n",
       "      <td>five</td>\n",
       "      <td>four</td>\n",
       "      <td>5500.0</td>\n",
       "      <td>17450.0</td>\n",
       "      <td>3.40</td>\n",
       "      <td>2</td>\n",
       "      <td>99.4</td>\n",
       "      <td>66.4</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  aspiration   body_style  bore  city_mpg  compression_ratio  curb_weight  \\\n",
       "0        std  convertible  3.47        21                9.0         2548   \n",
       "1        std  convertible  3.47        21                9.0         2548   \n",
       "2        std    hatchback  2.68        19                9.0         2823   \n",
       "3        std        sedan  3.19        24               10.0         2337   \n",
       "4        std        sedan  3.19        18                8.0         2824   \n",
       "\n",
       "  drive_wheels engine_location  engine_size engine_type fuel_system fuel_type  \\\n",
       "0          rwd           front          130        dohc        mpfi       gas   \n",
       "1          rwd           front          130        dohc        mpfi       gas   \n",
       "2          rwd           front          152        ohcv        mpfi       gas   \n",
       "3          fwd           front          109         ohc        mpfi       gas   \n",
       "4          4wd           front          136         ohc        mpfi       gas   \n",
       "\n",
       "   height  highway_mpg  horsepower  length           make  make_id  \\\n",
       "0    48.8           27       111.0   168.8  alfa-romero          1   \n",
       "1    48.8           27       111.0   168.8         audi          1   \n",
       "2    52.4           26       154.0   171.2          bmw          1   \n",
       "3    54.3           30       102.0   176.6    chevrolet          2   \n",
       "4    54.3           22       115.0   176.6        dodge          2   \n",
       "\n",
       "   normalized_losses num_of_cylinders num_of_doors  peak_rpm    price  stroke  \\\n",
       "0                NaN             four          two    5000.0  13495.0    2.68   \n",
       "1                NaN             four          two    5000.0  16500.0    2.68   \n",
       "2                NaN              six          two    5000.0  16500.0    3.47   \n",
       "3              164.0             four         four    5500.0  13950.0    3.40   \n",
       "4              164.0             five         four    5500.0  17450.0    3.40   \n",
       "\n",
       "   symboling  wheel_base  width  \n",
       "0          3        88.6   64.1  \n",
       "1          3        88.6   64.1  \n",
       "2          1        94.5   65.5  \n",
       "3          2        99.8   66.2  \n",
       "4          2        99.4   66.4  "
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 4. Types of Variables"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "aspiration            object\n",
       "body_style            object\n",
       "bore                 float64\n",
       "city_mpg               int64\n",
       "compression_ratio    float64\n",
       "curb_weight            int64\n",
       "drive_wheels          object\n",
       "engine_location       object\n",
       "engine_size            int64\n",
       "engine_type           object\n",
       "fuel_system           object\n",
       "fuel_type             object\n",
       "height               float64\n",
       "highway_mpg            int64\n",
       "horsepower           float64\n",
       "length               float64\n",
       "make                  object\n",
       "make_id                int64\n",
       "normalized_losses    float64\n",
       "num_of_cylinders      object\n",
       "num_of_doors          object\n",
       "peak_rpm             float64\n",
       "price                float64\n",
       "stroke               float64\n",
       "symboling              int64\n",
       "wheel_base           float64\n",
       "width                float64\n",
       "dtype: object"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# let's inspect the type of variables in pandas\n",
    "df.dtypes"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "There are a mixture of numerical and categorical variables. Normally object type determines categorical. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 4.1 Find categorical variables"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "There are 10 categorical variables\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "['aspiration',\n",
       " 'body_style',\n",
       " 'drive_wheels',\n",
       " 'engine_location',\n",
       " 'engine_type',\n",
       " 'fuel_system',\n",
       " 'fuel_type',\n",
       " 'make',\n",
       " 'num_of_cylinders',\n",
       " 'num_of_doors']"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# find categorical variables\n",
    "categorical = [var for var in df.columns if df[var].dtype=='O']\n",
    "print('There are {} categorical variables'.format(len(categorical)))\n",
    "categorical"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 4.2 Find temporal variables"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "There are 0 temporal variables\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[]"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# make a list of the numerical variables first\n",
    "numerical = [var for var in df.columns if df[var].dtype!='O']\n",
    "\n",
    "# list of variables that contain year information\n",
    "year_vars = [var for var in numerical if 'Yr' in var or 'Year' in var or 'Day'in var or 'Month'in var or 'Time'in var]\n",
    "print('There are {} temporal variables'.format(len(year_vars)))\n",
    "\n",
    "\n",
    "year_vars"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "No temporal variables"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 4.3 Find Discrete Varibles"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "symboling  values:  [ 3  1  2  0 -1 -2]\n",
      "\n",
      "There are 1 discrete variables\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "['symboling']"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# let's visualise the values of the discrete variables\n",
    "discrete = []\n",
    "\n",
    "for var in numerical:\n",
    "    if len(df[var].unique())<20 and var not in year_vars:\n",
    "        print(var, ' values: ', df[var].unique())\n",
    "        discrete.append(var)\n",
    "print()\n",
    "print('There are {} discrete variables'.format(len(discrete)))\n",
    "discrete"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "There are 16 numerical and continuous variables\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "['bore',\n",
       " 'city_mpg',\n",
       " 'compression_ratio',\n",
       " 'curb_weight',\n",
       " 'engine_size',\n",
       " 'height',\n",
       " 'highway_mpg',\n",
       " 'horsepower',\n",
       " 'length',\n",
       " 'make_id',\n",
       " 'normalized_losses',\n",
       " 'peak_rpm',\n",
       " 'price',\n",
       " 'stroke',\n",
       " 'wheel_base',\n",
       " 'width']"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "### Find Continuous\n",
    "# let's remember to skip the Id variable and the target variable SalePrice, which are both also numerical\n",
    "\n",
    "numerical = [var for var in numerical if var not in discrete and var not in ['Id', 'SalePrice'] and var not in year_vars]\n",
    "print('There are {} numerical and continuous variables'.format(len(numerical)))\n",
    "numerical"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "SO there are: \n",
    "\n",
    "- 1 discrete variables \n",
    "- 10 categorical \n",
    "- 0 temporal\n",
    "- 16 continuous\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 5.Types of problems within variables\n",
    "\n",
    "#### 5.1 Missing values"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "I have very few missing values in general, except for the 'normalized-losses'and 'make'columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total variables that contain missing information:  8\n"
     ]
    }
   ],
   "source": [
    "# let's no determine how many variables we have with missing information\n",
    "\n",
    "vars_with_na = [var for var in df.columns if df[var].isnull().sum()>0]\n",
    "print('Total variables that contain missing information: ', len(vars_with_na))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "There are a few variables with a lot of missing information, let's find out what kind of variables they are"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "make ['alfa-romero  ' 'audi  ' 'bmw  ' 'chevrolet  ' 'dodge  ' 'honda  '\n",
      " 'isuzu  ' 'jaguar  ' 'mazda  ' 'mercedes-benz  ' 'mercury  '\n",
      " 'mitsubishi  ' 'nissan  ' 'peugot  ' 'plymouth  ' 'porsche  ' 'renault  '\n",
      " 'saab  ' 'subaru  ' 'toyota  ' 'volkswagen  ' 'volvo ' nan]\n"
     ]
    }
   ],
   "source": [
    "# let's inspect the type of those variables with a lot of missing information\n",
    "for var in df.columns:\n",
    "    if df[var].isnull().mean()>0.80:\n",
    "        print(var, df[var].unique())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "bore 0.01951219512195122\n",
      "horsepower 0.00975609756097561\n",
      "make 0.8926829268292683\n",
      "normalized_losses 0.2\n",
      "num_of_doors 0.00975609756097561\n",
      "peak_rpm 0.00975609756097561\n",
      "price 0.01951219512195122\n",
      "stroke 0.01951219512195122\n"
     ]
    }
   ],
   "source": [
    "# let's visualise the percentage of missing values for each variable\n",
    "for var in df.columns:\n",
    "    if df[var].isnull().sum()>0:\n",
    "        print(var, df[var].isnull().mean())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "One variable with more than 80% missing values, it's categorical"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 5.2 Outliers\n",
    "\n",
    "#### 5.2.1 Start with numerical continuous"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['bore',\n",
       " 'city_mpg',\n",
       " 'compression_ratio',\n",
       " 'curb_weight',\n",
       " 'engine_size',\n",
       " 'height',\n",
       " 'highway_mpg',\n",
       " 'horsepower',\n",
       " 'length',\n",
       " 'make_id',\n",
       " 'normalized_losses',\n",
       " 'peak_rpm',\n",
       " 'price',\n",
       " 'stroke',\n",
       " 'wheel_base',\n",
       " 'width']"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# let's look at the numerical variables\n",
    "numerical"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The only real continuous numerical variable is DepDelay, the others are discrete or categorical (ID)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 5.2.2 Continue with discrete variables"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We know discrete varibles are: \n",
    "\n",
    "     - Symboling\n",
    "    \n",
    "    Actually, looking at the dataset, we can see antoher discrete variable (with more than 20 values)\n",
    "    \n",
    "    - make_id"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "discrete = ['symboling','make_id']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "# outlies in discrete variables\n",
    "for var in discrete:\n",
    "    (df.groupby(var)[var].count() / np.float(len(df))).plot.bar()\n",
    "    plt.ylabel('Percentage of observations per label')\n",
    "    plt.title(var)\n",
    "    plt.show()\n",
    "    #print(data[var].value_counts() / np.float(len(data)))\n",
    "    print()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 5.2.3 Number of labels: Cardinality"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'Number of different categories')"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "no_labels_ls = []\n",
    "for var in categorical:\n",
    "    no_labels_ls.append(len(df[var].unique()))\n",
    "    \n",
    " \n",
    "tmp = pd.Series(no_labels_ls)\n",
    "tmp.index = pd.Series(categorical)\n",
    "tmp.plot.bar(figsize=(12,8))\n",
    "plt.title('Number of categories in categorical variables')\n",
    "plt.xlabel('Categorical variables')\n",
    "plt.ylabel('Number of different categories')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Most of them have a few variables, except for make"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**5.2.4 Outliers in Continuous**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To be seen later"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "####  6. Split into train and test sets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((174, 27), (31, 27))"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Let's separate into train and test set\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(df, df.price, test_size=0.15,\n",
    "                                                    random_state=0)\n",
    "X_train.shape, X_test.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Part II: Feature engineering\n",
    "\n",
    "### 7. Engineering missing values\n",
    "\n",
    "**7.1 Continuous variables**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "bore 0.017241379310344827\n",
      "normalized_losses 0.1839080459770115\n",
      "price 0.011494252873563218\n",
      "stroke 0.017241379310344827\n"
     ]
    }
   ],
   "source": [
    "# print variables with missing data\n",
    "# keep in mind that now that we created those new temporal variables, we\n",
    "# are going to treat them as numerical and continuous as well:\n",
    "\n",
    "\n",
    "# examine percentage of missing values\n",
    "for col in numerical+year_vars:\n",
    "    if X_train[col].isnull().mean()>0:\n",
    "        print(col, X_train[col].isnull().mean())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Bore, stroke, price** and stroke missing values are less than 5%: median imputation\n",
    "\n",
    "**normalized losses** missing values are more than 5%: additional variable to capture NA + median imputation    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "# add variable indicating missingness + median imputation\n",
    "for df in [X_train, X_test]:\n",
    "    for var in ['normalized_losses']:\n",
    "        df[var+'_NA'] = np.where(df[var].isnull(), 1, 0)\n",
    "        df[var].fillna(X_train[var].median(), inplace=True) \n",
    "\n",
    "for df in [X_train, X_test]:\n",
    "    df['normalized_losses'].fillna(X_train['normalized_losses'].median(), inplace=True)\n",
    "    df['price'].fillna(X_train.price.median(), inplace=True)\n",
    "    df['stroke'].fillna(X_train.stroke.median(), inplace=True)\n",
    "\n",
    "\n",
    "    \n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**7.2 Discrete variables**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['symboling', 'make_id']\n",
      "['symboling', 'make_id']\n"
     ]
    }
   ],
   "source": [
    "# print variables with missing data\n",
    "for col in discrete:\n",
    "    if X_train[col].isnull().mean()>0:\n",
    "        print(col, X_train[col].isnull().mean())\n",
    "    print(discrete)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "There are no discrete variables variables with missing data\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 7.3 Engineering Missing data in Categorical Variables"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "make 0.896551724137931\n",
      "num_of_doors 0.011494252873563218\n"
     ]
    }
   ],
   "source": [
    "# print variables with missing data\n",
    "for col in categorical:\n",
    "    if X_train[col].isnull().mean()>0:\n",
    "        print(col, X_train[col].isnull().mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>aspiration</th>\n",
       "      <th>body_style</th>\n",
       "      <th>drive_wheels</th>\n",
       "      <th>engine_location</th>\n",
       "      <th>engine_type</th>\n",
       "      <th>fuel_system</th>\n",
       "      <th>fuel_type</th>\n",
       "      <th>make</th>\n",
       "      <th>num_of_cylinders</th>\n",
       "      <th>num_of_doors</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>193</th>\n",
       "      <td>std</td>\n",
       "      <td>wagon</td>\n",
       "      <td>fwd</td>\n",
       "      <td>front</td>\n",
       "      <td>ohc</td>\n",
       "      <td>mpfi</td>\n",
       "      <td>gas</td>\n",
       "      <td>NaN</td>\n",
       "      <td>four</td>\n",
       "      <td>four</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>125</th>\n",
       "      <td>std</td>\n",
       "      <td>hatchback</td>\n",
       "      <td>rwd</td>\n",
       "      <td>front</td>\n",
       "      <td>ohc</td>\n",
       "      <td>mpfi</td>\n",
       "      <td>gas</td>\n",
       "      <td>NaN</td>\n",
       "      <td>four</td>\n",
       "      <td>two</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>173</th>\n",
       "      <td>std</td>\n",
       "      <td>sedan</td>\n",
       "      <td>fwd</td>\n",
       "      <td>front</td>\n",
       "      <td>ohc</td>\n",
       "      <td>mpfi</td>\n",
       "      <td>gas</td>\n",
       "      <td>NaN</td>\n",
       "      <td>four</td>\n",
       "      <td>four</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>92</th>\n",
       "      <td>std</td>\n",
       "      <td>sedan</td>\n",
       "      <td>fwd</td>\n",
       "      <td>front</td>\n",
       "      <td>ohc</td>\n",
       "      <td>2bbl</td>\n",
       "      <td>gas</td>\n",
       "      <td>NaN</td>\n",
       "      <td>four</td>\n",
       "      <td>four</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>std</td>\n",
       "      <td>sedan</td>\n",
       "      <td>rwd</td>\n",
       "      <td>front</td>\n",
       "      <td>ohc</td>\n",
       "      <td>mpfi</td>\n",
       "      <td>gas</td>\n",
       "      <td>renault</td>\n",
       "      <td>six</td>\n",
       "      <td>two</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    aspiration body_style drive_wheels engine_location engine_type  \\\n",
       "193        std      wagon          fwd           front         ohc   \n",
       "125        std  hatchback          rwd           front         ohc   \n",
       "173        std      sedan          fwd           front         ohc   \n",
       "92         std      sedan          fwd           front         ohc   \n",
       "16         std      sedan          rwd           front         ohc   \n",
       "\n",
       "    fuel_system fuel_type       make num_of_cylinders num_of_doors  \n",
       "193        mpfi       gas        NaN             four         four  \n",
       "125        mpfi       gas        NaN             four          two  \n",
       "173        mpfi       gas        NaN             four         four  \n",
       "92         2bbl       gas        NaN             four         four  \n",
       "16         mpfi       gas  renault                six          two  "
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train[categorical].head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For the 'make'variable, I have over 89% of missing values. I shall delete this variable.\n",
    "\n",
    "As for 'num-of-doors', missing values are less than 5%, so I shall replace them with the most frequent category"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "#remove 'make'\n",
    "X_train = X_train.drop('make', axis = 1)\n",
    "X_test = X_test.drop('make', axis = 1)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'four'"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# find most frequent category in 'num-of-doors'\n",
    "X_train.groupby(['num_of_doors'])['num_of_doors'].count().sort_values(ascending=False).index[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "# let's create a variable to replace NA with the most frequent label\n",
    "# both in train and test set\n",
    "\n",
    "def impute_na(df_train, df_test, variable):\n",
    "    most_frequent_category = df_train.groupby([variable])[variable].count().sort_values(ascending=False).index[0]\n",
    "    df_train[variable].fillna(most_frequent_category, inplace=True)\n",
    "    df_test[variable].fillna(most_frequent_category, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "# and let's replace the NA\n",
    "for variable in ['num_of_doors',]:\n",
    "    impute_na(X_train, X_test, variable)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Finally let's check I have no more nulls"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "bore 3\n"
     ]
    }
   ],
   "source": [
    "# check absence of null values\n",
    "for var in X_train.columns:\n",
    "    if X_train[var].isnull().sum()>0:\n",
    "        print(var, X_train[var].isnull().sum())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "bore 3\n",
      "horsepower 0\n",
      "peak_rpm 0\n"
     ]
    }
   ],
   "source": [
    "# check absence of null values\n",
    "for var in X_test.columns:\n",
    "    if X_test[var].isnull().sum()>0:\n",
    "        print(var, X_train[var].isnull().sum())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "These are variables with null on the test set and not in the training set. I shall fill NA with median value"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "68.0"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "## find most frequent category in 'horsepower'\n",
    "X_train.groupby(['horsepower'])['horsepower'].count().sort_values(ascending=False).index[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "5500.0"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "## find most frequent category in 'peak-rpm'\n",
    "X_train.groupby(['peak_rpm'])['peak_rpm'].count().sort_values(ascending=False).index[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "3.62"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "## find most frequent category in 'peak-rpm'\n",
    "X_train.groupby(['bore'])['bore'].count().sort_values(ascending=False).index[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_test['horsepower'].fillna(68, inplace = True)\n",
    "X_test['peak_rpm'].fillna(5500, inplace = True)\n",
    "X_test['bore'].fillna(3.62, inplace = True)\n",
    "X_train['bore'].fillna(3.62, inplace = True)\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "# check absence of null values\n",
    "for var in X_train.columns:\n",
    "    if X_train[var].isnull().sum()>0:\n",
    "        print(var, X_train[var].isnull().sum())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "# check absence of null values\n",
    "for var in X_test.columns:\n",
    "    if X_test[var].isnull().sum()>0:\n",
    "        print(var, X_test[var].isnull().sum())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "aspiration              0\n",
       "body_style              0\n",
       "bore                    0\n",
       "city_mpg                0\n",
       "compression_ratio       0\n",
       "curb_weight             0\n",
       "drive_wheels            0\n",
       "engine_location         0\n",
       "engine_size             0\n",
       "engine_type             0\n",
       "fuel_system             0\n",
       "fuel_type               0\n",
       "height                  0\n",
       "highway_mpg             0\n",
       "horsepower              0\n",
       "length                  0\n",
       "make_id                 0\n",
       "normalized_losses       0\n",
       "num_of_cylinders        0\n",
       "num_of_doors            0\n",
       "peak_rpm                0\n",
       "price                   0\n",
       "stroke                  0\n",
       "symboling               0\n",
       "wheel_base              0\n",
       "width                   0\n",
       "normalized_losses_NA    0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "aspiration              0\n",
       "body_style              0\n",
       "bore                    0\n",
       "city_mpg                0\n",
       "compression_ratio       0\n",
       "curb_weight             0\n",
       "drive_wheels            0\n",
       "engine_location         0\n",
       "engine_size             0\n",
       "engine_type             0\n",
       "fuel_system             0\n",
       "fuel_type               0\n",
       "height                  0\n",
       "highway_mpg             0\n",
       "horsepower              0\n",
       "length                  0\n",
       "make_id                 0\n",
       "normalized_losses       0\n",
       "num_of_cylinders        0\n",
       "num_of_doors            0\n",
       "peak_rpm                0\n",
       "price                   0\n",
       "stroke                  0\n",
       "symboling               0\n",
       "wheel_base              0\n",
       "width                   0\n",
       "normalized_losses_NA    0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test.isnull().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Well done! I have no more NA in the variables."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 8. Outlier engineering\n",
    "\n",
    "#### 8.1 Outlier identification and strategy setting\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['bore',\n",
       " 'city_mpg',\n",
       " 'compression_ratio',\n",
       " 'curb_weight',\n",
       " 'engine_size',\n",
       " 'height',\n",
       " 'highway_mpg',\n",
       " 'horsepower',\n",
       " 'length',\n",
       " 'make_id',\n",
       " 'normalized_losses',\n",
       " 'peak_rpm',\n",
       " 'price',\n",
       " 'stroke',\n",
       " 'wheel_base',\n",
       " 'width']"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "numerical"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Note I remove the price as it is the target\n",
    "numerical = ['bore',\n",
    " 'city_mpg',\n",
    " 'compression_ratio',\n",
    " 'curb_weight',\n",
    " 'engine_size',\n",
    " 'height',\n",
    " 'highway_mpg',\n",
    " 'horsepower',\n",
    " 'length',\n",
    " 'make_id',\n",
    " 'normalized_losses',\n",
    " 'peak_rpm',\n",
    " 'stroke',\n",
    " 'wheel_base',\n",
    " 'width']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAbEAAAFpCAYAAAASvRbzAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvnQurowAAEstJREFUeJzt3X+MZXd53/H3B6/Bjm0Cxu7UsVMmFUhxohY7GVlEjqLRNrQRS4gDJEpqfqVUW6qkcQQEFvIDQrBkqsaQqFGTDZbZOE4INUZOvK5VY3yJ3LSudu0FY28qAtm0KQ6ExAYvchYbnv4xd5vRZnd27nrOmXnuvl/Slefee858n7Xu6K1z75kzqSokSeroGZs9gCRJp8qISZLaMmKSpLaMmCSpLSMmSWrLiEmS2jJikqS2jJgkqS0jJklqa9tmD7BeF1xwQS0uLm72GNKovvrVr3LOOeds9hjSqPbv3/+lqrpwPdu2idji4iL79u3b7DGkUU0mE5aXlzd7DGlUSf58vdv6dqIkqS0jJklqy4hJktoyYpKktoyYJKktIyZJasuISZLaMmKSpLaMmCSpLSMmSWrLiEmS2jJikqS2jJgkqa02V7GXOksy+ppVNfqa0tg8EpNGUFWndHv+224/5X2l04ERkyS1ZcQkSW0ZMUlSW0ZMktSWEZMktWXEJEltGTFJUltGTJLUlhGTJLVlxCRJbRkxSVJbRkyS1JYRkyS1ZcQkSW2NErEkZyR5IMnt0/sfTPJnSQ5Mb5eNMYckab6M9UcxrwEOAs9e9djPVtUtI60vSZpDgx+JJbkE2AF8YOi1JEmnlzGOxN4PvBU475jHr03yi8DdwK6qOnLsjkl2AjsBFhYWmEwmA48qbT2+7qUTGzRiSV4GfLGq9idZXvXU24G/BJ4J7AbeBrz72P2ravf0eZaWlmp5efnYTaT5dudefN1LJzb024lXAi9Pcgj4ELA9ye9U1SO14ghwI3DFwHNIkubQoBGrqrdX1SVVtQj8GPDxqnp1kosAkgS4Cvj0kHNIkubTWGcnHuvmJBcCAQ4Ab9ykOSRJjY0WsaqaAJPp19vHWleSNL+8YockqS0jJklqy4hJktoyYpKktoyYJKktIyZJasuISZLaMmKSpLaMmCSpLSMmSWrLiEmS2jJikqS2jJgkqS0jJklqy4hJktoyYpKktoyYJKktIyZJasuISZLaMmKSpLaMmCSpLSMmSWrLiEmS2jJikqS2jJgkqS0jJklqy4hJktoyYpKktoyYJKktIyZJasuISZLaMmKSpLaMmCSpLSMmSWrLiEmS2jJikqS2jJgkqS0jJklqy4hJktoyYpKktoyYJKmtUSKW5IwkDyS5fXr/25Lcl+QzSX4/yTPHmEOSNF/GOhK7Bji46v57gfdV1QuBR4E3jDSHJGmODB6xJJcAO4APTO8H2A7cMt1kD3DV0HNIkubPthHWeD/wVuC86f3nAY9V1VPT+38BXHy8HZPsBHYCLCwsMJlMhp1U2oJ83UsnNmjEkrwM+GJV7U+yfPTh42xax9u/qnYDuwGWlpZqeXn5eJtJ8+vOvfi6l05s6COxK4GXJ3kpcBbwbFaOzJ6TZNv0aOwS4PMDzyFJmkODfiZWVW+vqkuqahH4MeDjVXU1cA/wqulmrwNuG3IOSdJ82qzfE3sb8KYkf8rKZ2Q3bNIckqTGxjixA4CqmgCT6defA64Ya21J0nzyih2SpLaMmCSpLSMmSWrLiEmS2jJikqS2jJgkqS0jJklqy4hJktoyYpKktoyYJKktIyZJasuISZLaMmKSpLaMmCSpLSMmSWrLiEmS2jJikqS2jJgkqS0jJklqy4hJktoyYpKktoyYJKktIyZJasuISZLaMmKSpLaMmCSpLSMmSWrLiEmS2jJikqS2jJgkqS0jJklqy4hJktoyYpKktoyYJKktIyZJasuISZLaMmKSpLaMmCSpLSMmSWrLiEmS2ho0YknOSvI/k3wyyUNJfmn6+AeT/FmSA9PbZUPOIUmaT9sG/v5HgO1VdTjJmcC9Sf7L9LmfrapbBl5fkjTHBo1YVRVweHr3zOmthlxTknT6GPpIjCRnAPuBFwC/XlX3Jfm3wLVJfhG4G9hVVUeOs+9OYCfAwsICk8lk6HGlLcfXvXRiWTlYGmGh5DnAR4F/B/w18JfAM4HdwGer6t1r7b+0tFT79u0bfE5pK1nctZdD1+3Y7DGkUSXZX1VL69l2tLMTq+oxYAL8QFU9UiuOADcCV4w1hyRpfgx9duKF0yMwkpwNfD/wJ0kumj4W4Crg00POIUmaT0N/JnYRsGf6udgzgA9X1e1JPp7kQiDAAeCNA88hSZpDQ5+d+Cng8uM8vn3IdSVJpwev2CFJasuISZLaMmKSpLaMmCSprXWf2JHkTcd5+MvA/qo6sHEjSZK0PrMciS2xcir8xdPbTmAZ+K0kb9340SRJWtssp9g/D/iuqjoMkOSdwC3A97FybcR/v/HjSZJ0YrMcif0j4Gur7j8JPL+qnmDlT65IkjSqWY7Efhf4H0luY+VKGy8Dfi/JOcDDQwwnSdJa1h2xqvrlJHcA3zt96I1VdfSy8ldv+GSSJJ3ErKfYH/27Ld+Y3iRJ2jTrjtj0D1juAZ4LXADcmOTnhxpMkqSTmeUzsR8HLq+qvwVIch1wP/CeIQaTJOlkZnk78RBw1qr7zwI+u6HTSJI0g1mOxI4ADyW5i5XPxl4C3Jvk1wCq6qcHmE+SpBOaJWIfnd6OmmzsKJIkzWaWU+z3DDmIJEmzmuXsxJcleSDJ3yT5SpLHk3xlyOEkSVrLLG8nvh94BfBgVdXJNpYkaWiznJ34f4BPGzBJ0lYxy5HYW4E7knyCVRf8rarrN3wqSZLWYZaIXQscZuV3xZ45zDiSJK3fLBE7v6r++WCTSJI0o1k+E/tYEiMmSdoyZonYTwJ3JnnCU+wlSVvBLL/sfN5azyf5zqp66OmPJEnS+sz698TWctMGfi9Jkk5qIyOWDfxekiSd1EZGzF+CliSNaiMjJknSqDYyYl/bwO8lSdJJzXIV+48k2ZHkuPtU1Ys3bixJkk5uliOx/wT8S+AzSa5L8u0DzSRJ0rqsO2JV9bGquhr4LuAQcFeSP07yE0nOHGpASZJOZKbPxJI8D3g98K+BB4BfZSVqd234ZJIkncS6r9iR5Fbg21n5peYfrKpHpk/9fpJ9QwwnSdJaZrmK/Qeq6o7VDyR5VlUdqaqlDZ5LkqSTmuXtxPcc57H/vlGDSJI0q5MeiSX5h8DFwNlJLufvLi/1bOCbBpxNkqQ1reftxH/ByskclwDXr3r8ceAdA8wkbVkv+qX/ypefeHLUNRd37R11vW8++0w++U7/dKB6OGnEqmoPsCfJK6vqIyPMJG1ZX37iSQ5dt2O09SaTCcvLy6OtB+NHU3o61vN24qur6neAxSRvOvb5qrr+OLsd3fcs4I+AZ03XuqWq3pnk24APAecD9wOvqSovWyVJmsl6Tuw4Z/rfc4Hzjrmde5J9jwDbq+pFwGXADyR5MfBe4H1V9ULgUeANpzC7JOk0t563E39z+uU/Bq6pqscAkjwX+JWT7FvA4endM6e3ArazcgkrgD3Au1i5rJUkSes2y++J/dOjAQOoqkenZyuuKckZwH7gBcCvA58FHquqp6ab/AUrZz8eb9+dwE6AhYUFJpPJDONKwxjzdXj48OFNed37s6YuZonYM5I8t6oeBUhy/nr2r6qvA5cleQ7wUeDS4212gn13A7sBlpaWauwPuKW/5869o55osRkndoz9b5Sejlki9ivAHye5hZXo/Chw7Xp3rqrHkkyAFwPPSbJtejR2CfD5GeaQJAmY7Sr2vw28EvgC8FfAK6rqprX2SXLh9AiMJGcD3w8cBO4BXjXd7HXAbbOPLkk63c1yJEZVPQw8PMMuF7HyO2ZnsBLMD1fV7UkeBj6U5D2sXA3/hlnmkCQJZozYrKrqU8DfO/mjqj4HXDHk2pKk+TfT3xOTJGkrMWKSpLaMmCSpLSMmSWrLiEmS2jJikqS2jJgkqS0jJklqy4hJktoyYpKktoyYJKktIyZJasuISZLaMmKSpLaMmCSpLSMmSWrLiEmS2jJikqS2jJgkqS0jJklqy4hJktoyYpKktoyYJKktIyZJasuISZLaMmKSpLaMmCSpLSMmSWrLiEmS2jJikqS2jJgkqS0jJklqy4hJktoyYpKktoyYJKktIyZJasuISZLaMmKSpLaMmCSpLSMmSWpr0Igl+dYk9yQ5mOShJNdMH39Xkv+b5MD09tIh55AkzadtA3//p4A3V9X9Sc4D9ie5a/rc+6rqPwy8viRpjg0asap6BHhk+vXjSQ4CFw+5piTp9DH0kdj/l2QRuBy4D7gS+KkkrwX2sXK09uhx9tkJ7ARYWFhgMpmMNa50QmO+Dg8fPrwpr3t/1tRFqmr4RZJzgU8A11bVrUkWgC8BBfwycFFV/au1vsfS0lLt27dv8FmltSzu2suh63aMtt5kMmF5eXm09WD8f6N0rCT7q2ppPdsOfnZikjOBjwA3V9WtAFX1har6elV9A/gt4Iqh55AkzZ+hz04McANwsKquX/X4Ras2+2Hg00POIUmaT0N/JnYl8BrgwSQHpo+9A/jxJJex8nbiIeDfDDyHJGkODX124r1AjvPUHUOuK0k6PXjFDklSW0ZMktSWEZMktWXEJEltjXbFDmkenHfpLv7Jnl3jLrpn3OXOuxTAX3ZWD0ZMmsHjB687La7YIXXh24mSpLaMmCSpLSMmSWrLiEmS2jJikqS2jJgkqS0jJklqy4hJktoyYpKktoyYJKktIyZJasuISZLaMmKSpLaMmCSpLSMmSWrLiEmS2jJikqS2jJgkqS0jJklqy4hJktoyYpKktoyYJKktIyZJasuISZLaMmKSpLaMmCSpLSMmSWrLiEmS2jJikqS2jJgkqS0jJklqy4hJktoyYpKktoyYJKmtQSOW5FuT3JPkYJKHklwzffz8JHcl+cz0v88dcg5J0nwa+kjsKeDNVXUp8GLgJ5N8B7ALuLuqXgjcPb0vSdJMBo1YVT1SVfdPv34cOAhcDPwQsGe62R7gqiHnkCTNp21jLZRkEbgcuA9YqKpHYCV0Sf7BCfbZCewEWFhYYDKZjDKrtJYxX4eHDx/elNe9P2vqYpSIJTkX+AjwM1X1lSTr2q+qdgO7AZaWlmp5eXmwGaV1uXMvY74OJ5PJqOsBo/8bpadj8LMTk5zJSsBurqpbpw9/IclF0+cvAr449BySpPkz9NmJAW4ADlbV9aue+gPgddOvXwfcNuQckqT5NPTbiVcCrwEeTHJg+tg7gOuADyd5A/C/gR8ZeA5J0hwaNGJVdS9wog/A/tmQa0uS5p9X7JAktWXEJEltGTFJUltGTJLU1mhX7JDmxeKuveMueOe4633z2WeOup70dBgxaQaHrtsx6nqLu/aOvqbUiW8nSpLaMmKSpLaMmCSpLSMmSWrLiEmS2jJikqS2jJgkqS0jJklqy4hJktoyYpKktoyYJKktIyZJasuISZLaMmKSpLaMmCSpLSMmSWrLiEmS2jJikqS2tm32ANLpIMmp7/veU9uvqk55TakLj8SkEVTVKd3uueeeU95XOh0YMUlSW0ZMktSWEZMktWXEJEltGTFJUltGTJLUlhGTJLVlxCRJbRkxSVJbRkyS1JYRkyS1ZcQkSW0ZMUlSW+lyteskfwX8+WbPIY3sAuBLmz2ENLLnV9WF69mwTcSk01GSfVW1tNlzSFuVbydKktoyYpKktoyYtLXt3uwBpK3Mz8QkSW15JCZJasuISZLaMmKSpLaMmDSQJG9M8trp169P8i2bPZM0bzyxQxpBkgnwlqrat9mzSPPEIzFpgyR5bZJPJflkkpuSvCvJW5K8ClgCbk5yIMmOJB9dtd9Lkty6xvc9nOS9SfYn+ViSK5JMknwuycun27w+yW1J7kzyv5K8c9X+v5DkT5LcleT3krxlyP8P0piMmLQBknwn8HPA9qp6EXDN0eeq6hZgH3B1VV0G3AFcmuToteF+ArhxjW9/DjCpqu8GHgfeA7wE+GHg3au2uwK4GrgM+JEkS0mWgFcClwOvYCWm0tzYttkDSHNiO3BLVX0JoKr+JslxN6yqSnIT8OokNwLfA7x2je/9NeDO6dcPAkeq6skkDwKLq7a7q6r+GmB6ZPe908dvq6onpo//4an846StyohJGyPALB8w3wj8IfC3wH+uqqfW2PbJ+rsPr78BHAGoqm8kWf0zfOz6he+2aM75Apc2xt3AjyZ5HkCS8495/nHgvKN3qurzwOeBnwc+uEEzvCTJ+UnOBq4C/htwL/CDSc5Kci6wY4PWkrYEj8SkDVBVDyW5FvhEkq8DDwCHVm3yQeA3kjwBfM/07b2bgQur6uENGuNe4CbgBcDvHj0TMskfAJ9k5e/x7QO+vEHrSZvOU+ylTZLkPwIPVNUNG/C9Xg8sVdVPHee5c6vqcJJvAv4I2FlV9z/dNaWtwCMxaRMk2Q98FXjzCMvtTvIdwFnAHgOmeeKRmLRFJLkPeNYxD7+mqh7cjHmkDoyYJKktz06UJLVlxCRJbRkxSVJbRkyS1Nb/Ax/QvIs0IBv7AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 1080x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAbAAAAFpCAYAAAD9f33NAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvnQurowAAFIlJREFUeJzt3XGQnHV5wPHv05xRiMSoyIHSek2t8WZOxbA4pXPBO69YFUXbUsdMOwN4Mxnt9FrtqERvrPjHzaTa0cm0MzrISdM2vdEiVFsKhept4UaIvUC0LQs6MkGogFALEkBi0qd/3KZe0uRuE3h397f3/czc3O67++77ZObgO++7774bmYkkSaX5uU4PIEnSiTBgkqQiGTBJUpEMmCSpSAZMklQkAyZJKpIBkyQVyYBJkopkwCRJRTJgkqQi9XV6gFaceuqpOTAw0OkxpLZ64oknWLNmTafHkNpq9+7dj2TmS1p5bhEBGxgYYH5+vtNjSG1Vr9cZGRnp9BhSW0XEva0+10OIkqQiGTBJUpEMmCSpSAZMklQkAyZJKpIBkyQVyYBJkopkwCRJRTJgkqQiGTCpy8zMzDA0NMTY2BhDQ0PMzMx0eiSpKxVxKSlppZiZmWFycpLp6WkOHjzIqlWrGB8fB2Dz5s0dnk7qLu6BSV1kamqK6elpRkdH6evrY3R0lOnpaaampjo9mtR1DJjURRqNBsPDw4ctGx4eptFodGgiqXsZMKmLDA4OMjc3d9iyubk5BgcHOzSR1L0MmNRFJicnGR8fZ3Z2lgMHDjA7O8v4+DiTk5OdHk3qOp7EIXWRQydqTExM0Gg0GBwcZGpqyhM4pKOIzKzmhSM2AF9ctGg98MfAy4C3A/uB7wGXZuajS71WrVZLv9BSK41faKmVKCJ2Z2atledWdggxM+/OzLMy8yzgbOBJ4FrgJmAoM18DfAf4SFUzSJJ6V7veAxsDvpeZ92bmjZl5oLn8NuDMNs0gSeoh7QrYu4GjXU7gPcD1bZpBktRDKj+JIyJWAxdyxKHCiJgEDgA7j7HeFmALQH9/P/V6vdpBpS6zb98+/+6lJbTjLMS3ALdn5kOHFkTExcDbgLE8xlkkmXkFcAUsnMThm9laaTyJQ1paOwK2mUWHDyPizcBlwBsy88k2bF+S1IMqfQ8sIk4GzgeuWbT4z4FTgJsiYk9EfK7KGSRJvanSPbDmHtaLj1j2iiq3KUlaGbyUlCSpSAZMklQkAyZJKpIBkyQVyYBJkopkwCRJRTJgkqQiGTCpy8zMzDA0NMTY2BhDQ0PMzBztOtiS/EZmqYvMzMwwOTnJ9PQ0Bw8eZNWqVYyPjwP4rczSEdwDk7rI1NQU09PTjI6O0tfXx+joKNPT00xNTXV6NKnrGDCpizQaDYaHhw9bNjw8TKPR6NBEUvcyYFIXGRwcZG5u7rBlc3NzDA4OdmgiqXsZMKmLTE5OMj4+zuzsLAcOHGB2dpbx8XEmJyc7PZrUdTyJQ+oih07UmJiYoNFoMDg4yNTUlCdwSEcRx/hC5K5Sq9Vyfn6+02NIbeU3MmsliojdmVlr5bkeQpQkFcmASZKKZMAkSUUyYJKkIhkwSVKRDJgkqUgGTJJUJAMmSSqSAZMkFcmASZKKZMAkSUUyYJKkIhkwSVKRDJgkqUgGTJJUJAMmSSqSAZMkFcmASZKKZMAkSUUyYJKkIhkwSVKRDJgkqUgGTJJUJAMmSSqSAZMkFamygEXEhojYs+jnxxHx/oh4UUTcFBHfbf5+YVUzSJJ6V2UBy8y7M/OszDwLOBt4ErgW2Ap8LTN/Gfha874kScelXYcQx4DvZea9wDuAHc3lO4B3tmkGSVIPaVfA3g3MNG/3Z+YDAM3fp7VpBklSD+mregMRsRq4EPjIca63BdgC0N/fT71ef/aHk7rYvn37/LuXllB5wIC3ALdn5kPN+w9FxBmZ+UBEnAH88GgrZeYVwBUAtVotR0ZG2jCq1D3q9Tr+3UvH1o5DiJv52eFDgK8CFzdvXwx8pQ0zSJJ6TKUBi4iTgfOBaxYt3gacHxHfbT62rcoZJEm9qdJDiJn5JPDiI5b9FwtnJUqSdMK8EockqUgGTJJUJAMmSSqSAZMkFcmASZKKZMAkSUUyYJKkIhkwSVKRDJgkqUgGTJJUJAMmSSqSAZMkFcmASZKKZMAkSUUyYJKkIhkwSVKRDJgkqUgGTJJUJAMmSSqSAZMkFcmASZKKZMAkSUUyYJKkIhkwSVKRDJgkqUgGTJJUJAMmSSqSAZMkFcmASZKKZMAkSUUyYJKkIhkwSVKRDJgkqUgGTJJUJAMmSSpSX6cHkHpdRLR9m5nZ9m1K7eYemFSxzDyhn5df9g8nvK60EhgwSVKRDJgkqUgGTJJUJAMmSSpSpQGLiHURcXVE3BURjYg4NyLOiojbImJPRMxHxOurnEGS1JuqPo1+O3BDZl4UEauBk4EvAZ/IzOsj4q3AJ4GRiueQJPWYygIWEWuB84BLADJzP7A/IhJY23zaC4AfVDWDJKl3VbkHth54GLgqIl4L7Ab+EHg/8E8R8acsHML81QpnkCT1qCoD1gdsBCYyc1dEbAe2srDX9YHM/HJEvAuYBn7tyJUjYguwBaC/v596vV7hqFJ38u9eOrao6lP7EXE6cFtmDjTvb2IhYMPAuszMWLjGzmOZufbYrwS1Wi3n5+crmVPqVgNbr2Pvtgs6PYbUVhGxOzNrrTy3srMQM/NB4L6I2NBcNAbcycJ7Xm9oLnsj8N2qZpAk9a6qz0KcAHY2z0C8B7gU+AqwPSL6gJ/QPEwoSdLxqDRgmbkHOHJXcA44u8rtSpJ6n1fikCQVyYBJkopkwCRJRTJgkqQiGTBJUpEMmCSpSAZMklQkAyZJKpIBkyQVyYBJkopkwCRJRTJgkqQiGTBJUpEMmCSpSAZMklQkAyZJKpIBkyQVyYBJkopkwCRJRTJgkqQiGTBJUpEMmCSpSAZMklQkAyZJKpIBkyQVyYBJkorU18qTIuKVwIeAly9eJzPfWNFckiQtqaWAAX8LfA74PHCwunEkSWpNqwE7kJmfrXQSSZKOw5IBi4gXNW/+fUT8HnAt8PShxzPzRxXOJknSMS23B7YbSCCa9z+06LEE1lcxlCRJy1kyYJn5iwAR8bzM/MnixyLieVUOJknSUlp9D+wbwMYWlkk967WfuJHHnvppW7c5sPW6tm7vBSc9h299/E1t3aZ0opZ7D+x04GXASRHxOn52KHEtcHLFs0ld5bGnfsrebRe0bXv1ep2RkZG2bQ/aH0zpmVhuD+zXgUuAM4FPL1r+OPDRimaSJGlZy70HtgPYERG/lZlfbtNMkiQta7lDiH90tNuHZOanj1wmSVI7LHcI8ZTm7w3AOcBXm/ffDtxc1VCSJC1nuUOInwCIiBuBjZn5ePP+5SxcXkqSpI5o9Wr0vwDsX3R/PzDwrE8jSVKLWg3YXwHfjIjLI+LjwC7gL5dbKSLWRcTVEXFXRDQi4tzm8omIuDsi/iMiPnni40uSVqqWPsicmVMRcT2wqbno0sy8o4VVtwM3ZOZFEbEaODkiRoF3AK/JzKcj4rQTmlyStKItdxbi2sz8cfOivnubP4cee9FSF/ONiLXAeSx8jozM3A/sj4j3Adsy8+nm8h8+w3+DJGkFWu4Q4t80f+8G5ps/uxfdX8p64GHgqoi4IyKujIg1wCuBTRGxKyL+JSLOOfHxJUkr1XJnIb6teXOOhdPmb8nMu47jtTcCE5m5KyK2A1uby18I/AoLp+Z/KSLWZ2YuXjkitgBbAPr7+6nX6y1uVqpOO/8O9+3b15G/e/9bUylavZjvVcAw8GcRsR64g4WYbV9infuB+zNzV/P+1SwE7H7gmmawvhkR/wOcysLe2v/JzCuAKwBqtVq2+5pw0v9zw3VtvTZhJ66F2O5/o/RMtHQWYmZ+HZgCPgZcycKe0/uWWedB4L6I2NBcNAbcCfwd8EaAiHglsBp45ESGlyStXC3tgUXE14A1wK3ALcA5LZ58MQHsbJ6BeA9wKfAE8IWI+HcWPk928ZGHDyVJWk6rhxC/DZwNDAGPAY9GxK2Z+dRSK2XmHqB2lId+97imlCTpCK1+DuwDABHxfBb2oq4CTgeeW91okiQdW6uHEH+fhQ8xnw3cC3yBhUOJkiR1RKuHEE9i4Qstd2fmgQrnkSSpJa0eQvxU1YNIknQ8Wr2YryRJXcWASZKKZMAkSUVq9SQOacU7ZXArr96xtb0b3dHezZ0yCHBBezcqnSADJrXo8cY29m5r3//cO3EtxIGt17V1e9Iz4SFESVKRDJgkqUgGTJJUJAMmSSqSAZMkFcmASZKKZMAkSUUyYJKkIhkwSVKRDJgkqUgGTJJUJAMmSSqSAZMkFcmASZKKZMAkSUUyYJKkIhkwSVKRDJgkqUgGTJJUJAMmSSqSAZMkFcmASZKKZMAkSUUyYJKkIhkwSVKRDJgkqUgGTJJUJAMmSSpSX6cHkEoysPW69m7whvZu7wUnPaet25OeCQMmtWjvtgvaur2Brde1fZtSSTyEKEkqkgGTJBWp0oBFxLqIuDoi7oqIRkScu+ixD0ZERsSpVc4gSepNVb8Hth24ITMviojVwMkAEfHzwPnA9yveviSpR1W2BxYRa4HzgGmAzNyfmY82H/4M8GEgq9q+JKm3VXkIcT3wMHBVRNwREVdGxJqIuBD4z8z8VoXbliT1uCoPIfYBG4GJzNwVEduBy1nYK3vTcitHxBZgC0B/fz/1er26SaUu5d+9dGyRWc1RvIg4HbgtMwea9zexELBXA082n3Ym8APg9Zn54LFeq1ar5fz8fCVzSt3Kz4FpJYqI3ZlZa+W5lR1CbAbpvojY0Fw0Btyemadl5kAzbPcDG5eKlyRJR1P1WYgTwM7mGYj3AJdWvD1J0gpRacAycw9wzF3BQ4cXJUk6Xl6JQ5JUJAMmSSqSAZMkFcmASZKKZMAkSUUyYJKkIhkwSVKRDJgkqUgGTJJUpKovJSWteBFx4uv+yYmtV9VFuqVu4h6YVLHMPKGf2dnZE15XWgkMmCSpSAZMklQkAyZJKpIBkyQVyYBJkopkwCRJRTJgkqQiGTBJUpEMmCSpSAZMklQkAyZJKpIBkyQVyYBJkopkwCRJRTJgkqQiGTBJUpEMmCSpSAZMklQkAyZJKpIBkyQVyYBJkopkwCRJRTJgkqQiGTBJUpEMmCSpSAZMklQkAyZJKpIBkyQVyYBJkopUacAiYl1EXB0Rd0VEIyLOjYhPNe9/OyKujYh1Vc4gSepNVe+BbQduyMxXAa8FGsBNwFBmvgb4DvCRimeQJPWgygIWEWuB84BpgMzcn5mPZuaNmXmg+bTbgDOrmkGS1Luq3ANbDzwMXBURd0TElRGx5ojnvAe4vsIZJEk9qq/i194ITGTmrojYDmwFPgYQEZPAAWDn0VaOiC3AFoD+/n7q9XqFo0rdZ9++ff7dS0uIzKzmhSNOB27LzIHm/U3A1sy8ICIuBt4LjGXmk8u9Vq1Wy/n5+UrmlLpVvV5nZGSk02NIbRURuzOz1spzKzuEmJkPAvdFxIbmojHgzoh4M3AZcGEr8ZIk6WiqPIQIMAHsjIjVwD3ApcC/As8FbooIWNhLe2/Fc0iSekylAcvMPcCRu4KvqHKbkqSVwStxSJKKZMAkSUUyYJKkIhkwSVKRDJgkqUgGTJJUJAMmSSqSAZMkFcmASZKKZMAkSUUyYJKkIhkwSVKRDJgkqUgGTJJUJAMmSSqSAZMkFcmASZKKZMAkSUUyYJKkIhkwSVKRDJgkqUgGTJJUJAMmSSqSAZMkFcmASZKKZMAkSUUyYJKkIhkwSVKRDJgkqUgGTJJUJAMmSSqSAZMkFcmASZKKZMAkSUUyYFKXmZmZYWhoiLGxMYaGhpiZmen0SFJX6uv0AJJ+ZmZmhsnJSaanpzl48CCrVq1ifHwcgM2bN3d4Oqm7uAcmdZGpqSmmp6cZHR2lr6+P0dFRpqenmZqa6vRoUtcxYFIXaTQaDA8PH7ZseHiYRqPRoYmk7mXApC4yODjI3NzcYcvm5uYYHBzs0ERS9zJgUheZnJxkfHyc2dlZDhw4wOzsLOPj40xOTnZ6NKnrVHoSR0SsA64EhoAE3gPcDXwRGAD2Au/KzP+ucg6pFIdO1JiYmKDRaDA4OMjU1JQncEhHEZlZ3YtH7ABuycwrI2I1cDLwUeBHmbktIrYCL8zMy5Z6nVqtlvPz85XNKXWjer3OyMhIp8eQ2ioidmdmrZXnVnYIMSLWAucB0wCZuT8zHwXeAexoPm0H8M6qZpAk9a4q3wNbDzwMXBURd0TElRGxBujPzAcAmr9Pq3AGSVKPqvI9sD5gIzCRmbsiYjuwtdWVI2ILsAWgv7+fer1eyZBSt9q3b59/99ISKnsPLCJOB27LzIHm/U0sBOwVwEhmPhARZwD1zNyw1Gv5HphWIt8D00rUFe+BZeaDwH0RcShOY8CdwFeBi5vLLga+UtUMkqTeVfW1ECeAnc0zEO8BLmUhml+KiHHg+8BvVzyDJKkHVRqwzNwDHG1XcKzK7UqSep9X4pAkFcmASZKKZMAkSUWq9FJSz5aIeBi4t9NzSG12KvBIp4eQ2uzlmfmSVp5YRMCklSgi5lv9PIy0EnkIUZJUJAMmSSqSAZO61xWdHkDqZr4HJkkqkntgkqQiGTCpC0TEP0bEuqMsvzwiPti8fUlEvHTRY3sj4tR2zil1EwMmdYHMfGvzG8uXcgnw0mWeI60YBkxqg4j4cET8QfP2ZyLi683bYxHx14v3piJiMiLujoh/BjY0l13EwoWxd0bEnog4qfnSExFxe0T8W0S8qv3/MqlzDJjUHjcDm5q3a8DzI+I5wDBwy6EnRcTZwLuB1wG/CZwDkJlXA/PA72TmWZn5VHOVRzJzI/BZ4IPt+IdI3cKASe2xGzg7Ik4BngZuZSFkm1gUsOb9azPzycz8MQtfALuUaxa9/sCzOrHU5ar+QktJQGb+NCL2svClrt8Avg2MAr8ENI58+nG89NPN3wfxv2etMO6BSe1zMwuH+W5mYa/rvcCePPzDmDcDvxERJzX31t6+6LHHgVPaNazU7QyY1D63AGcAt2bmQ8BPOPzwIZl5O/BFYA/w5SMe/wvgc0ecxCGtWF6JQ5JUJPfAJElFMmCSpCIZMElSkQyYJKlIBkySVCQDJkkqkgGTJBXJgEmSivS/0a1pxnZA0OIAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 1080x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "for var in numerical:\n",
    "    plt.figure(figsize=(15,6))\n",
    "    plt.subplot(1, 2, 1)\n",
    "    fig = df.boxplot(column=var)\n",
    "    fig.set_title('')\n",
    "    fig.set_ylabel(var)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The variables presenting outliers are:\n",
    "    - bore (top)\n",
    "    - compression_ratio (both)\n",
    "    - Height (top)\n",
    "    - highway_mpg (top)\n",
    "    - horse_power (top)\n",
    "    - lenght (both)\n",
    "    - normalized_losses (top)\n",
    "    - stroke (bottom)\n",
    "    - wheel_base (both)\n",
    "    - width (both)\n",
    "   "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "bore skewness is 0.6257923004524101\n",
      "compression_ratio skewness is 5.190962998330603\n",
      "height skewness is 0.27350172478135837\n",
      "highway_mpg skewness is 0.954946036021284\n",
      "horsepower skewness is 2.4448428600495635\n",
      "length skewness is -0.16070860345381996\n",
      "normalized_losses skewness is 0.8716855672848831\n",
      "stroke skewness is -1.212937019722827\n",
      "wheel_base skewness is 1.2071569265180229\n",
      "width skewness is 0.8650397680572235\n"
     ]
    }
   ],
   "source": [
    "# let's find the skewness of above variables\n",
    "for var in ['bore',\n",
    " 'compression_ratio',\n",
    " 'height',\n",
    " 'highway_mpg',\n",
    " 'horsepower',\n",
    " 'length',\n",
    " 'normalized_losses',\n",
    " 'stroke',\n",
    " 'wheel_base',\n",
    " 'width']:\n",
    "    print(var, 'skewness is', df[var].skew() )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "If we consider an absolute value of 1 to be normal skewness. \n",
    "\n",
    "I shall deal with outliers as follows:\n",
    "\n",
    "    - Equal binning for variables with a skewness above 1:\n",
    "        - stroke -done\n",
    "        - horsepower -done\n",
    "        - compression_ratio - not yet\n",
    "        - wheel_balse -not yet\n",
    "        \n",
    "    - Bottom/top capping for others. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 8.2 Equal binning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['Q1', 'Q2', 'Q3', 'Q4', 'Q5', 'Q6', 'Q7', 'Q8', 'Q9', 'Q10']\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>aspiration</th>\n",
       "      <th>body_style</th>\n",
       "      <th>bore</th>\n",
       "      <th>city_mpg</th>\n",
       "      <th>compression_ratio</th>\n",
       "      <th>curb_weight</th>\n",
       "      <th>drive_wheels</th>\n",
       "      <th>engine_location</th>\n",
       "      <th>engine_size</th>\n",
       "      <th>engine_type</th>\n",
       "      <th>fuel_system</th>\n",
       "      <th>fuel_type</th>\n",
       "      <th>height</th>\n",
       "      <th>highway_mpg</th>\n",
       "      <th>horsepower</th>\n",
       "      <th>length</th>\n",
       "      <th>make_id</th>\n",
       "      <th>normalized_losses</th>\n",
       "      <th>num_of_cylinders</th>\n",
       "      <th>num_of_doors</th>\n",
       "      <th>peak_rpm</th>\n",
       "      <th>price</th>\n",
       "      <th>stroke</th>\n",
       "      <th>symboling</th>\n",
       "      <th>wheel_base</th>\n",
       "      <th>width</th>\n",
       "      <th>normalized_losses_NA</th>\n",
       "      <th>stroke_label</th>\n",
       "      <th>stroke_disc</th>\n",
       "      <th>horsepower_label</th>\n",
       "      <th>horsepower_disc</th>\n",
       "      <th>wheel_base_label</th>\n",
       "      <th>wheel_base_disc</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>193</th>\n",
       "      <td>std</td>\n",
       "      <td>wagon</td>\n",
       "      <td>3.19</td>\n",
       "      <td>25</td>\n",
       "      <td>9.0</td>\n",
       "      <td>2563</td>\n",
       "      <td>fwd</td>\n",
       "      <td>front</td>\n",
       "      <td>109</td>\n",
       "      <td>ohc</td>\n",
       "      <td>mpfi</td>\n",
       "      <td>gas</td>\n",
       "      <td>55.1</td>\n",
       "      <td>31</td>\n",
       "      <td>88.0</td>\n",
       "      <td>183.1</td>\n",
       "      <td>21</td>\n",
       "      <td>114.0</td>\n",
       "      <td>four</td>\n",
       "      <td>four</td>\n",
       "      <td>5500.0</td>\n",
       "      <td>12290.0</td>\n",
       "      <td>3.40</td>\n",
       "      <td>0</td>\n",
       "      <td>100.4</td>\n",
       "      <td>66.9</td>\n",
       "      <td>1</td>\n",
       "      <td>Q7</td>\n",
       "      <td>(3.39, 3.4]</td>\n",
       "      <td>Q4</td>\n",
       "      <td>(77.8, 88.0]</td>\n",
       "      <td>Q7</td>\n",
       "      <td>(98.8, 100.4]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>125</th>\n",
       "      <td>std</td>\n",
       "      <td>hatchback</td>\n",
       "      <td>3.94</td>\n",
       "      <td>19</td>\n",
       "      <td>9.5</td>\n",
       "      <td>2778</td>\n",
       "      <td>rwd</td>\n",
       "      <td>front</td>\n",
       "      <td>151</td>\n",
       "      <td>ohc</td>\n",
       "      <td>mpfi</td>\n",
       "      <td>gas</td>\n",
       "      <td>50.2</td>\n",
       "      <td>27</td>\n",
       "      <td>143.0</td>\n",
       "      <td>168.9</td>\n",
       "      <td>16</td>\n",
       "      <td>186.0</td>\n",
       "      <td>four</td>\n",
       "      <td>two</td>\n",
       "      <td>5500.0</td>\n",
       "      <td>22018.0</td>\n",
       "      <td>3.11</td>\n",
       "      <td>3</td>\n",
       "      <td>94.5</td>\n",
       "      <td>68.3</td>\n",
       "      <td>0</td>\n",
       "      <td>Q3</td>\n",
       "      <td>(3.07, 3.15]</td>\n",
       "      <td>Q9</td>\n",
       "      <td>(140.8, 160.0]</td>\n",
       "      <td>Q2</td>\n",
       "      <td>(93.7, 94.5]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>173</th>\n",
       "      <td>std</td>\n",
       "      <td>sedan</td>\n",
       "      <td>3.31</td>\n",
       "      <td>29</td>\n",
       "      <td>8.7</td>\n",
       "      <td>2326</td>\n",
       "      <td>fwd</td>\n",
       "      <td>front</td>\n",
       "      <td>122</td>\n",
       "      <td>ohc</td>\n",
       "      <td>mpfi</td>\n",
       "      <td>gas</td>\n",
       "      <td>54.9</td>\n",
       "      <td>34</td>\n",
       "      <td>92.0</td>\n",
       "      <td>175.6</td>\n",
       "      <td>20</td>\n",
       "      <td>65.0</td>\n",
       "      <td>four</td>\n",
       "      <td>four</td>\n",
       "      <td>4200.0</td>\n",
       "      <td>8948.0</td>\n",
       "      <td>3.54</td>\n",
       "      <td>-1</td>\n",
       "      <td>102.4</td>\n",
       "      <td>66.5</td>\n",
       "      <td>0</td>\n",
       "      <td>Q9</td>\n",
       "      <td>(3.46, 3.54]</td>\n",
       "      <td>Q5</td>\n",
       "      <td>(88.0, 95.0]</td>\n",
       "      <td>Q8</td>\n",
       "      <td>(100.4, 103.5]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>92</th>\n",
       "      <td>std</td>\n",
       "      <td>sedan</td>\n",
       "      <td>3.15</td>\n",
       "      <td>31</td>\n",
       "      <td>9.4</td>\n",
       "      <td>1938</td>\n",
       "      <td>fwd</td>\n",
       "      <td>front</td>\n",
       "      <td>97</td>\n",
       "      <td>ohc</td>\n",
       "      <td>2bbl</td>\n",
       "      <td>gas</td>\n",
       "      <td>54.5</td>\n",
       "      <td>37</td>\n",
       "      <td>69.0</td>\n",
       "      <td>165.3</td>\n",
       "      <td>13</td>\n",
       "      <td>122.0</td>\n",
       "      <td>four</td>\n",
       "      <td>four</td>\n",
       "      <td>5200.0</td>\n",
       "      <td>6849.0</td>\n",
       "      <td>3.29</td>\n",
       "      <td>1</td>\n",
       "      <td>94.5</td>\n",
       "      <td>63.8</td>\n",
       "      <td>0</td>\n",
       "      <td>Q5</td>\n",
       "      <td>(3.23, 3.29]</td>\n",
       "      <td>Q2</td>\n",
       "      <td>(68.0, 69.0]</td>\n",
       "      <td>Q2</td>\n",
       "      <td>(93.7, 94.5]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>std</td>\n",
       "      <td>sedan</td>\n",
       "      <td>3.62</td>\n",
       "      <td>16</td>\n",
       "      <td>8.0</td>\n",
       "      <td>3380</td>\n",
       "      <td>rwd</td>\n",
       "      <td>front</td>\n",
       "      <td>209</td>\n",
       "      <td>ohc</td>\n",
       "      <td>mpfi</td>\n",
       "      <td>gas</td>\n",
       "      <td>53.7</td>\n",
       "      <td>22</td>\n",
       "      <td>182.0</td>\n",
       "      <td>193.8</td>\n",
       "      <td>3</td>\n",
       "      <td>114.0</td>\n",
       "      <td>six</td>\n",
       "      <td>two</td>\n",
       "      <td>5400.0</td>\n",
       "      <td>41315.0</td>\n",
       "      <td>3.39</td>\n",
       "      <td>0</td>\n",
       "      <td>103.5</td>\n",
       "      <td>67.9</td>\n",
       "      <td>1</td>\n",
       "      <td>Q6</td>\n",
       "      <td>(3.29, 3.39]</td>\n",
       "      <td>Q10</td>\n",
       "      <td>(160.0, 262.0]</td>\n",
       "      <td>Q8</td>\n",
       "      <td>(100.4, 103.5]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>189</th>\n",
       "      <td>std</td>\n",
       "      <td>convertible</td>\n",
       "      <td>3.19</td>\n",
       "      <td>24</td>\n",
       "      <td>8.5</td>\n",
       "      <td>2254</td>\n",
       "      <td>fwd</td>\n",
       "      <td>front</td>\n",
       "      <td>109</td>\n",
       "      <td>ohc</td>\n",
       "      <td>mpfi</td>\n",
       "      <td>gas</td>\n",
       "      <td>55.6</td>\n",
       "      <td>29</td>\n",
       "      <td>90.0</td>\n",
       "      <td>159.3</td>\n",
       "      <td>21</td>\n",
       "      <td>114.0</td>\n",
       "      <td>four</td>\n",
       "      <td>two</td>\n",
       "      <td>5500.0</td>\n",
       "      <td>11595.0</td>\n",
       "      <td>3.40</td>\n",
       "      <td>3</td>\n",
       "      <td>94.5</td>\n",
       "      <td>64.2</td>\n",
       "      <td>1</td>\n",
       "      <td>Q7</td>\n",
       "      <td>(3.39, 3.4]</td>\n",
       "      <td>Q5</td>\n",
       "      <td>(88.0, 95.0]</td>\n",
       "      <td>Q2</td>\n",
       "      <td>(93.7, 94.5]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>136</th>\n",
       "      <td>turbo</td>\n",
       "      <td>hatchback</td>\n",
       "      <td>3.54</td>\n",
       "      <td>19</td>\n",
       "      <td>9.0</td>\n",
       "      <td>2808</td>\n",
       "      <td>fwd</td>\n",
       "      <td>front</td>\n",
       "      <td>121</td>\n",
       "      <td>dohc</td>\n",
       "      <td>mpfi</td>\n",
       "      <td>gas</td>\n",
       "      <td>56.1</td>\n",
       "      <td>26</td>\n",
       "      <td>160.0</td>\n",
       "      <td>186.6</td>\n",
       "      <td>18</td>\n",
       "      <td>150.0</td>\n",
       "      <td>four</td>\n",
       "      <td>two</td>\n",
       "      <td>5500.0</td>\n",
       "      <td>18150.0</td>\n",
       "      <td>3.07</td>\n",
       "      <td>3</td>\n",
       "      <td>99.1</td>\n",
       "      <td>66.5</td>\n",
       "      <td>0</td>\n",
       "      <td>Q2</td>\n",
       "      <td>(2.879, 3.07]</td>\n",
       "      <td>Q9</td>\n",
       "      <td>(140.8, 160.0]</td>\n",
       "      <td>Q7</td>\n",
       "      <td>(98.8, 100.4]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>std</td>\n",
       "      <td>hatchback</td>\n",
       "      <td>2.97</td>\n",
       "      <td>31</td>\n",
       "      <td>9.4</td>\n",
       "      <td>1876</td>\n",
       "      <td>fwd</td>\n",
       "      <td>front</td>\n",
       "      <td>90</td>\n",
       "      <td>ohc</td>\n",
       "      <td>2bbl</td>\n",
       "      <td>gas</td>\n",
       "      <td>50.8</td>\n",
       "      <td>38</td>\n",
       "      <td>68.0</td>\n",
       "      <td>157.3</td>\n",
       "      <td>5</td>\n",
       "      <td>118.0</td>\n",
       "      <td>four</td>\n",
       "      <td>two</td>\n",
       "      <td>5500.0</td>\n",
       "      <td>6377.0</td>\n",
       "      <td>3.23</td>\n",
       "      <td>1</td>\n",
       "      <td>93.7</td>\n",
       "      <td>63.8</td>\n",
       "      <td>0</td>\n",
       "      <td>Q4</td>\n",
       "      <td>(3.15, 3.23]</td>\n",
       "      <td>Q1</td>\n",
       "      <td>(51.999, 68.0]</td>\n",
       "      <td>Q1</td>\n",
       "      <td>(86.59899999999999, 93.7]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>74</th>\n",
       "      <td>std</td>\n",
       "      <td>hardtop</td>\n",
       "      <td>3.80</td>\n",
       "      <td>14</td>\n",
       "      <td>8.0</td>\n",
       "      <td>3715</td>\n",
       "      <td>rwd</td>\n",
       "      <td>front</td>\n",
       "      <td>304</td>\n",
       "      <td>ohcv</td>\n",
       "      <td>mpfi</td>\n",
       "      <td>gas</td>\n",
       "      <td>55.4</td>\n",
       "      <td>16</td>\n",
       "      <td>184.0</td>\n",
       "      <td>199.2</td>\n",
       "      <td>10</td>\n",
       "      <td>114.0</td>\n",
       "      <td>eight</td>\n",
       "      <td>two</td>\n",
       "      <td>4500.0</td>\n",
       "      <td>45400.0</td>\n",
       "      <td>3.35</td>\n",
       "      <td>1</td>\n",
       "      <td>112.0</td>\n",
       "      <td>72.0</td>\n",
       "      <td>1</td>\n",
       "      <td>Q6</td>\n",
       "      <td>(3.29, 3.39]</td>\n",
       "      <td>Q10</td>\n",
       "      <td>(160.0, 262.0]</td>\n",
       "      <td>Q10</td>\n",
       "      <td>(107.97, 120.9]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44</th>\n",
       "      <td>std</td>\n",
       "      <td>sedan</td>\n",
       "      <td>3.03</td>\n",
       "      <td>38</td>\n",
       "      <td>9.6</td>\n",
       "      <td>1874</td>\n",
       "      <td>fwd</td>\n",
       "      <td>front</td>\n",
       "      <td>90</td>\n",
       "      <td>ohc</td>\n",
       "      <td>2bbl</td>\n",
       "      <td>gas</td>\n",
       "      <td>52.0</td>\n",
       "      <td>43</td>\n",
       "      <td>70.0</td>\n",
       "      <td>155.9</td>\n",
       "      <td>7</td>\n",
       "      <td>114.0</td>\n",
       "      <td>four</td>\n",
       "      <td>two</td>\n",
       "      <td>5400.0</td>\n",
       "      <td>10646.5</td>\n",
       "      <td>3.11</td>\n",
       "      <td>1</td>\n",
       "      <td>94.5</td>\n",
       "      <td>63.6</td>\n",
       "      <td>1</td>\n",
       "      <td>Q3</td>\n",
       "      <td>(3.07, 3.15]</td>\n",
       "      <td>Q3</td>\n",
       "      <td>(69.0, 77.8]</td>\n",
       "      <td>Q2</td>\n",
       "      <td>(93.7, 94.5]</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    aspiration   body_style  bore  city_mpg  compression_ratio  curb_weight  \\\n",
       "193        std        wagon  3.19        25                9.0         2563   \n",
       "125        std    hatchback  3.94        19                9.5         2778   \n",
       "173        std        sedan  3.31        29                8.7         2326   \n",
       "92         std        sedan  3.15        31                9.4         1938   \n",
       "16         std        sedan  3.62        16                8.0         3380   \n",
       "189        std  convertible  3.19        24                8.5         2254   \n",
       "136      turbo    hatchback  3.54        19                9.0         2808   \n",
       "22         std    hatchback  2.97        31                9.4         1876   \n",
       "74         std      hardtop  3.80        14                8.0         3715   \n",
       "44         std        sedan  3.03        38                9.6         1874   \n",
       "\n",
       "    drive_wheels engine_location  engine_size engine_type fuel_system  \\\n",
       "193          fwd           front          109         ohc        mpfi   \n",
       "125          rwd           front          151         ohc        mpfi   \n",
       "173          fwd           front          122         ohc        mpfi   \n",
       "92           fwd           front           97         ohc        2bbl   \n",
       "16           rwd           front          209         ohc        mpfi   \n",
       "189          fwd           front          109         ohc        mpfi   \n",
       "136          fwd           front          121        dohc        mpfi   \n",
       "22           fwd           front           90         ohc        2bbl   \n",
       "74           rwd           front          304        ohcv        mpfi   \n",
       "44           fwd           front           90         ohc        2bbl   \n",
       "\n",
       "    fuel_type  height  highway_mpg  horsepower  length  make_id  \\\n",
       "193       gas    55.1           31        88.0   183.1       21   \n",
       "125       gas    50.2           27       143.0   168.9       16   \n",
       "173       gas    54.9           34        92.0   175.6       20   \n",
       "92        gas    54.5           37        69.0   165.3       13   \n",
       "16        gas    53.7           22       182.0   193.8        3   \n",
       "189       gas    55.6           29        90.0   159.3       21   \n",
       "136       gas    56.1           26       160.0   186.6       18   \n",
       "22        gas    50.8           38        68.0   157.3        5   \n",
       "74        gas    55.4           16       184.0   199.2       10   \n",
       "44        gas    52.0           43        70.0   155.9        7   \n",
       "\n",
       "     normalized_losses num_of_cylinders num_of_doors  peak_rpm    price  \\\n",
       "193              114.0             four         four    5500.0  12290.0   \n",
       "125              186.0             four          two    5500.0  22018.0   \n",
       "173               65.0             four         four    4200.0   8948.0   \n",
       "92               122.0             four         four    5200.0   6849.0   \n",
       "16               114.0              six          two    5400.0  41315.0   \n",
       "189              114.0             four          two    5500.0  11595.0   \n",
       "136              150.0             four          two    5500.0  18150.0   \n",
       "22               118.0             four          two    5500.0   6377.0   \n",
       "74               114.0            eight          two    4500.0  45400.0   \n",
       "44               114.0             four          two    5400.0  10646.5   \n",
       "\n",
       "     stroke  symboling  wheel_base  width  normalized_losses_NA stroke_label  \\\n",
       "193    3.40          0       100.4   66.9                     1           Q7   \n",
       "125    3.11          3        94.5   68.3                     0           Q3   \n",
       "173    3.54         -1       102.4   66.5                     0           Q9   \n",
       "92     3.29          1        94.5   63.8                     0           Q5   \n",
       "16     3.39          0       103.5   67.9                     1           Q6   \n",
       "189    3.40          3        94.5   64.2                     1           Q7   \n",
       "136    3.07          3        99.1   66.5                     0           Q2   \n",
       "22     3.23          1        93.7   63.8                     0           Q4   \n",
       "74     3.35          1       112.0   72.0                     1           Q6   \n",
       "44     3.11          1        94.5   63.6                     1           Q3   \n",
       "\n",
       "       stroke_disc horsepower_label horsepower_disc wheel_base_label  \\\n",
       "193    (3.39, 3.4]               Q4    (77.8, 88.0]               Q7   \n",
       "125   (3.07, 3.15]               Q9  (140.8, 160.0]               Q2   \n",
       "173   (3.46, 3.54]               Q5    (88.0, 95.0]               Q8   \n",
       "92    (3.23, 3.29]               Q2    (68.0, 69.0]               Q2   \n",
       "16    (3.29, 3.39]              Q10  (160.0, 262.0]               Q8   \n",
       "189    (3.39, 3.4]               Q5    (88.0, 95.0]               Q2   \n",
       "136  (2.879, 3.07]               Q9  (140.8, 160.0]               Q7   \n",
       "22    (3.15, 3.23]               Q1  (51.999, 68.0]               Q1   \n",
       "74    (3.29, 3.39]              Q10  (160.0, 262.0]              Q10   \n",
       "44    (3.07, 3.15]               Q3    (69.0, 77.8]               Q2   \n",
       "\n",
       "               wheel_base_disc  \n",
       "193              (98.8, 100.4]  \n",
       "125               (93.7, 94.5]  \n",
       "173             (100.4, 103.5]  \n",
       "92                (93.7, 94.5]  \n",
       "16              (100.4, 103.5]  \n",
       "189               (93.7, 94.5]  \n",
       "136              (98.8, 100.4]  \n",
       "22   (86.59899999999999, 93.7]  \n",
       "74             (107.97, 120.9]  \n",
       "44                (93.7, 94.5]  "
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# and now, I will divide into 10 quantiles for the rest of the exercise. I will leave the quantile\n",
    "# boundary and generate labels as well for comparison\n",
    "\n",
    "# create 10 labels, one for each quantile\n",
    "labels = ['Q'+str(i+1) for i in range(0,10)]\n",
    "print(labels)\n",
    "\n",
    "\n",
    "# bins with labels\n",
    "X_train['stroke_label'], bins = pd.qcut(x=X_train.stroke, q=10, labels=labels, retbins=True, precision=3, duplicates='raise')\n",
    "\n",
    "# bins with boundaries\n",
    "X_train['stroke_disc'], bins = pd.qcut(x=X_train.stroke, q=10, retbins=True, precision=3, duplicates='raise')\n",
    "\n",
    "# bins with labels\n",
    "X_train['horsepower_label'], bins = pd.qcut(x=X_train.horsepower, q=10, labels=labels, retbins=True, precision=3, duplicates='raise')\n",
    "\n",
    "# bins with boundaries\n",
    "X_train['horsepower_disc'], bins = pd.qcut(x=X_train.horsepower, q=10, retbins=True, precision=3, duplicates='raise')\n",
    "\n",
    "# bins with labels\n",
    "X_train['wheel_base_label'], bins = pd.qcut(x=X_train.wheel_base, q=10, labels=labels, retbins=True, precision=3, duplicates='raise')\n",
    "\n",
    "# bins with boundaries\n",
    "X_train['wheel_base_disc'], bins = pd.qcut(x=X_train.wheel_base, q=10, retbins=True, precision=3, duplicates='raise')\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "X_train.head(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['Q1', 'Q2', 'Q3', 'Q4', 'Q5']\n"
     ]
    }
   ],
   "source": [
    "# create 10 labels, one for each quantile\n",
    "labels = ['Q'+str(i+1) for i in range(0,5)]\n",
    "print(labels)\n",
    "\n",
    "# bins with labels\n",
    "X_train['compression_ratio_label'], bins = pd.qcut(x=X_train.compression_ratio, q=5, labels=labels, retbins=True, precision=3, duplicates='raise')\n",
    "\n",
    "# bins with boundaries\n",
    "X_train['compression_ratio_disc'], bins = pd.qcut(x=X_train.compression_ratio, q=5, retbins=True, precision=3, duplicates='raise')\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>aspiration</th>\n",
       "      <th>body_style</th>\n",
       "      <th>bore</th>\n",
       "      <th>city_mpg</th>\n",
       "      <th>compression_ratio</th>\n",
       "      <th>curb_weight</th>\n",
       "      <th>drive_wheels</th>\n",
       "      <th>engine_location</th>\n",
       "      <th>engine_size</th>\n",
       "      <th>engine_type</th>\n",
       "      <th>fuel_system</th>\n",
       "      <th>fuel_type</th>\n",
       "      <th>height</th>\n",
       "      <th>highway_mpg</th>\n",
       "      <th>horsepower</th>\n",
       "      <th>length</th>\n",
       "      <th>make_id</th>\n",
       "      <th>normalized_losses</th>\n",
       "      <th>num_of_cylinders</th>\n",
       "      <th>num_of_doors</th>\n",
       "      <th>peak_rpm</th>\n",
       "      <th>price</th>\n",
       "      <th>stroke</th>\n",
       "      <th>symboling</th>\n",
       "      <th>wheel_base</th>\n",
       "      <th>width</th>\n",
       "      <th>normalized_losses_NA</th>\n",
       "      <th>stroke_label</th>\n",
       "      <th>stroke_disc</th>\n",
       "      <th>horsepower_label</th>\n",
       "      <th>horsepower_disc</th>\n",
       "      <th>wheel_base_label</th>\n",
       "      <th>wheel_base_disc</th>\n",
       "      <th>compression_ratio_label</th>\n",
       "      <th>compression_ratio_disc</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>193</th>\n",
       "      <td>std</td>\n",
       "      <td>wagon</td>\n",
       "      <td>3.19</td>\n",
       "      <td>25</td>\n",
       "      <td>9.0</td>\n",
       "      <td>2563</td>\n",
       "      <td>fwd</td>\n",
       "      <td>front</td>\n",
       "      <td>109</td>\n",
       "      <td>ohc</td>\n",
       "      <td>mpfi</td>\n",
       "      <td>gas</td>\n",
       "      <td>55.1</td>\n",
       "      <td>31</td>\n",
       "      <td>88.0</td>\n",
       "      <td>183.1</td>\n",
       "      <td>21</td>\n",
       "      <td>114.0</td>\n",
       "      <td>four</td>\n",
       "      <td>four</td>\n",
       "      <td>5500.0</td>\n",
       "      <td>12290.0</td>\n",
       "      <td>3.40</td>\n",
       "      <td>0</td>\n",
       "      <td>100.4</td>\n",
       "      <td>66.9</td>\n",
       "      <td>1</td>\n",
       "      <td>Q7</td>\n",
       "      <td>(3.39, 3.4]</td>\n",
       "      <td>Q4</td>\n",
       "      <td>(77.8, 88.0]</td>\n",
       "      <td>Q7</td>\n",
       "      <td>(98.8, 100.4]</td>\n",
       "      <td>Q2</td>\n",
       "      <td>(8.5, 9.0]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>125</th>\n",
       "      <td>std</td>\n",
       "      <td>hatchback</td>\n",
       "      <td>3.94</td>\n",
       "      <td>19</td>\n",
       "      <td>9.5</td>\n",
       "      <td>2778</td>\n",
       "      <td>rwd</td>\n",
       "      <td>front</td>\n",
       "      <td>151</td>\n",
       "      <td>ohc</td>\n",
       "      <td>mpfi</td>\n",
       "      <td>gas</td>\n",
       "      <td>50.2</td>\n",
       "      <td>27</td>\n",
       "      <td>143.0</td>\n",
       "      <td>168.9</td>\n",
       "      <td>16</td>\n",
       "      <td>186.0</td>\n",
       "      <td>four</td>\n",
       "      <td>two</td>\n",
       "      <td>5500.0</td>\n",
       "      <td>22018.0</td>\n",
       "      <td>3.11</td>\n",
       "      <td>3</td>\n",
       "      <td>94.5</td>\n",
       "      <td>68.3</td>\n",
       "      <td>0</td>\n",
       "      <td>Q3</td>\n",
       "      <td>(3.07, 3.15]</td>\n",
       "      <td>Q9</td>\n",
       "      <td>(140.8, 160.0]</td>\n",
       "      <td>Q2</td>\n",
       "      <td>(93.7, 94.5]</td>\n",
       "      <td>Q4</td>\n",
       "      <td>(9.28, 9.5]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>173</th>\n",
       "      <td>std</td>\n",
       "      <td>sedan</td>\n",
       "      <td>3.31</td>\n",
       "      <td>29</td>\n",
       "      <td>8.7</td>\n",
       "      <td>2326</td>\n",
       "      <td>fwd</td>\n",
       "      <td>front</td>\n",
       "      <td>122</td>\n",
       "      <td>ohc</td>\n",
       "      <td>mpfi</td>\n",
       "      <td>gas</td>\n",
       "      <td>54.9</td>\n",
       "      <td>34</td>\n",
       "      <td>92.0</td>\n",
       "      <td>175.6</td>\n",
       "      <td>20</td>\n",
       "      <td>65.0</td>\n",
       "      <td>four</td>\n",
       "      <td>four</td>\n",
       "      <td>4200.0</td>\n",
       "      <td>8948.0</td>\n",
       "      <td>3.54</td>\n",
       "      <td>-1</td>\n",
       "      <td>102.4</td>\n",
       "      <td>66.5</td>\n",
       "      <td>0</td>\n",
       "      <td>Q9</td>\n",
       "      <td>(3.46, 3.54]</td>\n",
       "      <td>Q5</td>\n",
       "      <td>(88.0, 95.0]</td>\n",
       "      <td>Q8</td>\n",
       "      <td>(100.4, 103.5]</td>\n",
       "      <td>Q2</td>\n",
       "      <td>(8.5, 9.0]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>92</th>\n",
       "      <td>std</td>\n",
       "      <td>sedan</td>\n",
       "      <td>3.15</td>\n",
       "      <td>31</td>\n",
       "      <td>9.4</td>\n",
       "      <td>1938</td>\n",
       "      <td>fwd</td>\n",
       "      <td>front</td>\n",
       "      <td>97</td>\n",
       "      <td>ohc</td>\n",
       "      <td>2bbl</td>\n",
       "      <td>gas</td>\n",
       "      <td>54.5</td>\n",
       "      <td>37</td>\n",
       "      <td>69.0</td>\n",
       "      <td>165.3</td>\n",
       "      <td>13</td>\n",
       "      <td>122.0</td>\n",
       "      <td>four</td>\n",
       "      <td>four</td>\n",
       "      <td>5200.0</td>\n",
       "      <td>6849.0</td>\n",
       "      <td>3.29</td>\n",
       "      <td>1</td>\n",
       "      <td>94.5</td>\n",
       "      <td>63.8</td>\n",
       "      <td>0</td>\n",
       "      <td>Q5</td>\n",
       "      <td>(3.23, 3.29]</td>\n",
       "      <td>Q2</td>\n",
       "      <td>(68.0, 69.0]</td>\n",
       "      <td>Q2</td>\n",
       "      <td>(93.7, 94.5]</td>\n",
       "      <td>Q4</td>\n",
       "      <td>(9.28, 9.5]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>std</td>\n",
       "      <td>sedan</td>\n",
       "      <td>3.62</td>\n",
       "      <td>16</td>\n",
       "      <td>8.0</td>\n",
       "      <td>3380</td>\n",
       "      <td>rwd</td>\n",
       "      <td>front</td>\n",
       "      <td>209</td>\n",
       "      <td>ohc</td>\n",
       "      <td>mpfi</td>\n",
       "      <td>gas</td>\n",
       "      <td>53.7</td>\n",
       "      <td>22</td>\n",
       "      <td>182.0</td>\n",
       "      <td>193.8</td>\n",
       "      <td>3</td>\n",
       "      <td>114.0</td>\n",
       "      <td>six</td>\n",
       "      <td>two</td>\n",
       "      <td>5400.0</td>\n",
       "      <td>41315.0</td>\n",
       "      <td>3.39</td>\n",
       "      <td>0</td>\n",
       "      <td>103.5</td>\n",
       "      <td>67.9</td>\n",
       "      <td>1</td>\n",
       "      <td>Q6</td>\n",
       "      <td>(3.29, 3.39]</td>\n",
       "      <td>Q10</td>\n",
       "      <td>(160.0, 262.0]</td>\n",
       "      <td>Q8</td>\n",
       "      <td>(100.4, 103.5]</td>\n",
       "      <td>Q1</td>\n",
       "      <td>(6.999, 8.5]</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    aspiration body_style  bore  city_mpg  compression_ratio  curb_weight  \\\n",
       "193        std      wagon  3.19        25                9.0         2563   \n",
       "125        std  hatchback  3.94        19                9.5         2778   \n",
       "173        std      sedan  3.31        29                8.7         2326   \n",
       "92         std      sedan  3.15        31                9.4         1938   \n",
       "16         std      sedan  3.62        16                8.0         3380   \n",
       "\n",
       "    drive_wheels engine_location  engine_size engine_type fuel_system  \\\n",
       "193          fwd           front          109         ohc        mpfi   \n",
       "125          rwd           front          151         ohc        mpfi   \n",
       "173          fwd           front          122         ohc        mpfi   \n",
       "92           fwd           front           97         ohc        2bbl   \n",
       "16           rwd           front          209         ohc        mpfi   \n",
       "\n",
       "    fuel_type  height  highway_mpg  horsepower  length  make_id  \\\n",
       "193       gas    55.1           31        88.0   183.1       21   \n",
       "125       gas    50.2           27       143.0   168.9       16   \n",
       "173       gas    54.9           34        92.0   175.6       20   \n",
       "92        gas    54.5           37        69.0   165.3       13   \n",
       "16        gas    53.7           22       182.0   193.8        3   \n",
       "\n",
       "     normalized_losses num_of_cylinders num_of_doors  peak_rpm    price  \\\n",
       "193              114.0             four         four    5500.0  12290.0   \n",
       "125              186.0             four          two    5500.0  22018.0   \n",
       "173               65.0             four         four    4200.0   8948.0   \n",
       "92               122.0             four         four    5200.0   6849.0   \n",
       "16               114.0              six          two    5400.0  41315.0   \n",
       "\n",
       "     stroke  symboling  wheel_base  width  normalized_losses_NA stroke_label  \\\n",
       "193    3.40          0       100.4   66.9                     1           Q7   \n",
       "125    3.11          3        94.5   68.3                     0           Q3   \n",
       "173    3.54         -1       102.4   66.5                     0           Q9   \n",
       "92     3.29          1        94.5   63.8                     0           Q5   \n",
       "16     3.39          0       103.5   67.9                     1           Q6   \n",
       "\n",
       "      stroke_disc horsepower_label horsepower_disc wheel_base_label  \\\n",
       "193   (3.39, 3.4]               Q4    (77.8, 88.0]               Q7   \n",
       "125  (3.07, 3.15]               Q9  (140.8, 160.0]               Q2   \n",
       "173  (3.46, 3.54]               Q5    (88.0, 95.0]               Q8   \n",
       "92   (3.23, 3.29]               Q2    (68.0, 69.0]               Q2   \n",
       "16   (3.29, 3.39]              Q10  (160.0, 262.0]               Q8   \n",
       "\n",
       "    wheel_base_disc compression_ratio_label compression_ratio_disc  \n",
       "193   (98.8, 100.4]                      Q2             (8.5, 9.0]  \n",
       "125    (93.7, 94.5]                      Q4            (9.28, 9.5]  \n",
       "173  (100.4, 103.5]                      Q2             (8.5, 9.0]  \n",
       "92     (93.7, 94.5]                      Q4            (9.28, 9.5]  \n",
       "16   (100.4, 103.5]                      Q1           (6.999, 8.5]  "
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# we use pandas cut method and pass the quantile edges calculated in the training set\n",
    "\n",
    "X_test['stroke_disc'] = pd.cut(x = X_test.stroke, bins=bins, labels=labels)\n",
    "X_test['horsepower_disc'] = pd.cut(x = X_test.horsepower, bins=bins)\n",
    "X_test['compression_ratio_disc'] = pd.cut(x = X_test.compression_ratio, bins=bins)\n",
    "X_test['wheel_base_disc'] = pd.cut(x = X_test.wheel_base, bins=bins)\n",
    "\n",
    "\n",
    "\n",
    "X_test['stroke_label'] = pd.cut(x = X_test.stroke, bins=bins, labels=labels)\n",
    "X_test['horsepower_label'] = pd.cut(x = X_test.horsepower, bins=bins, labels=labels)\n",
    "X_test['compression_ratio_label'] = pd.cut(x = X_test.compression_ratio, bins=bins, labels=labels)\n",
    "X_test['wheel_base_label'] = pd.cut(x = X_test.wheel_base, bins=bins, labels=labels)\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "X_train.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Check for NA before I continue"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "aspiration                  0\n",
       "body_style                  0\n",
       "bore                        0\n",
       "city_mpg                    0\n",
       "compression_ratio           0\n",
       "curb_weight                 0\n",
       "drive_wheels                0\n",
       "engine_location             0\n",
       "engine_size                 0\n",
       "engine_type                 0\n",
       "fuel_system                 0\n",
       "fuel_type                   0\n",
       "height                      0\n",
       "highway_mpg                 0\n",
       "horsepower                  0\n",
       "length                      0\n",
       "make_id                     0\n",
       "normalized_losses           0\n",
       "num_of_cylinders            0\n",
       "num_of_doors                0\n",
       "peak_rpm                    0\n",
       "price                       0\n",
       "stroke                      0\n",
       "symboling                   0\n",
       "wheel_base                  0\n",
       "width                       0\n",
       "normalized_losses_NA        0\n",
       "stroke_disc                31\n",
       "horsepower_disc            31\n",
       "compression_ratio_disc      0\n",
       "wheel_base_disc            31\n",
       "stroke_label               31\n",
       "horsepower_label           31\n",
       "compression_ratio_label     0\n",
       "wheel_base_label           31\n",
       "dtype: int64"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'Q3'"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Replace missing values with most common label\n",
    "# with this command we capture the most frequent label (check output with plot above)\n",
    "X_train.groupby(['stroke_label'])['stroke_label'].count().sort_values(ascending=False).index[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'Q1'"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Replace missing values with most common label\n",
    "# with this command we capture the most frequent label (check output with plot above)\n",
    "X_train.groupby(['horsepower_label'])['horsepower_label'].count().sort_values(ascending=False).index[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'Q2'"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Replace missing values with most common label\n",
    "# with this command we capture the most frequent label (check output with plot above)\n",
    "X_train.groupby(['compression_ratio_label'])['compression_ratio_label'].count().sort_values(ascending=False).index[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'Q1'"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Replace missing values with most common label\n",
    "# with this command we capture the most frequent label (check output with plot above)\n",
    "X_train.groupby(['wheel_base_label'])['wheel_base_label'].count().sort_values(ascending=False).index[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Interval(3.07, 3.15, closed='right')"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Replace missing values with most common label\n",
    "# with this command we capture the most frequent label (check output with plot above)\n",
    "X_train.groupby(['stroke_disc'])['stroke_disc'].count().sort_values(ascending=False).index[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Interval(51.999, 68.0, closed='right')"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Replace missing values with most common label\n",
    "# with this command we capture the most frequent label (check output with plot above)\n",
    "X_train.groupby(['horsepower_disc'])['horsepower_disc'].count().sort_values(ascending=False).index[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Interval(8.5, 9.0, closed='right')"
      ]
     },
     "execution_count": 60,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Replace missing values with most common label\n",
    "# with this command we capture the most frequent label (check output with plot above)\n",
    "X_train.groupby(['compression_ratio_disc'])['compression_ratio_disc'].count().sort_values(ascending=False).index[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Interval(86.59899999999999, 93.7, closed='right')"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Replace missing values with most common label\n",
    "# with this command we capture the most frequent label (check output with plot above)\n",
    "X_train.groupby(['wheel_base_disc'])['wheel_base_disc'].count().sort_values(ascending=False).index[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [],
   "source": [
    "# let's create a variable to replace NA with the most frequent label\n",
    "# both in train and test set\n",
    "\n",
    "def impute_na(df_train, df_test, variable):\n",
    "    most_frequent_category = df_train.groupby([variable])[variable].count().sort_values(ascending=False).index[0]\n",
    "    df_train[variable].fillna(most_frequent_category, inplace=True)\n",
    "    df_test[variable].fillna(most_frequent_category, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [],
   "source": [
    "# and let's replace the NA\n",
    "for variable in ['stroke_label','horsepower_label', \n",
    "                 'compression_ratio_label', 'wheel_base_label']:\n",
    "    impute_na(X_train, X_test, variable)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now I will drop the var_disc, as I will use the varibles_label. \n",
    "\n",
    "I shall also drop the original variables to prevent overfitting:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train = X_train.drop('stroke_disc', axis = 1)\n",
    "X_test = X_test.drop('stroke_disc', axis = 1)\n",
    "\n",
    "X_train = X_train.drop('horsepower_disc', axis = 1)\n",
    "X_test = X_test.drop('horsepower_disc', axis = 1)\n",
    "\n",
    "X_train = X_train.drop('compression_ratio_disc', axis = 1)\n",
    "X_test = X_test.drop('compression_ratio_disc', axis = 1)\n",
    "\n",
    "X_train = X_train.drop('wheel_base_disc', axis = 1)\n",
    "X_test = X_test.drop('wheel_base_disc', axis = 1)\n",
    "\n",
    "######\n",
    "X_train = X_train.drop('stroke', axis = 1)\n",
    "X_test = X_test.drop('stroke', axis = 1)\n",
    "\n",
    "X_train = X_train.drop('horsepower', axis = 1)\n",
    "X_test = X_test.drop('horsepower', axis = 1)\n",
    "\n",
    "X_train = X_train.drop('compression_ratio', axis = 1)\n",
    "X_test = X_test.drop('compression_ratio', axis = 1)\n",
    "\n",
    "X_train = X_train.drop('wheel_base', axis = 1)\n",
    "X_test = X_test.drop('wheel_base', axis = 1)\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "aspiration                 0\n",
       "body_style                 0\n",
       "bore                       0\n",
       "city_mpg                   0\n",
       "curb_weight                0\n",
       "drive_wheels               0\n",
       "engine_location            0\n",
       "engine_size                0\n",
       "engine_type                0\n",
       "fuel_system                0\n",
       "fuel_type                  0\n",
       "height                     0\n",
       "highway_mpg                0\n",
       "length                     0\n",
       "make_id                    0\n",
       "normalized_losses          0\n",
       "num_of_cylinders           0\n",
       "num_of_doors               0\n",
       "peak_rpm                   0\n",
       "price                      0\n",
       "symboling                  0\n",
       "width                      0\n",
       "normalized_losses_NA       0\n",
       "stroke_label               0\n",
       "horsepower_label           0\n",
       "wheel_base_label           0\n",
       "compression_ratio_label    0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "aspiration                 0\n",
       "body_style                 0\n",
       "bore                       0\n",
       "city_mpg                   0\n",
       "curb_weight                0\n",
       "drive_wheels               0\n",
       "engine_location            0\n",
       "engine_size                0\n",
       "engine_type                0\n",
       "fuel_system                0\n",
       "fuel_type                  0\n",
       "height                     0\n",
       "highway_mpg                0\n",
       "length                     0\n",
       "make_id                    0\n",
       "normalized_losses          0\n",
       "num_of_cylinders           0\n",
       "num_of_doors               0\n",
       "peak_rpm                   0\n",
       "price                      0\n",
       "symboling                  0\n",
       "width                      0\n",
       "normalized_losses_NA       0\n",
       "stroke_label               0\n",
       "horsepower_label           0\n",
       "compression_ratio_label    0\n",
       "wheel_base_label           0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 66,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test.isnull().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 8.2.1 Combine discretisation with mean encoding (Risk factor)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [],
   "source": [
    "ordered_labels1 = X_train.groupby(['stroke_label'])['price'].mean().to_dict()\n",
    "ordered_labels2 = X_train.groupby(['horsepower_label'])['price'].mean().to_dict()\n",
    "ordered_labels3 = X_train.groupby(['compression_ratio_label'])['price'].mean().to_dict()\n",
    "ordered_labels4 = X_train.groupby(['wheel_base_label'])['price'].mean().to_dict()\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [],
   "source": [
    "# replace the labels with the 'risk' (target frequency)\n",
    "# note that we calculated the frequencies based on the training set only\n",
    "\n",
    "X_train['stroke_ordered'] = X_train.stroke_label.map(ordered_labels1)\n",
    "X_test['stroke_ordered'] = X_test.stroke_label.map(ordered_labels1)\n",
    "\n",
    "X_train['horsepower_ordered'] = X_train.horsepower_label.map(ordered_labels2)\n",
    "X_test['horsepower_ordered'] = X_test.horsepower_label.map(ordered_labels2)\n",
    "\n",
    "\n",
    "X_train['compression_ratio_ordered'] = X_train.compression_ratio_label.map(ordered_labels3)\n",
    "X_test['compression_ratio_ordered'] = X_test.compression_ratio_label.map(ordered_labels3)\n",
    "\n",
    "X_train['wheel_base_ordered'] = X_train.wheel_base_label.map(ordered_labels4)\n",
    "X_test['wheel_base_ordered'] = X_test.wheel_base_label.map(ordered_labels4)\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To prevent overfitting, now I can delete those columns used to created the var_label"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train = X_train.drop('stroke_label', axis = 1)\n",
    "X_test = X_test.drop('stroke_label', axis = 1)\n",
    "\n",
    "X_train = X_train.drop('horsepower_label', axis = 1)\n",
    "X_test = X_test.drop('horsepower_label', axis = 1)\n",
    "\n",
    "X_train = X_train.drop('compression_ratio_label', axis = 1)\n",
    "X_test = X_test.drop('compression_ratio_label', axis = 1)\n",
    "\n",
    "X_train = X_train.drop('wheel_base_label', axis = 1)\n",
    "X_test = X_test.drop('wheel_base_label', axis = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "aspiration                   0\n",
       "body_style                   0\n",
       "bore                         0\n",
       "city_mpg                     0\n",
       "curb_weight                  0\n",
       "drive_wheels                 0\n",
       "engine_location              0\n",
       "engine_size                  0\n",
       "engine_type                  0\n",
       "fuel_system                  0\n",
       "fuel_type                    0\n",
       "height                       0\n",
       "highway_mpg                  0\n",
       "length                       0\n",
       "make_id                      0\n",
       "normalized_losses            0\n",
       "num_of_cylinders             0\n",
       "num_of_doors                 0\n",
       "peak_rpm                     0\n",
       "price                        0\n",
       "symboling                    0\n",
       "width                        0\n",
       "normalized_losses_NA         0\n",
       "stroke_ordered               0\n",
       "horsepower_ordered           0\n",
       "compression_ratio_ordered    0\n",
       "wheel_base_ordered           0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 70,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.isnull().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**8.3 Outlier handling with Top/Bottom caping**\n",
    "\n",
    "For variables with skew values less than absolute 1"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    " - bore (top)\n",
    " - height (top)\n",
    " - highway-mpg (top)\n",
    " - length (bottom, top)\n",
    " - normalized-losses(top)\n",
    " - peak-rpm(bottom)\n",
    " - width(bottom,top)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [],
   "source": [
    "outliertopbot = ['bore',\n",
    "'height',\n",
    "'highway_mpg',\n",
    "'length',\n",
    "'peak_rpm',\n",
    "'normalized_losses',\n",
    "'width']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['bore'] 2.2199999999999998 & 4.390000000000001\n",
      "['height'] 44.5 & 62.0\n",
      "['highway_mpg'] -3.5 & 66.5\n",
      "['length'] 132.8 & 209.8\n",
      "['peak_rpm'] 2700.0 & 7600.0\n",
      "['normalized_losses'] 18.0 & 221.0\n",
      "['width'] 57.700000000000045 & 73.09999999999997\n"
     ]
    }
   ],
   "source": [
    "for var in outliertopbot:\n",
    "    IQR = df[var].quantile(0.75) - df[var].quantile(0.25)\n",
    "    Lower_fence = df[var].quantile(0.25) - (IQR * 3)\n",
    "    Upper_fence = df[var].quantile(0.75) + (IQR * 3)\n",
    "    print([var],Lower_fence,'&', Upper_fence)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [],
   "source": [
    "def bottom_code(df, variable, bottom):\n",
    "    return np.where(df[variable]<=bottom, bottom, df[variable])\n",
    "\n",
    "for df in [X_train, X_test]:\n",
    "    df['length'] = bottom_code(df, 'length', 132.8)\n",
    "    df['peak_rpm'] = bottom_code(df, 'peak_rpm', 2700)\n",
    "    df['width'] = bottom_code(df, 'width', 57.70)    \n",
    "\n",
    "def top_code(df, variable, top):\n",
    "    return np.where(df[variable]>=top, top, df[variable])\n",
    "\n",
    "for df in [X_train, X_test]:\n",
    "    df['bore'] = top_code(df, 'bore', 4.39)\n",
    "    df['height'] = top_code(df, 'height', 62)\n",
    "    df['highway_mpg'] = top_code(df, 'highway_mpg', 66.5)\n",
    "    df['length'] = top_code(df, 'length', 209.8)\n",
    "    df['normalized_losses'] = top_code(df, 'normalized_losses', 221)\n",
    "    df['widht'] = top_code(df, 'width', 73.09)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    31.000000\n",
       "mean     65.700000\n",
       "std       2.344071\n",
       "min      60.300000\n",
       "25%      64.300000\n",
       "50%      65.500000\n",
       "75%      66.500000\n",
       "max      72.300000\n",
       "Name: width, dtype: float64"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['width'].describe()\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 9. Engineer rare lables in categorical variables/discrete"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [],
   "source": [
    "# the following vars in the data set are encoded the wrong way:\n",
    "\n",
    "X_train['symboling'] = X_train['symboling'].astype('category')\n",
    "X_test['symboling'] = X_test['symboling'].astype('category')\n",
    "\n",
    "X_train['make_id'] = X_train['make_id'].astype('category')\n",
    "X_test['make_id'] = X_test['make_id'].astype('category')    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [],
   "source": [
    "def rare_imputation(variable):\n",
    "    # find frequent labels / discrete numbers\n",
    "    temp = X_train.groupby([variable])[variable].count()/np.float(len(X_train))\n",
    "    frequent_cat = [x for x in temp.loc[temp>0.03].index.values]\n",
    "    \n",
    "    X_train[variable] = np.where(X_train[variable].isin(frequent_cat), X_train[variable], 'Rare')\n",
    "    X_test[variable] = np.where(X_test[variable].isin(frequent_cat), X_test[variable], 'Rare')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['aspiration',\n",
       " 'body_style',\n",
       " 'drive_wheels',\n",
       " 'engine_location',\n",
       " 'engine_type',\n",
       " 'fuel_system',\n",
       " 'fuel_type',\n",
       " 'make',\n",
       " 'num_of_cylinders',\n",
       " 'num_of_doors']"
      ]
     },
     "execution_count": 77,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "categorical"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [],
   "source": [
    "categorical = ['aspiration',\n",
    " 'body_style',\n",
    " 'drive_wheels',\n",
    " 'engine_location',\n",
    " 'engine_type',\n",
    " 'fuel_system',\n",
    " 'fuel_type',\n",
    " 'make',\n",
    " 'num_of_cylinders',\n",
    " 'num_of_doors', 'symboling','make_id']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [],
   "source": [
    "# remove make because it is no longer in our dataset\n",
    "categorical.remove('make')\n",
    "\n",
    "# find infrequent labels in categorical variables and replace by Rare\n",
    "for var in categorical:\n",
    "    rare_imputation(var)\n",
    "    \n",
    "# find infrequent labels in categorical variables and replace by Rare\n",
    "# remember that we are treating discrete variables as if they were categorical\n",
    "for var in discrete:\n",
    "    rare_imputation(var)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "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": [
    "# let's check that it worked\n",
    "for var in categorical:\n",
    "    (X_train.groupby(var)[var].count() / np.float(len(X_train))).plot.bar()\n",
    "    plt.ylabel('Percentage of observations per label')\n",
    "    plt.title(var)\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "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": [
    "# let's check that it worked\n",
    "for var in discrete:\n",
    "    (X_train.groupby(var)[var].count() / np.float(len(X_train))).plot.bar()\n",
    "    plt.ylabel('Percentage of observations per label')\n",
    "    plt.title(var)\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "YES! It did work well!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 10. Encoding Categorical Variables\n",
    "\n",
    "I will use target encoding\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [],
   "source": [
    "def encode_categorical_variables(var, target):\n",
    "        # make label to price dictionary\n",
    "        ordered_labels = X_train.groupby([var])[target].mean().to_dict()\n",
    "        \n",
    "        # encode variables\n",
    "        X_train[var] = X_train[var].map(ordered_labels)\n",
    "        X_test[var] = X_test[var].map(ordered_labels)\n",
    "\n",
    "# encode labels in categorical vars\n",
    "for var in categorical:\n",
    "    encode_categorical_variables(var, 'price')\n",
    "    \n",
    "# encode labels in discrete vars\n",
    "for var in discrete:\n",
    "    encode_categorical_variables(var, 'price')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>aspiration</th>\n",
       "      <th>body_style</th>\n",
       "      <th>bore</th>\n",
       "      <th>city_mpg</th>\n",
       "      <th>curb_weight</th>\n",
       "      <th>drive_wheels</th>\n",
       "      <th>engine_location</th>\n",
       "      <th>engine_size</th>\n",
       "      <th>engine_type</th>\n",
       "      <th>fuel_system</th>\n",
       "      <th>fuel_type</th>\n",
       "      <th>height</th>\n",
       "      <th>highway_mpg</th>\n",
       "      <th>length</th>\n",
       "      <th>make_id</th>\n",
       "      <th>normalized_losses</th>\n",
       "      <th>num_of_cylinders</th>\n",
       "      <th>num_of_doors</th>\n",
       "      <th>peak_rpm</th>\n",
       "      <th>price</th>\n",
       "      <th>symboling</th>\n",
       "      <th>width</th>\n",
       "      <th>normalized_losses_NA</th>\n",
       "      <th>stroke_ordered</th>\n",
       "      <th>horsepower_ordered</th>\n",
       "      <th>compression_ratio_ordered</th>\n",
       "      <th>wheel_base_ordered</th>\n",
       "      <th>widht</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>193</th>\n",
       "      <td>12844.206522</td>\n",
       "      <td>12037.050000</td>\n",
       "      <td>3.19</td>\n",
       "      <td>25</td>\n",
       "      <td>2563</td>\n",
       "      <td>9373.348039</td>\n",
       "      <td>13185.830409</td>\n",
       "      <td>109</td>\n",
       "      <td>11773.261905</td>\n",
       "      <td>18257.185897</td>\n",
       "      <td>13289.948387</td>\n",
       "      <td>55.1</td>\n",
       "      <td>31.0</td>\n",
       "      <td>183.1</td>\n",
       "      <td>10296.000000</td>\n",
       "      <td>114.0</td>\n",
       "      <td>10444.281022</td>\n",
       "      <td>13634.740385</td>\n",
       "      <td>5500.0</td>\n",
       "      <td>12290.0</td>\n",
       "      <td>14640.078947</td>\n",
       "      <td>66.9</td>\n",
       "      <td>1</td>\n",
       "      <td>13066.343750</td>\n",
       "      <td>8843.363636</td>\n",
       "      <td>11295.254902</td>\n",
       "      <td>14538.28125</td>\n",
       "      <td>66.9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>125</th>\n",
       "      <td>12844.206522</td>\n",
       "      <td>10194.361111</td>\n",
       "      <td>3.94</td>\n",
       "      <td>19</td>\n",
       "      <td>2778</td>\n",
       "      <td>20587.546875</td>\n",
       "      <td>13185.830409</td>\n",
       "      <td>151</td>\n",
       "      <td>11773.261905</td>\n",
       "      <td>18257.185897</td>\n",
       "      <td>13289.948387</td>\n",
       "      <td>50.2</td>\n",
       "      <td>27.0</td>\n",
       "      <td>168.9</td>\n",
       "      <td>19875.285714</td>\n",
       "      <td>186.0</td>\n",
       "      <td>10444.281022</td>\n",
       "      <td>13433.542857</td>\n",
       "      <td>5500.0</td>\n",
       "      <td>22018.0</td>\n",
       "      <td>17864.086957</td>\n",
       "      <td>68.3</td>\n",
       "      <td>0</td>\n",
       "      <td>15010.613636</td>\n",
       "      <td>17661.605263</td>\n",
       "      <td>12200.340909</td>\n",
       "      <td>9203.42500</td>\n",
       "      <td>68.3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>173</th>\n",
       "      <td>12844.206522</td>\n",
       "      <td>14589.798851</td>\n",
       "      <td>3.31</td>\n",
       "      <td>29</td>\n",
       "      <td>2326</td>\n",
       "      <td>9373.348039</td>\n",
       "      <td>13185.830409</td>\n",
       "      <td>122</td>\n",
       "      <td>11773.261905</td>\n",
       "      <td>18257.185897</td>\n",
       "      <td>13289.948387</td>\n",
       "      <td>54.9</td>\n",
       "      <td>34.0</td>\n",
       "      <td>175.6</td>\n",
       "      <td>9659.956522</td>\n",
       "      <td>65.0</td>\n",
       "      <td>10444.281022</td>\n",
       "      <td>13634.740385</td>\n",
       "      <td>4200.0</td>\n",
       "      <td>8948.0</td>\n",
       "      <td>17405.952381</td>\n",
       "      <td>66.5</td>\n",
       "      <td>0</td>\n",
       "      <td>12370.062500</td>\n",
       "      <td>12346.285714</td>\n",
       "      <td>11295.254902</td>\n",
       "      <td>18111.43750</td>\n",
       "      <td>66.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>92</th>\n",
       "      <td>12844.206522</td>\n",
       "      <td>14589.798851</td>\n",
       "      <td>3.15</td>\n",
       "      <td>31</td>\n",
       "      <td>1938</td>\n",
       "      <td>9373.348039</td>\n",
       "      <td>13185.830409</td>\n",
       "      <td>97</td>\n",
       "      <td>11773.261905</td>\n",
       "      <td>7656.026786</td>\n",
       "      <td>13289.948387</td>\n",
       "      <td>54.5</td>\n",
       "      <td>37.0</td>\n",
       "      <td>165.3</td>\n",
       "      <td>10299.000000</td>\n",
       "      <td>122.0</td>\n",
       "      <td>10444.281022</td>\n",
       "      <td>13634.740385</td>\n",
       "      <td>5200.0</td>\n",
       "      <td>6849.0</td>\n",
       "      <td>9753.031250</td>\n",
       "      <td>63.8</td>\n",
       "      <td>0</td>\n",
       "      <td>11270.125000</td>\n",
       "      <td>7461.500000</td>\n",
       "      <td>12200.340909</td>\n",
       "      <td>9203.42500</td>\n",
       "      <td>63.8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>12844.206522</td>\n",
       "      <td>14589.798851</td>\n",
       "      <td>3.62</td>\n",
       "      <td>16</td>\n",
       "      <td>3380</td>\n",
       "      <td>20587.546875</td>\n",
       "      <td>13185.830409</td>\n",
       "      <td>209</td>\n",
       "      <td>11773.261905</td>\n",
       "      <td>18257.185897</td>\n",
       "      <td>13289.948387</td>\n",
       "      <td>53.7</td>\n",
       "      <td>22.0</td>\n",
       "      <td>193.8</td>\n",
       "      <td>26854.285714</td>\n",
       "      <td>114.0</td>\n",
       "      <td>24910.350000</td>\n",
       "      <td>13433.542857</td>\n",
       "      <td>5400.0</td>\n",
       "      <td>41315.0</td>\n",
       "      <td>14640.078947</td>\n",
       "      <td>67.9</td>\n",
       "      <td>1</td>\n",
       "      <td>17316.400000</td>\n",
       "      <td>30801.812500</td>\n",
       "      <td>17305.815217</td>\n",
       "      <td>18111.43750</td>\n",
       "      <td>67.9</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       aspiration    body_style  bore  city_mpg  curb_weight  drive_wheels  \\\n",
       "193  12844.206522  12037.050000  3.19        25         2563   9373.348039   \n",
       "125  12844.206522  10194.361111  3.94        19         2778  20587.546875   \n",
       "173  12844.206522  14589.798851  3.31        29         2326   9373.348039   \n",
       "92   12844.206522  14589.798851  3.15        31         1938   9373.348039   \n",
       "16   12844.206522  14589.798851  3.62        16         3380  20587.546875   \n",
       "\n",
       "     engine_location  engine_size   engine_type   fuel_system     fuel_type  \\\n",
       "193     13185.830409          109  11773.261905  18257.185897  13289.948387   \n",
       "125     13185.830409          151  11773.261905  18257.185897  13289.948387   \n",
       "173     13185.830409          122  11773.261905  18257.185897  13289.948387   \n",
       "92      13185.830409           97  11773.261905   7656.026786  13289.948387   \n",
       "16      13185.830409          209  11773.261905  18257.185897  13289.948387   \n",
       "\n",
       "     height  highway_mpg  length       make_id  normalized_losses  \\\n",
       "193    55.1         31.0   183.1  10296.000000              114.0   \n",
       "125    50.2         27.0   168.9  19875.285714              186.0   \n",
       "173    54.9         34.0   175.6   9659.956522               65.0   \n",
       "92     54.5         37.0   165.3  10299.000000              122.0   \n",
       "16     53.7         22.0   193.8  26854.285714              114.0   \n",
       "\n",
       "     num_of_cylinders  num_of_doors  peak_rpm    price     symboling  width  \\\n",
       "193      10444.281022  13634.740385    5500.0  12290.0  14640.078947   66.9   \n",
       "125      10444.281022  13433.542857    5500.0  22018.0  17864.086957   68.3   \n",
       "173      10444.281022  13634.740385    4200.0   8948.0  17405.952381   66.5   \n",
       "92       10444.281022  13634.740385    5200.0   6849.0   9753.031250   63.8   \n",
       "16       24910.350000  13433.542857    5400.0  41315.0  14640.078947   67.9   \n",
       "\n",
       "     normalized_losses_NA  stroke_ordered  horsepower_ordered  \\\n",
       "193                     1    13066.343750         8843.363636   \n",
       "125                     0    15010.613636        17661.605263   \n",
       "173                     0    12370.062500        12346.285714   \n",
       "92                      0    11270.125000         7461.500000   \n",
       "16                      1    17316.400000        30801.812500   \n",
       "\n",
       "     compression_ratio_ordered  wheel_base_ordered  widht  \n",
       "193               11295.254902         14538.28125   66.9  \n",
       "125               12200.340909          9203.42500   68.3  \n",
       "173               11295.254902         18111.43750   66.5  \n",
       "92                12200.340909          9203.42500   63.8  \n",
       "16                17305.815217         18111.43750   67.9  "
      ]
     },
     "execution_count": 83,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#let's inspect the dataset\n",
    "X_train.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 11.Feature Scaling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "aspiration                   float64\n",
       "body_style                   float64\n",
       "bore                         float64\n",
       "city_mpg                       int64\n",
       "curb_weight                    int64\n",
       "drive_wheels                 float64\n",
       "engine_location              float64\n",
       "engine_size                    int64\n",
       "engine_type                  float64\n",
       "fuel_system                  float64\n",
       "fuel_type                    float64\n",
       "height                       float64\n",
       "highway_mpg                  float64\n",
       "length                       float64\n",
       "make_id                      float64\n",
       "normalized_losses            float64\n",
       "num_of_cylinders             float64\n",
       "num_of_doors                 float64\n",
       "peak_rpm                     float64\n",
       "price                        float64\n",
       "symboling                    float64\n",
       "width                        float64\n",
       "normalized_losses_NA           int64\n",
       "stroke_ordered               float64\n",
       "horsepower_ordered           float64\n",
       "compression_ratio_ordered    float64\n",
       "wheel_base_ordered           float64\n",
       "widht                        float64\n",
       "dtype: object"
      ]
     },
     "execution_count": 84,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Check before it is all numerical\n",
    "X_train.dtypes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "aspiration                   float64\n",
       "body_style                   float64\n",
       "bore                         float64\n",
       "city_mpg                       int64\n",
       "curb_weight                    int64\n",
       "drive_wheels                 float64\n",
       "engine_location              float64\n",
       "engine_size                    int64\n",
       "engine_type                  float64\n",
       "fuel_system                  float64\n",
       "fuel_type                    float64\n",
       "height                       float64\n",
       "highway_mpg                  float64\n",
       "length                       float64\n",
       "make_id                      float64\n",
       "normalized_losses            float64\n",
       "num_of_cylinders             float64\n",
       "num_of_doors                 float64\n",
       "peak_rpm                     float64\n",
       "price                        float64\n",
       "symboling                    float64\n",
       "width                        float64\n",
       "normalized_losses_NA           int64\n",
       "stroke_ordered               float64\n",
       "horsepower_ordered           float64\n",
       "compression_ratio_ordered    float64\n",
       "wheel_base_ordered           float64\n",
       "widht                        float64\n",
       "dtype: object"
      ]
     },
     "execution_count": 85,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Check before it is all numerical\n",
    "X_test.dtypes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['aspiration', 'body_style', 'bore', 'city_mpg', 'curb_weight',\n",
       "       'drive_wheels', 'engine_location', 'engine_size', 'engine_type',\n",
       "       'fuel_system', 'fuel_type', 'height', 'highway_mpg', 'length',\n",
       "       'make_id', 'normalized_losses', 'num_of_cylinders', 'num_of_doors',\n",
       "       'peak_rpm', 'price', 'symboling', 'width', 'normalized_losses_NA',\n",
       "       'stroke_ordered', 'horsepower_ordered', 'compression_ratio_ordered',\n",
       "       'wheel_base_ordered', 'widht'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 86,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Find column names\n",
    "X_train.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "total number of variables to use for training:  28\n"
     ]
    }
   ],
   "source": [
    "# let's create a list of the training variables\n",
    "training_vars = ['aspiration', 'body_style', 'bore', 'city_mpg', 'curb_weight',\n",
    "       'drive_wheels', 'engine_location', 'engine_size', 'engine_type',\n",
    "       'fuel_system', 'fuel_type', 'height', 'highway_mpg', 'length',\n",
    "       'make_id', 'normalized_losses', 'num_of_cylinders', 'num_of_doors',\n",
    "       'peak_rpm', 'price', 'symboling', 'width', 'normalized_losses_NA',\n",
    "       'stroke_ordered', 'horsepower_ordered', 'compression_ratio_ordered',\n",
    "       'wheel_base_ordered', 'widht']\n",
    "\n",
    "print('total number of variables to use for training: ', len(training_vars))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "aspiration skewness is 5.5677643628300215\n",
      "body_style skewness is 2.2846623412884255\n",
      "bore skewness is 0.527700822150885\n",
      "city_mpg skewness is 1.0403231549722474\n",
      "curb_weight skewness is 0.2120018148039922\n",
      "drive_wheels skewness is 0.4829064772716164\n",
      "engine_location skewness is 0\n",
      "engine_size skewness is 0.5635017615197339\n",
      "engine_type skewness is 2.868683399748382\n",
      "fuel_system skewness is -0.27588835177803833\n",
      "fuel_type skewness is 5.567764362830074\n",
      "height skewness is 0.27350172478135837\n",
      "highway_mpg skewness is 0.954946036021284\n",
      "length skewness is -0.16070860345381996\n",
      "make_id skewness is 1.1874675504165326\n",
      "normalized_losses skewness is 0.8716855672848831\n",
      "num_of_cylinders skewness is 1.1198901422468581\n",
      "num_of_doors skewness is 0.4874994082774718\n",
      "peak_rpm skewness is -0.15997438759248278\n",
      "price skewness is 0.6255002331118181\n",
      "symboling skewness is 0.4544326061084143\n",
      "width skewness is 0.8650397680572235\n",
      "normalized_losses_NA skewness is 0.9715255657145386\n",
      "stroke_ordered skewness is 0\n",
      "horsepower_ordered skewness is 0\n",
      "compression_ratio_ordered skewness is 0.9318789035528492\n",
      "wheel_base_ordered skewness is 0\n",
      "widht skewness is 0.8650397680572235\n"
     ]
    }
   ],
   "source": [
    "# let's find the skewness of above variables\n",
    "for var in training_vars:\n",
    "    print(var, 'skewness is', df[var].skew() )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "OK, I shall do :\n",
    "   - Robust Scaling for variables with an absolute skew above 2\n",
    "   - MinMax Scaling for variables with an absoulute skew between 1 and 2\n",
    "   - Standard for the rest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "aspiration\n",
      "body_style\n",
      "engine_type\n",
      "fuel_type\n"
     ]
    }
   ],
   "source": [
    "for var in training_vars:\n",
    "    max_abs_value = 2\n",
    "    if abs(df[var].skew()) >= max_abs_value:\n",
    "        print (var)\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {},
   "outputs": [],
   "source": [
    "# I Will start with Robust Scaling\n",
    "scalerR = RobustScaler() # call the object\n",
    "X_train_scaledR = scalerR.fit_transform(X_train[['aspiration', 'body_style', 'engine_type', 'fuel_type']]) # fit the scaler to the train set, and then scale it\n",
    "X_test_scaledR= scalerR.transform(X_test[['aspiration', 'body_style', 'engine_type', 'fuel_type']]) # scale the test set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "city_mpg\n",
      "make_id\n",
      "num_of_cylinders\n"
     ]
    }
   ],
   "source": [
    "for var in training_vars:\n",
    "    max_abs_value = 2 \n",
    "    if (1 <= abs(df[var].skew()) <=2):  \n",
    "        print (var)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Continue with MinMax Scaling\n",
    "scalerMM = MinMaxScaler() # create an instance\n",
    "\n",
    "X_train_scaledMM= scalerMM.fit_transform(X_train[['num_of_cylinders','make_id','city_mpg']]) #  fit  the scaler to the train set and then transform it\n",
    "X_test_scaledMM= scalerMM.transform(X_test[['num_of_cylinders','make_id','city_mpg']]) # transform (scale) the test set\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "bore\n",
      "curb_weight\n",
      "drive_wheels\n",
      "engine_location\n",
      "engine_size\n",
      "fuel_system\n",
      "height\n",
      "highway_mpg\n",
      "length\n",
      "normalized_losses\n",
      "num_of_doors\n",
      "peak_rpm\n",
      "price\n",
      "symboling\n",
      "width\n",
      "normalized_losses_NA\n",
      "stroke_ordered\n",
      "horsepower_ordered\n",
      "compression_ratio_ordered\n",
      "wheel_base_ordered\n",
      "widht\n"
     ]
    }
   ],
   "source": [
    "for var in training_vars:\n",
    "    max_abs_value = 2 \n",
    "    if ((df[var].skew()) <=1):  \n",
    "        print (var)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Normal Standarization\n",
    "scalerN = StandardScaler() # create an object\n",
    "X_train_scaledN = scalerN.fit_transform(X_train[['bore',\n",
    "'curb_weight',\n",
    "'drive_wheels',\n",
    "'engine_location',\n",
    "'engine_size',\n",
    "'fuel_system',\n",
    "'height',\n",
    "'highway_mpg',\n",
    "'length',\n",
    "'normalized_losses',\n",
    "'num_of_doors',\n",
    "'peak_rpm',\n",
    "'symboling',\n",
    "'width',\n",
    "'normalized_losses_NA',\n",
    "'stroke_ordered',\n",
    "'horsepower_ordered',\n",
    "'compression_ratio_ordered',\n",
    "'wheel_base_ordered',\n",
    "'widht' ]]) # fit the scaler to the train set, and then transform it\n",
    "X_test_scaledN = scalerN.transform(X_test[['bore',\n",
    "'curb_weight',\n",
    "'drive_wheels',\n",
    "'engine_location',\n",
    "'engine_size',\n",
    "'fuel_system',\n",
    "'height',\n",
    "'highway_mpg',\n",
    "'length',\n",
    "'normalized_losses',\n",
    "'num_of_doors',\n",
    "'peak_rpm',\n",
    "'symboling',\n",
    "'width',\n",
    "'normalized_losses_NA',\n",
    "'stroke_ordered',\n",
    "'horsepower_ordered',\n",
    "'compression_ratio_ordered',\n",
    "'wheel_base_ordered',\n",
    "'widht']]) # transform the test set"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 11. Sanity check (nulls)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train = pd.DataFrame(X_train)\n",
    "X_test = pd.DataFrame(X_test)\n",
    "y_train = pd.DataFrame(y_train)\n",
    "y_test = pd.DataFrame(y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "aspiration                   0\n",
       "body_style                   0\n",
       "bore                         0\n",
       "city_mpg                     0\n",
       "curb_weight                  0\n",
       "drive_wheels                 0\n",
       "engine_location              0\n",
       "engine_size                  0\n",
       "engine_type                  0\n",
       "fuel_system                  0\n",
       "fuel_type                    0\n",
       "height                       0\n",
       "highway_mpg                  0\n",
       "length                       0\n",
       "make_id                      0\n",
       "normalized_losses            0\n",
       "num_of_cylinders             0\n",
       "num_of_doors                 0\n",
       "peak_rpm                     0\n",
       "price                        0\n",
       "symboling                    0\n",
       "width                        0\n",
       "normalized_losses_NA         0\n",
       "stroke_ordered               0\n",
       "horsepower_ordered           0\n",
       "compression_ratio_ordered    0\n",
       "wheel_base_ordered           0\n",
       "widht                        0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 96,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "aspiration                   0\n",
       "body_style                   0\n",
       "bore                         0\n",
       "city_mpg                     0\n",
       "curb_weight                  0\n",
       "drive_wheels                 0\n",
       "engine_location              0\n",
       "engine_size                  0\n",
       "engine_type                  0\n",
       "fuel_system                  0\n",
       "fuel_type                    0\n",
       "height                       0\n",
       "highway_mpg                  0\n",
       "length                       0\n",
       "make_id                      0\n",
       "normalized_losses            0\n",
       "num_of_cylinders             0\n",
       "num_of_doors                 0\n",
       "peak_rpm                     0\n",
       "price                        0\n",
       "symboling                    0\n",
       "width                        0\n",
       "normalized_losses_NA         0\n",
       "stroke_ordered               0\n",
       "horsepower_ordered           0\n",
       "compression_ratio_ordered    0\n",
       "wheel_base_ordered           0\n",
       "widht                        0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 97,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "price    2\n",
       "dtype: int64"
      ]
     },
     "execution_count": 98,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "price    2\n",
       "dtype: int64"
      ]
     },
     "execution_count": 99,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_test.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count      172.000000\n",
       "mean     13587.604651\n",
       "std       8320.024245\n",
       "min       5195.000000\n",
       "25%       7784.750000\n",
       "50%      10646.500000\n",
       "75%      16525.750000\n",
       "max      45400.000000\n",
       "Name: price, dtype: float64"
      ]
     },
     "execution_count": 100,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train['price'].describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_train['price'].fillna(10646.5, inplace = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 102,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train['price'].isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count       29.000000\n",
       "mean     10950.517241\n",
       "std       4711.299795\n",
       "min       5118.000000\n",
       "25%       6938.000000\n",
       "50%       9959.000000\n",
       "75%      15580.000000\n",
       "max      20970.000000\n",
       "Name: price, dtype: float64"
      ]
     },
     "execution_count": 103,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_test['price'].describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_test['price'].fillna(9959, inplace = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 105,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train['price'].isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 106,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_test['price'].isnull().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 12. Save selected variables for features predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_scaledR = pd.DataFrame(X_train_scaledR, columns = ['aspiration', 'body_style', 'engine_type', 'fuel_type'])\n",
    "X_train_scaledMM = pd.DataFrame(X_train_scaledMM, columns = ['num_of_cylinders','make_id','city_mpg'])\n",
    "X_train_scaledN = pd.DataFrame(X_train_scaledN, columns = ['bore',\n",
    "'curb_weight',\n",
    "'drive_wheels',\n",
    "'engine_location',\n",
    "'engine_size',\n",
    "'fuel_system',\n",
    "'height',\n",
    "'highway_mpg',\n",
    "'length',\n",
    "'normalized_losses',\n",
    "'num_of_doors',\n",
    "'peak_rpm',\n",
    "'symboling',\n",
    "'width',\n",
    "'normalized_losses_NA',\n",
    "'stroke_ordered',\n",
    "'horsepower_ordered',\n",
    "'compression_ratio_ordered',\n",
    "'wheel_base_ordered',\n",
    "'widht'])\n",
    "\n",
    "######\n",
    "X_test_scaledR = pd.DataFrame(X_test_scaledR, columns = ['aspiration', 'body_style', 'engine_type', 'fuel_type'])\n",
    "X_test_scaledMM = pd.DataFrame(X_test_scaledMM, columns = ['num_of_cylinders','make_id','city_mpg'])\n",
    "X_test_scaledN = pd.DataFrame(X_test_scaledN, columns = ['bore',\n",
    "'curb_weight',\n",
    "'drive_wheels',\n",
    "'engine_location',\n",
    "'engine_size',\n",
    "'fuel_system',\n",
    "'height',\n",
    "'highway_mpg',\n",
    "'length',\n",
    "'normalized_losses',\n",
    "'num_of_doors',\n",
    "'peak_rpm',\n",
    "'symboling',\n",
    "'width',\n",
    "'normalized_losses_NA',\n",
    "'stroke_ordered',\n",
    "'horsepower_ordered',\n",
    "'compression_ratio_ordered',\n",
    "'wheel_base_ordered',\n",
    "'widht'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train = pd.concat([X_train_scaledR, X_train_scaledMM, X_train_scaledN], axis=1)\n",
    "X_test = pd.concat([X_test_scaledR, X_test_scaledMM, X_test_scaledN], axis=1)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>aspiration</th>\n",
       "      <th>body_style</th>\n",
       "      <th>engine_type</th>\n",
       "      <th>fuel_type</th>\n",
       "      <th>num_of_cylinders</th>\n",
       "      <th>make_id</th>\n",
       "      <th>city_mpg</th>\n",
       "      <th>bore</th>\n",
       "      <th>curb_weight</th>\n",
       "      <th>drive_wheels</th>\n",
       "      <th>engine_location</th>\n",
       "      <th>engine_size</th>\n",
       "      <th>fuel_system</th>\n",
       "      <th>height</th>\n",
       "      <th>highway_mpg</th>\n",
       "      <th>length</th>\n",
       "      <th>normalized_losses</th>\n",
       "      <th>num_of_doors</th>\n",
       "      <th>peak_rpm</th>\n",
       "      <th>symboling</th>\n",
       "      <th>width</th>\n",
       "      <th>normalized_losses_NA</th>\n",
       "      <th>stroke_ordered</th>\n",
       "      <th>horsepower_ordered</th>\n",
       "      <th>compression_ratio_ordered</th>\n",
       "      <th>wheel_base_ordered</th>\n",
       "      <th>widht</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.580772</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.093924</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>-0.550694</td>\n",
       "      <td>-0.021960</td>\n",
       "      <td>-0.778344</td>\n",
       "      <td>-0.132453</td>\n",
       "      <td>-0.444750</td>\n",
       "      <td>0.965406</td>\n",
       "      <td>0.529179</td>\n",
       "      <td>0.069328</td>\n",
       "      <td>0.698583</td>\n",
       "      <td>-0.181097</td>\n",
       "      <td>0.820413</td>\n",
       "      <td>0.784944</td>\n",
       "      <td>0.346142</td>\n",
       "      <td>0.453360</td>\n",
       "      <td>2.106537</td>\n",
       "      <td>-0.169365</td>\n",
       "      <td>-0.687669</td>\n",
       "      <td>-0.879240</td>\n",
       "      <td>0.185701</td>\n",
       "      <td>0.453360</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.0</td>\n",
       "      <td>-1.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.465623</td>\n",
       "      <td>0.166667</td>\n",
       "      <td>2.156108</td>\n",
       "      <td>0.386956</td>\n",
       "      <td>1.309589</td>\n",
       "      <td>-0.132453</td>\n",
       "      <td>0.536008</td>\n",
       "      <td>0.965406</td>\n",
       "      <td>-1.461019</td>\n",
       "      <td>-0.512024</td>\n",
       "      <td>-0.459224</td>\n",
       "      <td>2.150649</td>\n",
       "      <td>-1.218899</td>\n",
       "      <td>0.784944</td>\n",
       "      <td>1.373468</td>\n",
       "      <td>1.117853</td>\n",
       "      <td>-0.474713</td>\n",
       "      <td>0.506168</td>\n",
       "      <td>0.599692</td>\n",
       "      <td>-0.526895</td>\n",
       "      <td>-0.820603</td>\n",
       "      <td>1.117853</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.069243</td>\n",
       "      <td>0.444444</td>\n",
       "      <td>-0.117606</td>\n",
       "      <td>-0.472718</td>\n",
       "      <td>-0.778344</td>\n",
       "      <td>-0.132453</td>\n",
       "      <td>-0.141182</td>\n",
       "      <td>0.965406</td>\n",
       "      <td>0.447946</td>\n",
       "      <td>0.505342</td>\n",
       "      <td>0.087065</td>\n",
       "      <td>-1.767980</td>\n",
       "      <td>0.820413</td>\n",
       "      <td>-1.901203</td>\n",
       "      <td>1.227484</td>\n",
       "      <td>0.263506</td>\n",
       "      <td>-0.474713</td>\n",
       "      <td>-0.411287</td>\n",
       "      <td>-0.176283</td>\n",
       "      <td>-0.879240</td>\n",
       "      <td>0.859699</td>\n",
       "      <td>0.263506</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.094040</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>-0.695057</td>\n",
       "      <td>-1.210668</td>\n",
       "      <td>-0.778344</td>\n",
       "      <td>-0.132453</td>\n",
       "      <td>-0.724967</td>\n",
       "      <td>-1.210562</td>\n",
       "      <td>0.285481</td>\n",
       "      <td>0.941356</td>\n",
       "      <td>-0.752753</td>\n",
       "      <td>0.077985</td>\n",
       "      <td>0.820413</td>\n",
       "      <td>0.165064</td>\n",
       "      <td>-1.211110</td>\n",
       "      <td>-1.018015</td>\n",
       "      <td>-0.474713</td>\n",
       "      <td>-0.793459</td>\n",
       "      <td>-0.889405</td>\n",
       "      <td>-0.526895</td>\n",
       "      <td>-0.820603</td>\n",
       "      <td>-1.018015</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.775761</td>\n",
       "      <td>0.736426</td>\n",
       "      <td>0.083333</td>\n",
       "      <td>1.001206</td>\n",
       "      <td>1.531919</td>\n",
       "      <td>1.309589</td>\n",
       "      <td>-0.132453</td>\n",
       "      <td>1.890389</td>\n",
       "      <td>0.965406</td>\n",
       "      <td>-0.039449</td>\n",
       "      <td>-1.238715</td>\n",
       "      <td>1.571016</td>\n",
       "      <td>-0.181097</td>\n",
       "      <td>-1.218899</td>\n",
       "      <td>0.578317</td>\n",
       "      <td>0.346142</td>\n",
       "      <td>0.927998</td>\n",
       "      <td>2.106537</td>\n",
       "      <td>1.307310</td>\n",
       "      <td>2.518011</td>\n",
       "      <td>1.460642</td>\n",
       "      <td>0.859699</td>\n",
       "      <td>0.927998</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   aspiration  body_style  engine_type  fuel_type  num_of_cylinders   make_id  \\\n",
       "0         0.0   -0.580772          0.0        0.0          0.000000  0.093924   \n",
       "1         0.0   -1.000000          0.0        0.0          0.000000  0.465623   \n",
       "2         0.0    0.000000          0.0        0.0          0.000000  0.069243   \n",
       "3         0.0    0.000000          0.0        0.0          0.000000  0.094040   \n",
       "4         0.0    0.000000          0.0        0.0          0.775761  0.736426   \n",
       "\n",
       "   city_mpg      bore  curb_weight  drive_wheels  engine_location  \\\n",
       "0  0.333333 -0.550694    -0.021960     -0.778344        -0.132453   \n",
       "1  0.166667  2.156108     0.386956      1.309589        -0.132453   \n",
       "2  0.444444 -0.117606    -0.472718     -0.778344        -0.132453   \n",
       "3  0.500000 -0.695057    -1.210668     -0.778344        -0.132453   \n",
       "4  0.083333  1.001206     1.531919      1.309589        -0.132453   \n",
       "\n",
       "   engine_size  fuel_system    height  highway_mpg    length  \\\n",
       "0    -0.444750     0.965406  0.529179     0.069328  0.698583   \n",
       "1     0.536008     0.965406 -1.461019    -0.512024 -0.459224   \n",
       "2    -0.141182     0.965406  0.447946     0.505342  0.087065   \n",
       "3    -0.724967    -1.210562  0.285481     0.941356 -0.752753   \n",
       "4     1.890389     0.965406 -0.039449    -1.238715  1.571016   \n",
       "\n",
       "   normalized_losses  num_of_doors  peak_rpm  symboling     width  \\\n",
       "0          -0.181097      0.820413  0.784944   0.346142  0.453360   \n",
       "1           2.150649     -1.218899  0.784944   1.373468  1.117853   \n",
       "2          -1.767980      0.820413 -1.901203   1.227484  0.263506   \n",
       "3           0.077985      0.820413  0.165064  -1.211110 -1.018015   \n",
       "4          -0.181097     -1.218899  0.578317   0.346142  0.927998   \n",
       "\n",
       "   normalized_losses_NA  stroke_ordered  horsepower_ordered  \\\n",
       "0              2.106537       -0.169365           -0.687669   \n",
       "1             -0.474713        0.506168            0.599692   \n",
       "2             -0.474713       -0.411287           -0.176283   \n",
       "3             -0.474713       -0.793459           -0.889405   \n",
       "4              2.106537        1.307310            2.518011   \n",
       "\n",
       "   compression_ratio_ordered  wheel_base_ordered     widht  \n",
       "0                  -0.879240            0.185701  0.453360  \n",
       "1                  -0.526895           -0.820603  1.117853  \n",
       "2                  -0.879240            0.859699  0.263506  \n",
       "3                  -0.526895           -0.820603 -1.018015  \n",
       "4                   1.460642            0.859699  0.927998  "
      ]
     },
     "execution_count": 109,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train = pd.DataFrame(X_train)\n",
    "X_train.to_csv('X_train.csv', index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>aspiration</th>\n",
       "      <th>body_style</th>\n",
       "      <th>engine_type</th>\n",
       "      <th>fuel_type</th>\n",
       "      <th>num_of_cylinders</th>\n",
       "      <th>make_id</th>\n",
       "      <th>city_mpg</th>\n",
       "      <th>bore</th>\n",
       "      <th>curb_weight</th>\n",
       "      <th>drive_wheels</th>\n",
       "      <th>engine_location</th>\n",
       "      <th>engine_size</th>\n",
       "      <th>fuel_system</th>\n",
       "      <th>height</th>\n",
       "      <th>highway_mpg</th>\n",
       "      <th>length</th>\n",
       "      <th>normalized_losses</th>\n",
       "      <th>num_of_doors</th>\n",
       "      <th>peak_rpm</th>\n",
       "      <th>symboling</th>\n",
       "      <th>width</th>\n",
       "      <th>normalized_losses_NA</th>\n",
       "      <th>stroke_ordered</th>\n",
       "      <th>horsepower_ordered</th>\n",
       "      <th>compression_ratio_ordered</th>\n",
       "      <th>wheel_base_ordered</th>\n",
       "      <th>widht</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.0</td>\n",
       "      <td>-1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.096338</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>-1.128145</td>\n",
       "      <td>-1.273431</td>\n",
       "      <td>-0.778344</td>\n",
       "      <td>-0.132453</td>\n",
       "      <td>-0.865075</td>\n",
       "      <td>-1.210562</td>\n",
       "      <td>0.123016</td>\n",
       "      <td>1.086694</td>\n",
       "      <td>-1.258275</td>\n",
       "      <td>-0.504951</td>\n",
       "      <td>-1.218899</td>\n",
       "      <td>-0.248189</td>\n",
       "      <td>-1.211110</td>\n",
       "      <td>-0.828160</td>\n",
       "      <td>-0.474713</td>\n",
       "      <td>0.506168</td>\n",
       "      <td>-0.927738</td>\n",
       "      <td>-0.879240</td>\n",
       "      <td>-0.592607</td>\n",
       "      <td>-0.828160</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.580772</td>\n",
       "      <td>2.631301</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.775761</td>\n",
       "      <td>0.069243</td>\n",
       "      <td>0.166667</td>\n",
       "      <td>-0.261969</td>\n",
       "      <td>1.096377</td>\n",
       "      <td>1.309589</td>\n",
       "      <td>-0.132453</td>\n",
       "      <td>0.769522</td>\n",
       "      <td>0.965406</td>\n",
       "      <td>0.123016</td>\n",
       "      <td>-0.948038</td>\n",
       "      <td>1.081801</td>\n",
       "      <td>-0.181097</td>\n",
       "      <td>0.820413</td>\n",
       "      <td>0.165064</td>\n",
       "      <td>1.227484</td>\n",
       "      <td>0.263506</td>\n",
       "      <td>2.106537</td>\n",
       "      <td>0.506168</td>\n",
       "      <td>-0.927738</td>\n",
       "      <td>-1.625007</td>\n",
       "      <td>-0.592607</td>\n",
       "      <td>0.263506</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.611065</td>\n",
       "      <td>0.465623</td>\n",
       "      <td>0.166667</td>\n",
       "      <td>-0.550694</td>\n",
       "      <td>-0.128468</td>\n",
       "      <td>-0.778344</td>\n",
       "      <td>-0.132453</td>\n",
       "      <td>0.185738</td>\n",
       "      <td>0.965406</td>\n",
       "      <td>-0.283147</td>\n",
       "      <td>-0.802700</td>\n",
       "      <td>0.225676</td>\n",
       "      <td>-0.181097</td>\n",
       "      <td>-1.218899</td>\n",
       "      <td>0.784944</td>\n",
       "      <td>-0.958794</td>\n",
       "      <td>0.168578</td>\n",
       "      <td>2.106537</td>\n",
       "      <td>0.506168</td>\n",
       "      <td>-0.927738</td>\n",
       "      <td>1.460642</td>\n",
       "      <td>-0.592607</td>\n",
       "      <td>0.168578</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.0</td>\n",
       "      <td>-1.000000</td>\n",
       "      <td>1.383807</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.465623</td>\n",
       "      <td>0.944444</td>\n",
       "      <td>-1.561233</td>\n",
       "      <td>-2.066537</td>\n",
       "      <td>-0.778344</td>\n",
       "      <td>-0.132453</td>\n",
       "      <td>-1.565617</td>\n",
       "      <td>-1.210562</td>\n",
       "      <td>-0.242531</td>\n",
       "      <td>3.266765</td>\n",
       "      <td>-2.725918</td>\n",
       "      <td>0.045600</td>\n",
       "      <td>-1.218899</td>\n",
       "      <td>-0.041563</td>\n",
       "      <td>-0.958794</td>\n",
       "      <td>-2.679246</td>\n",
       "      <td>-0.474713</td>\n",
       "      <td>0.506168</td>\n",
       "      <td>-0.927738</td>\n",
       "      <td>-0.526895</td>\n",
       "      <td>-0.592607</td>\n",
       "      <td>-2.679246</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.093924</td>\n",
       "      <td>0.361111</td>\n",
       "      <td>-0.550694</td>\n",
       "      <td>-0.522168</td>\n",
       "      <td>-0.778344</td>\n",
       "      <td>-0.132453</td>\n",
       "      <td>-0.444750</td>\n",
       "      <td>0.965406</td>\n",
       "      <td>0.772876</td>\n",
       "      <td>0.214666</td>\n",
       "      <td>-0.230924</td>\n",
       "      <td>-0.828805</td>\n",
       "      <td>0.820413</td>\n",
       "      <td>0.784944</td>\n",
       "      <td>-0.958794</td>\n",
       "      <td>-0.211132</td>\n",
       "      <td>-0.474713</td>\n",
       "      <td>0.506168</td>\n",
       "      <td>-0.927738</td>\n",
       "      <td>0.469621</td>\n",
       "      <td>-0.592607</td>\n",
       "      <td>-0.211132</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   aspiration  body_style  engine_type  fuel_type  num_of_cylinders   make_id  \\\n",
       "0         0.0   -1.000000     0.000000        0.0          0.000000  0.096338   \n",
       "1         0.0   -0.580772     2.631301        0.0          0.775761  0.069243   \n",
       "2         0.0    0.000000     0.000000        0.0          0.611065  0.465623   \n",
       "3         0.0   -1.000000     1.383807        0.0          1.000000  0.465623   \n",
       "4         0.0    0.000000     0.000000        0.0          0.000000  0.093924   \n",
       "\n",
       "   city_mpg      bore  curb_weight  drive_wheels  engine_location  \\\n",
       "0  0.500000 -1.128145    -1.273431     -0.778344        -0.132453   \n",
       "1  0.166667 -0.261969     1.096377      1.309589        -0.132453   \n",
       "2  0.166667 -0.550694    -0.128468     -0.778344        -0.132453   \n",
       "3  0.944444 -1.561233    -2.066537     -0.778344        -0.132453   \n",
       "4  0.361111 -0.550694    -0.522168     -0.778344        -0.132453   \n",
       "\n",
       "   engine_size  fuel_system    height  highway_mpg    length  \\\n",
       "0    -0.865075    -1.210562  0.123016     1.086694 -1.258275   \n",
       "1     0.769522     0.965406  0.123016    -0.948038  1.081801   \n",
       "2     0.185738     0.965406 -0.283147    -0.802700  0.225676   \n",
       "3    -1.565617    -1.210562 -0.242531     3.266765 -2.725918   \n",
       "4    -0.444750     0.965406  0.772876     0.214666 -0.230924   \n",
       "\n",
       "   normalized_losses  num_of_doors  peak_rpm  symboling     width  \\\n",
       "0          -0.504951     -1.218899 -0.248189  -1.211110 -0.828160   \n",
       "1          -0.181097      0.820413  0.165064   1.227484  0.263506   \n",
       "2          -0.181097     -1.218899  0.784944  -0.958794  0.168578   \n",
       "3           0.045600     -1.218899 -0.041563  -0.958794 -2.679246   \n",
       "4          -0.828805      0.820413  0.784944  -0.958794 -0.211132   \n",
       "\n",
       "   normalized_losses_NA  stroke_ordered  horsepower_ordered  \\\n",
       "0             -0.474713        0.506168           -0.927738   \n",
       "1              2.106537        0.506168           -0.927738   \n",
       "2              2.106537        0.506168           -0.927738   \n",
       "3             -0.474713        0.506168           -0.927738   \n",
       "4             -0.474713        0.506168           -0.927738   \n",
       "\n",
       "   compression_ratio_ordered  wheel_base_ordered     widht  \n",
       "0                  -0.879240           -0.592607 -0.828160  \n",
       "1                  -1.625007           -0.592607  0.263506  \n",
       "2                   1.460642           -0.592607  0.168578  \n",
       "3                  -0.526895           -0.592607 -2.679246  \n",
       "4                   0.469621           -0.592607 -0.211132  "
      ]
     },
     "execution_count": 111,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test = pd.DataFrame(X_test)\n",
    "X_test.to_csv('X_test.csv', index=False)\n",
    "\n",
    "X_test.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>price</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>193</th>\n",
       "      <td>12290.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>125</th>\n",
       "      <td>22018.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>173</th>\n",
       "      <td>8948.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>92</th>\n",
       "      <td>6849.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>41315.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       price\n",
       "193  12290.0\n",
       "125  22018.0\n",
       "173   8948.0\n",
       "92    6849.0\n",
       "16   41315.0"
      ]
     },
     "execution_count": 112,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train = pd.DataFrame(y_train)\n",
    "y_train.to_csv('y_train.csv', index=False)\n",
    "\n",
    "y_train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_test = pd.DataFrame(y_test)\n",
    "y_test.to_csv('y_test.csv', index=False)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}