Tennis_Ensembles_Hot3_Batch49_2383.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 267,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "\n",
    "#Randomforest\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.ensemble import BaggingClassifier\n",
    "from xgboost import XGBClassifier\n",
    "from sklearn.ensemble import AdaBoostClassifier\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.metrics import accuracy_score, log_loss\n",
    "from catboost import CatBoostClassifier\n",
    "\n",
    "\n",
    "import math\n",
    "from copy import deepcopy\n",
    "from collections import defaultdict\n",
    "\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "\n",
    "import os"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "#change directory\n",
    "os.chdir(\"D:/HoTes/CSC7305c_ML_Hot_2383_Batch49/Tennis dataset_stacking_adab_xgb\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "#read datasets\n",
    "tennis_data= pd.read_csv(\"Batch49_CSE7305c_CuTE_train-1544001028037.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(8001, 27)"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#shape\n",
    "tennis_data.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 8001 entries, 0 to 8000\n",
      "Data columns (total 27 columns):\n",
      "rally                                 8001 non-null int64\n",
      "serve                                 8001 non-null int64\n",
      "hitpoint                              8001 non-null object\n",
      "speed                                 8001 non-null float64\n",
      "net.clearance                         8001 non-null float64\n",
      "distance.from.sideline                8001 non-null float64\n",
      "depth                                 8001 non-null float64\n",
      "outside.sideline                      8001 non-null bool\n",
      "outside.baseline                      8001 non-null bool\n",
      "player.distance.travelled             8001 non-null float64\n",
      "player.impact.depth                   8001 non-null float64\n",
      "player.impact.distance.from.center    8001 non-null float64\n",
      "player.depth                          8001 non-null float64\n",
      "player.distance.from.center           8001 non-null float64\n",
      "previous.speed                        8001 non-null float64\n",
      "previous.net.clearance                8001 non-null float64\n",
      "previous.distance.from.sideline       8001 non-null float64\n",
      "previous.depth                        8001 non-null float64\n",
      "opponent.depth                        8001 non-null float64\n",
      "opponent.distance.from.center         8001 non-null float64\n",
      "same.side                             8001 non-null bool\n",
      "previous.hitpoint                     8001 non-null object\n",
      "previous.time.to.net                  8001 non-null float64\n",
      "server.is.impact.player               8001 non-null bool\n",
      "outcome                               8001 non-null object\n",
      "gender                                8001 non-null object\n",
      "ID                                    8001 non-null int64\n",
      "dtypes: bool(4), float64(16), int64(3), object(4)\n",
      "memory usage: 1.4+ MB\n"
     ]
    }
   ],
   "source": [
    "#datatypes\n",
    "tennis_data.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Columns in Data With Categorical Values- Must LabelEncode them\n",
    "categorical_cols = ['hitpoint', 'outside.sideline', 'outside.baseline', 'same.side', \n",
    "                    'previous.hitpoint', 'server.is.impact.player', 'outcome']\n",
    "\n",
    "# Columns in the Data That Should Be Scaled\n",
    "scaled_data = ['serve', 'rally', 'speed', 'net.clearance', 'distance.from.sideline', 'depth', 'player.distance.travelled', \n",
    "               'player.impact.depth', 'player.impact.distance.from.center', 'player.depth', \n",
    "               'player.distance.from.center', 'previous.speed', 'previous.net.clearance', \n",
    "               'previous.distance.from.sideline', 'previous.depth', 'opponent.depth', \n",
    "               'opponent.distance.from.center', 'previous.time.to.net']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "tennis_data = tennis_data.drop(['gender', 'ID'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "tennis_data[categorical_cols]= tennis_data[categorical_cols].astype('category')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "rally                                 0\n",
       "serve                                 0\n",
       "hitpoint                              0\n",
       "speed                                 0\n",
       "net.clearance                         0\n",
       "distance.from.sideline                0\n",
       "depth                                 0\n",
       "outside.sideline                      0\n",
       "outside.baseline                      0\n",
       "player.distance.travelled             0\n",
       "player.impact.depth                   0\n",
       "player.impact.distance.from.center    0\n",
       "player.depth                          0\n",
       "player.distance.from.center           0\n",
       "previous.speed                        0\n",
       "previous.net.clearance                0\n",
       "previous.distance.from.sideline       0\n",
       "previous.depth                        0\n",
       "opponent.depth                        0\n",
       "opponent.distance.from.center         0\n",
       "same.side                             0\n",
       "previous.hitpoint                     0\n",
       "previous.time.to.net                  0\n",
       "server.is.impact.player               0\n",
       "outcome                               0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tennis_data.isnull().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Feature Engineering"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "####################### Feature Engineering #################\n",
    "def feature_engineer(data):\n",
    "    \n",
    "    # Speed difference between previous and last shot\n",
    "    data['speed.difference'] = data['speed'] - data['previous.speed']\n",
    "    \n",
    "    # Net clearance difference between previous and last shot\n",
    "    data['net.clearance.difference'] = data['net.clearance'] - data['previous.net.clearance']\n",
    "    \n",
    "    # The actual distance the player was from the sideline\n",
    "    data['true.distance.from.sideline'] = [dist if not boolean else (dist * -1) \n",
    "                                           for dist, boolean in \n",
    "                                           zip(data['distance.from.sideline'], \n",
    "                                               data['outside.sideline'])]\n",
    "\n",
    "    # The actual distance the player was from the baseline\n",
    "    data['true.distance.from.baseline'] = [dist if not boolean else (dist * -1) \n",
    "                                           for dist, boolean in \n",
    "                                           zip(data['depth'], \n",
    "                                               data['outside.baseline'])]\n",
    "\n",
    "    # Opponent distance to net + player distance to net\n",
    "    data['player.distance.from.other.length'] = data['player.depth'] + data['opponent.depth']\n",
    "\n",
    "    # Opponent distance from center + player distance to center\n",
    "    data['player.distance.from.other.width'] = [math.fabs(p_center - o_center)\n",
    "                                                     if boolean else p_center + o_center\n",
    "                                                     for p_center, o_center, boolean in \n",
    "                                                     zip(data['player.distance.from.center'], \n",
    "                                                         data['opponent.distance.from.center'], \n",
    "                                                         data['same.side'])]\n",
    "\n",
    "    # Straight line distance between players\n",
    "    data['bw.player.distance.penultimate'] = [math.hypot(length, width)\n",
    "                                                 for length, width in \n",
    "                                                 zip(data['player.distance.from.other.length'], \n",
    "                                                     data['player.distance.from.other.width'])]\n",
    "\n",
    "    # Distance penultimate shot was made from net- previous to net (s) * penultimate shot (m/s)\n",
    "    data['player.from.net.penultimate'] = data['previous.speed'] * data['previous.time.to.net']\n",
    "    \n",
    "    return data\n",
    "\n",
    "# Perform Feature Engineering\n",
    "tennis_data_fe = feature_engineer(tennis_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "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>rally</th>\n",
       "      <th>serve</th>\n",
       "      <th>hitpoint</th>\n",
       "      <th>speed</th>\n",
       "      <th>net.clearance</th>\n",
       "      <th>distance.from.sideline</th>\n",
       "      <th>depth</th>\n",
       "      <th>outside.sideline</th>\n",
       "      <th>outside.baseline</th>\n",
       "      <th>player.distance.travelled</th>\n",
       "      <th>...</th>\n",
       "      <th>server.is.impact.player</th>\n",
       "      <th>outcome</th>\n",
       "      <th>speed.difference</th>\n",
       "      <th>net.clearance.difference</th>\n",
       "      <th>true.distance.from.sideline</th>\n",
       "      <th>true.distance.from.baseline</th>\n",
       "      <th>player.distance.from.other.length</th>\n",
       "      <th>player.distance.from.other.width</th>\n",
       "      <th>bw.player.distance.penultimate</th>\n",
       "      <th>player.from.net.penultimate</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>B</td>\n",
       "      <td>35.515042</td>\n",
       "      <td>-0.021725</td>\n",
       "      <td>3.474766</td>\n",
       "      <td>6.797621</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>1.467570</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>UE</td>\n",
       "      <td>3.201450</td>\n",
       "      <td>-1.372257</td>\n",
       "      <td>3.474766</td>\n",
       "      <td>6.797621</td>\n",
       "      <td>26.2332</td>\n",
       "      <td>1.8394</td>\n",
       "      <td>26.297608</td>\n",
       "      <td>14.389823</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>4</td>\n",
       "      <td>2</td>\n",
       "      <td>B</td>\n",
       "      <td>33.382640</td>\n",
       "      <td>1.114202</td>\n",
       "      <td>2.540801</td>\n",
       "      <td>2.608708</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>2.311931</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>FE</td>\n",
       "      <td>-0.678148</td>\n",
       "      <td>0.353704</td>\n",
       "      <td>2.540801</td>\n",
       "      <td>-2.608708</td>\n",
       "      <td>24.9902</td>\n",
       "      <td>7.4286</td>\n",
       "      <td>26.070945</td>\n",
       "      <td>14.729042</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>23</td>\n",
       "      <td>1</td>\n",
       "      <td>B</td>\n",
       "      <td>22.316690</td>\n",
       "      <td>-0.254046</td>\n",
       "      <td>3.533166</td>\n",
       "      <td>9.435749</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>3.903728</td>\n",
       "      <td>...</td>\n",
       "      <td>True</td>\n",
       "      <td>FE</td>\n",
       "      <td>-17.489938</td>\n",
       "      <td>-0.631778</td>\n",
       "      <td>3.533166</td>\n",
       "      <td>9.435749</td>\n",
       "      <td>27.5616</td>\n",
       "      <td>3.7472</td>\n",
       "      <td>27.815163</td>\n",
       "      <td>15.824638</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>9</td>\n",
       "      <td>1</td>\n",
       "      <td>F</td>\n",
       "      <td>36.837309</td>\n",
       "      <td>0.766694</td>\n",
       "      <td>0.586885</td>\n",
       "      <td>3.342180</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>0.583745</td>\n",
       "      <td>...</td>\n",
       "      <td>True</td>\n",
       "      <td>UE</td>\n",
       "      <td>11.818491</td>\n",
       "      <td>-0.472339</td>\n",
       "      <td>-0.586885</td>\n",
       "      <td>3.342180</td>\n",
       "      <td>27.4088</td>\n",
       "      <td>0.1418</td>\n",
       "      <td>27.409167</td>\n",
       "      <td>16.812251</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>B</td>\n",
       "      <td>35.544208</td>\n",
       "      <td>0.116162</td>\n",
       "      <td>0.918725</td>\n",
       "      <td>5.499119</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>2.333456</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>W</td>\n",
       "      <td>-1.246771</td>\n",
       "      <td>-0.049569</td>\n",
       "      <td>0.918725</td>\n",
       "      <td>5.499119</td>\n",
       "      <td>25.4968</td>\n",
       "      <td>1.4232</td>\n",
       "      <td>25.536490</td>\n",
       "      <td>12.524057</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 33 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   rally  serve hitpoint      speed  net.clearance  distance.from.sideline  \\\n",
       "0      4      1        B  35.515042      -0.021725                3.474766   \n",
       "1      4      2        B  33.382640       1.114202                2.540801   \n",
       "2     23      1        B  22.316690      -0.254046                3.533166   \n",
       "3      9      1        F  36.837309       0.766694                0.586885   \n",
       "4      4      1        B  35.544208       0.116162                0.918725   \n",
       "\n",
       "      depth outside.sideline outside.baseline  player.distance.travelled  \\\n",
       "0  6.797621            False            False                   1.467570   \n",
       "1  2.608708            False             True                   2.311931   \n",
       "2  9.435749            False            False                   3.903728   \n",
       "3  3.342180             True            False                   0.583745   \n",
       "4  5.499119            False            False                   2.333456   \n",
       "\n",
       "              ...               server.is.impact.player  outcome  \\\n",
       "0             ...                                 False       UE   \n",
       "1             ...                                 False       FE   \n",
       "2             ...                                  True       FE   \n",
       "3             ...                                  True       UE   \n",
       "4             ...                                 False        W   \n",
       "\n",
       "   speed.difference  net.clearance.difference  true.distance.from.sideline  \\\n",
       "0          3.201450                 -1.372257                     3.474766   \n",
       "1         -0.678148                  0.353704                     2.540801   \n",
       "2        -17.489938                 -0.631778                     3.533166   \n",
       "3         11.818491                 -0.472339                    -0.586885   \n",
       "4         -1.246771                 -0.049569                     0.918725   \n",
       "\n",
       "   true.distance.from.baseline  player.distance.from.other.length  \\\n",
       "0                     6.797621                            26.2332   \n",
       "1                    -2.608708                            24.9902   \n",
       "2                     9.435749                            27.5616   \n",
       "3                     3.342180                            27.4088   \n",
       "4                     5.499119                            25.4968   \n",
       "\n",
       "   player.distance.from.other.width  bw.player.distance.penultimate  \\\n",
       "0                            1.8394                       26.297608   \n",
       "1                            7.4286                       26.070945   \n",
       "2                            3.7472                       27.815163   \n",
       "3                            0.1418                       27.409167   \n",
       "4                            1.4232                       25.536490   \n",
       "\n",
       "   player.from.net.penultimate  \n",
       "0                    14.389823  \n",
       "1                    14.729042  \n",
       "2                    15.824638  \n",
       "3                    16.812251  \n",
       "4                    12.524057  \n",
       "\n",
       "[5 rows x 33 columns]"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tennis_data_fe.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(8001, 33)"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tennis_data_fe.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Columns to be dropped\n",
    "dropcols = ['same.side', 'server.is.impact.player', \n",
    "                  'outside.baseline', 'previous.hitpoint', 'hitpoint', 'rally', 'serve', \n",
    "                  'outside.sideline', 'player.distance.from.center', 'depth', \n",
    "                  'player.distance.from.other.length', 'bw.player.distance.penultimate',\n",
    "                  'player.distance.from.other.width', 'previous.time.to.net',\n",
    "                  'player.impact.distance.from.center', 'opponent.distance.from.center', \n",
    "                  'player.from.net.penultimate','net.clearance.difference']\n",
    "\n",
    "tennis_data_fe = tennis_data_fe.drop(dropcols, axis=1) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(8001, 15)"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tennis_data_fe.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "target = tennis_data_fe['outcome']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [],
   "source": [
    "tennis_data_fe= tennis_data_fe.drop('outcome', axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 8001 entries, 0 to 8000\n",
      "Data columns (total 14 columns):\n",
      "speed                              8001 non-null float64\n",
      "net.clearance                      8001 non-null float64\n",
      "distance.from.sideline             8001 non-null float64\n",
      "player.distance.travelled          8001 non-null float64\n",
      "player.impact.depth                8001 non-null float64\n",
      "player.depth                       8001 non-null float64\n",
      "previous.speed                     8001 non-null float64\n",
      "previous.net.clearance             8001 non-null float64\n",
      "previous.distance.from.sideline    8001 non-null float64\n",
      "previous.depth                     8001 non-null float64\n",
      "opponent.depth                     8001 non-null float64\n",
      "speed.difference                   8001 non-null float64\n",
      "true.distance.from.sideline        8001 non-null float64\n",
      "true.distance.from.baseline        8001 non-null float64\n",
      "dtypes: float64(14)\n",
      "memory usage: 875.2 KB\n"
     ]
    }
   ],
   "source": [
    "tennis_data_fe.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "tennis_data_fe= pd.get_dummies(tennis_data_fe, drop_first=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(8001, 16)"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tennis_data_fe.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "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>speed</th>\n",
       "      <th>net.clearance</th>\n",
       "      <th>distance.from.sideline</th>\n",
       "      <th>player.distance.travelled</th>\n",
       "      <th>player.impact.depth</th>\n",
       "      <th>player.depth</th>\n",
       "      <th>previous.speed</th>\n",
       "      <th>previous.net.clearance</th>\n",
       "      <th>previous.distance.from.sideline</th>\n",
       "      <th>previous.depth</th>\n",
       "      <th>opponent.depth</th>\n",
       "      <th>speed.difference</th>\n",
       "      <th>true.distance.from.sideline</th>\n",
       "      <th>true.distance.from.baseline</th>\n",
       "      <th>outcome_UE</th>\n",
       "      <th>outcome_W</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>35.515042</td>\n",
       "      <td>-0.021725</td>\n",
       "      <td>3.474766</td>\n",
       "      <td>1.467570</td>\n",
       "      <td>14.1758</td>\n",
       "      <td>13.6704</td>\n",
       "      <td>32.313592</td>\n",
       "      <td>1.350533</td>\n",
       "      <td>2.449182</td>\n",
       "      <td>0.705435</td>\n",
       "      <td>12.5628</td>\n",
       "      <td>3.201450</td>\n",
       "      <td>3.474766</td>\n",
       "      <td>6.797621</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>33.382640</td>\n",
       "      <td>1.114202</td>\n",
       "      <td>2.540801</td>\n",
       "      <td>2.311931</td>\n",
       "      <td>12.7396</td>\n",
       "      <td>12.6358</td>\n",
       "      <td>34.060788</td>\n",
       "      <td>0.760499</td>\n",
       "      <td>0.583291</td>\n",
       "      <td>3.856600</td>\n",
       "      <td>12.3544</td>\n",
       "      <td>-0.678148</td>\n",
       "      <td>2.540801</td>\n",
       "      <td>-2.608708</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>22.316690</td>\n",
       "      <td>-0.254046</td>\n",
       "      <td>3.533166</td>\n",
       "      <td>3.903728</td>\n",
       "      <td>15.3782</td>\n",
       "      <td>13.6996</td>\n",
       "      <td>39.806628</td>\n",
       "      <td>0.377732</td>\n",
       "      <td>1.115250</td>\n",
       "      <td>2.908892</td>\n",
       "      <td>13.8620</td>\n",
       "      <td>-17.489938</td>\n",
       "      <td>3.533166</td>\n",
       "      <td>9.435749</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>36.837309</td>\n",
       "      <td>0.766694</td>\n",
       "      <td>0.586885</td>\n",
       "      <td>0.583745</td>\n",
       "      <td>13.7308</td>\n",
       "      <td>13.1492</td>\n",
       "      <td>25.018818</td>\n",
       "      <td>1.239033</td>\n",
       "      <td>3.256695</td>\n",
       "      <td>0.557554</td>\n",
       "      <td>14.2596</td>\n",
       "      <td>11.818491</td>\n",
       "      <td>-0.586885</td>\n",
       "      <td>3.342180</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>35.544208</td>\n",
       "      <td>0.116162</td>\n",
       "      <td>0.918725</td>\n",
       "      <td>2.333456</td>\n",
       "      <td>15.1992</td>\n",
       "      <td>14.1310</td>\n",
       "      <td>36.790979</td>\n",
       "      <td>0.165731</td>\n",
       "      <td>1.431146</td>\n",
       "      <td>3.945317</td>\n",
       "      <td>11.3658</td>\n",
       "      <td>-1.246771</td>\n",
       "      <td>0.918725</td>\n",
       "      <td>5.499119</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       speed  net.clearance  distance.from.sideline  \\\n",
       "0  35.515042      -0.021725                3.474766   \n",
       "1  33.382640       1.114202                2.540801   \n",
       "2  22.316690      -0.254046                3.533166   \n",
       "3  36.837309       0.766694                0.586885   \n",
       "4  35.544208       0.116162                0.918725   \n",
       "\n",
       "   player.distance.travelled  player.impact.depth  player.depth  \\\n",
       "0                   1.467570              14.1758       13.6704   \n",
       "1                   2.311931              12.7396       12.6358   \n",
       "2                   3.903728              15.3782       13.6996   \n",
       "3                   0.583745              13.7308       13.1492   \n",
       "4                   2.333456              15.1992       14.1310   \n",
       "\n",
       "   previous.speed  previous.net.clearance  previous.distance.from.sideline  \\\n",
       "0       32.313592                1.350533                         2.449182   \n",
       "1       34.060788                0.760499                         0.583291   \n",
       "2       39.806628                0.377732                         1.115250   \n",
       "3       25.018818                1.239033                         3.256695   \n",
       "4       36.790979                0.165731                         1.431146   \n",
       "\n",
       "   previous.depth  opponent.depth  speed.difference  \\\n",
       "0        0.705435         12.5628          3.201450   \n",
       "1        3.856600         12.3544         -0.678148   \n",
       "2        2.908892         13.8620        -17.489938   \n",
       "3        0.557554         14.2596         11.818491   \n",
       "4        3.945317         11.3658         -1.246771   \n",
       "\n",
       "   true.distance.from.sideline  true.distance.from.baseline  outcome_UE  \\\n",
       "0                     3.474766                     6.797621           1   \n",
       "1                     2.540801                    -2.608708           0   \n",
       "2                     3.533166                     9.435749           0   \n",
       "3                    -0.586885                     3.342180           1   \n",
       "4                     0.918725                     5.499119           0   \n",
       "\n",
       "   outcome_W  \n",
       "0          0  \n",
       "1          0  \n",
       "2          0  \n",
       "3          0  \n",
       "4          1  "
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tennis_data_fe.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 210,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x22c607bab70>"
      ]
     },
     "execution_count": 210,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "colormap = plt.cm.RdBu\n",
    "plt.figure(figsize=(10,10))\n",
    "plt.title('Pearson Correlation of Features', y=1.05, size=15)\n",
    "sns.heatmap(tennis_data_fe.corr(),linewidths=0.1,vmax=1.0, \n",
    "            square=True, cmap=colormap, linecolor='white', annot=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## splitting"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(tennis_data_fe, target, test_size = 0.25, random_state = 786)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#standardization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 162,
   "metadata": {},
   "outputs": [],
   "source": [
    "#train data standardization\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "standardizer = StandardScaler()\n",
    "standardizer.fit(X_train)\n",
    "X_train = pd.DataFrame(standardizer.transform(X_train), columns=tennis_data_fe.columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 164,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Test data standardization\n",
    "standardizer = StandardScaler()\n",
    "standardizer.fit(X_test)\n",
    "X_test = pd.DataFrame(standardizer.transform(X_test), columns=tennis_data_fe.columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 165,
   "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>speed</th>\n",
       "      <th>net.clearance</th>\n",
       "      <th>distance.from.sideline</th>\n",
       "      <th>player.distance.travelled</th>\n",
       "      <th>player.impact.depth</th>\n",
       "      <th>player.depth</th>\n",
       "      <th>previous.speed</th>\n",
       "      <th>previous.net.clearance</th>\n",
       "      <th>previous.distance.from.sideline</th>\n",
       "      <th>previous.depth</th>\n",
       "      <th>opponent.depth</th>\n",
       "      <th>speed.difference</th>\n",
       "      <th>true.distance.from.sideline</th>\n",
       "      <th>true.distance.from.baseline</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.251841</td>\n",
       "      <td>0.172085</td>\n",
       "      <td>0.317546</td>\n",
       "      <td>0.172419</td>\n",
       "      <td>0.639689</td>\n",
       "      <td>0.552107</td>\n",
       "      <td>0.482919</td>\n",
       "      <td>-0.381947</td>\n",
       "      <td>-0.446773</td>\n",
       "      <td>-0.451524</td>\n",
       "      <td>0.181863</td>\n",
       "      <td>0.600468</td>\n",
       "      <td>0.439649</td>\n",
       "      <td>-1.122823</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.385337</td>\n",
       "      <td>-0.741854</td>\n",
       "      <td>0.315626</td>\n",
       "      <td>0.176469</td>\n",
       "      <td>0.456398</td>\n",
       "      <td>-0.057737</td>\n",
       "      <td>1.079780</td>\n",
       "      <td>-0.886178</td>\n",
       "      <td>-0.412138</td>\n",
       "      <td>-0.318133</td>\n",
       "      <td>0.486530</td>\n",
       "      <td>-0.405127</td>\n",
       "      <td>0.438103</td>\n",
       "      <td>0.725961</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.166139</td>\n",
       "      <td>-0.374440</td>\n",
       "      <td>-0.855796</td>\n",
       "      <td>-1.054190</td>\n",
       "      <td>0.376936</td>\n",
       "      <td>0.503636</td>\n",
       "      <td>-0.736615</td>\n",
       "      <td>1.155253</td>\n",
       "      <td>1.393809</td>\n",
       "      <td>-0.676829</td>\n",
       "      <td>0.635215</td>\n",
       "      <td>1.311106</td>\n",
       "      <td>-0.505484</td>\n",
       "      <td>0.286552</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.730793</td>\n",
       "      <td>-0.194214</td>\n",
       "      <td>0.448117</td>\n",
       "      <td>-0.952160</td>\n",
       "      <td>-0.251170</td>\n",
       "      <td>-0.463607</td>\n",
       "      <td>0.453810</td>\n",
       "      <td>-0.717097</td>\n",
       "      <td>0.853026</td>\n",
       "      <td>0.693034</td>\n",
       "      <td>-0.005939</td>\n",
       "      <td>0.241626</td>\n",
       "      <td>0.544824</td>\n",
       "      <td>-0.419230</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.781429</td>\n",
       "      <td>2.228198</td>\n",
       "      <td>-0.979172</td>\n",
       "      <td>1.041107</td>\n",
       "      <td>1.220581</td>\n",
       "      <td>1.336942</td>\n",
       "      <td>1.149344</td>\n",
       "      <td>-0.580086</td>\n",
       "      <td>-1.062001</td>\n",
       "      <td>-1.221139</td>\n",
       "      <td>-0.127378</td>\n",
       "      <td>-0.162397</td>\n",
       "      <td>-0.604864</td>\n",
       "      <td>-2.882417</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      speed  net.clearance  distance.from.sideline  player.distance.travelled  \\\n",
       "0  1.251841       0.172085                0.317546                   0.172419   \n",
       "1  0.385337      -0.741854                0.315626                   0.176469   \n",
       "2  1.166139      -0.374440               -0.855796                  -1.054190   \n",
       "3  0.730793      -0.194214                0.448117                  -0.952160   \n",
       "4  0.781429       2.228198               -0.979172                   1.041107   \n",
       "\n",
       "   player.impact.depth  player.depth  previous.speed  previous.net.clearance  \\\n",
       "0             0.639689      0.552107        0.482919               -0.381947   \n",
       "1             0.456398     -0.057737        1.079780               -0.886178   \n",
       "2             0.376936      0.503636       -0.736615                1.155253   \n",
       "3            -0.251170     -0.463607        0.453810               -0.717097   \n",
       "4             1.220581      1.336942        1.149344               -0.580086   \n",
       "\n",
       "   previous.distance.from.sideline  previous.depth  opponent.depth  \\\n",
       "0                        -0.446773       -0.451524        0.181863   \n",
       "1                        -0.412138       -0.318133        0.486530   \n",
       "2                         1.393809       -0.676829        0.635215   \n",
       "3                         0.853026        0.693034       -0.005939   \n",
       "4                        -1.062001       -1.221139       -0.127378   \n",
       "\n",
       "   speed.difference  true.distance.from.sideline  true.distance.from.baseline  \n",
       "0          0.600468                     0.439649                    -1.122823  \n",
       "1         -0.405127                     0.438103                     0.725961  \n",
       "2          1.311106                    -0.505484                     0.286552  \n",
       "3          0.241626                     0.544824                    -0.419230  \n",
       "4         -0.162397                    -0.604864                    -2.882417  "
      ]
     },
     "execution_count": 165,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Model Building"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 172,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.9165\n",
      "Accuracy: 0.8655672163918041\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\preprocessing\\label.py:151: DeprecationWarning: The truth value of an empty array is ambiguous. Returning False, but in future this will result in an error. Use `array.size > 0` to check that an array is not empty.\n",
      "  if diff:\n",
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\preprocessing\\label.py:151: DeprecationWarning: The truth value of an empty array is ambiguous. Returning False, but in future this will result in an error. Use `array.size > 0` to check that an array is not empty.\n",
      "  if diff:\n"
     ]
    }
   ],
   "source": [
    "#model\n",
    "model_xgb= XGBClassifier(n_estimators=1000, random_state= 786, learning_rate= 0.1, \n",
    "                               subsample= 0.85, colsample_bytree= 0.8, \n",
    "                               eval_metric= 'mlogloss', max_depth= 6, \n",
    "                               min_child_weight= 1, gamma= 5)\n",
    "#fit model\n",
    "model_xgb.fit(X_train, y_train)\n",
    "\n",
    "#predict\n",
    "predicted_xgb = model_xgb.predict(X_train)\n",
    "preds_val_xgb=model_xgb.predict(X_test)\n",
    "\n",
    "#Accuracy\n",
    "print ('Accuracy:', accuracy_score(y_train, predicted_xgb))\n",
    "print ('Accuracy:', accuracy_score(y_test, preds_val_xgb))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Adaboost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "AdaBoostClassifier(algorithm='SAMME.R',\n",
       "          base_estimator=DecisionTreeClassifier(class_weight=None, criterion='gini', max_depth=6,\n",
       "            max_features=12, max_leaf_nodes=6, min_impurity_decrease=0.0,\n",
       "            min_impurity_split=None, min_samples_leaf=2,\n",
       "            min_samples_split=2, min_weight_fraction_leaf=0.1,\n",
       "            presort=False, random_state=123456, splitter='best'),\n",
       "          learning_rate=0.05, n_estimators=1000, random_state=None)"
      ]
     },
     "execution_count": 92,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Adaboost_model = AdaBoostClassifier(\n",
    "    DecisionTreeClassifier(max_depth=6,max_features=12, max_leaf_nodes=6,\n",
    "            min_samples_leaf=2, min_samples_split=2,\n",
    "            min_weight_fraction_leaf=0.1, presort=False, random_state=123456,\n",
    "            splitter='best'),\n",
    "    n_estimators=1000,\n",
    "    learning_rate=0.05)\n",
    "\n",
    "\n",
    "Adaboost_model.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Confusion matrix [[1046  278   34]\n",
      " [ 230 2322   70]\n",
      " [  15   37 1968]]\n",
      "Test Confusion matrix [[336 108  16]\n",
      " [107 743  29]\n",
      " [ 15  31 616]]\n",
      "Train Accuracy Score 0.8893333333333333\n",
      "Test Accuracy Score 0.8470764617691154\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "y_pred=Adaboost_model.predict(X_train)\n",
    "y_pred1=Adaboost_model.predict(X_test)\n",
    "\n",
    "print(\"Train Confusion matrix\",confusion_matrix(y_train,y_pred))\n",
    "print(\"Test Confusion matrix\",confusion_matrix(y_test,y_pred1))\n",
    "\n",
    "print(\"Train Accuracy Score\",accuracy_score(y_train,y_pred))\n",
    "print(\"Test Accuracy Score\",accuracy_score(y_test,y_pred1))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Grid search cv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.8668333333333333 {'learning_rate': 0.9, 'n_estimators': 1000}\n"
     ]
    }
   ],
   "source": [
    "#from sklearn.model_selection import GridSearchCV\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "param_grid = {'n_estimators':[100,150,200,500,1000], 'learning_rate' : [0.01,0.05,0.1,0.5,0.9]}\n",
    "clf = GridSearchCV(AdaBoostClassifier(\n",
    "    DecisionTreeClassifier(max_depth=6)), param_grid, n_jobs=15)\n",
    "clf.fit(X=X_train, y=y_train)\n",
    "ada_model = clf.best_estimator_\n",
    "print (clf.best_score_, clf.best_params_) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 253,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[1337   21    0]\n",
      " [  14 2607    1]\n",
      " [   0    0 2020]]\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.907"
      ]
     },
     "execution_count": 253,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pred_cv=ada_model.predict(X_train)\n",
    "print(confusion_matrix(y_train,y_pred_cv))\n",
    "accuracy_score(y_train,y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 254,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[316 127  17]\n",
      " [ 82 769  28]\n",
      " [  8  32 622]]\n",
      "0.8530734632683659\n"
     ]
    }
   ],
   "source": [
    "y_pred_test_cv=ada_model.predict(X_test)\n",
    "print(confusion_matrix(y_test,y_pred_test_cv))\n",
    "print(accuracy_score(y_test,y_pred_test_cv))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 167,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "features = X_train.columns\n",
    "importances = ada_model.feature_importances_\n",
    "indices = np.argsort(importances)\n",
    "\n",
    "plt.title('Feature Importances')\n",
    "plt.barh(range(len(indices)), importances[indices], color='b', align='center')\n",
    "plt.yticks(range(len(indices)), [features[i] for i in indices])\n",
    "plt.xlabel('Relative Importance')\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 168,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 6000 entries, 0 to 5999\n",
      "Data columns (total 14 columns):\n",
      "speed                              6000 non-null float64\n",
      "net.clearance                      6000 non-null float64\n",
      "distance.from.sideline             6000 non-null float64\n",
      "player.distance.travelled          6000 non-null float64\n",
      "player.impact.depth                6000 non-null float64\n",
      "player.depth                       6000 non-null float64\n",
      "previous.speed                     6000 non-null float64\n",
      "previous.net.clearance             6000 non-null float64\n",
      "previous.distance.from.sideline    6000 non-null float64\n",
      "previous.depth                     6000 non-null float64\n",
      "opponent.depth                     6000 non-null float64\n",
      "speed.difference                   6000 non-null float64\n",
      "true.distance.from.sideline        6000 non-null float64\n",
      "true.distance.from.baseline        6000 non-null float64\n",
      "dtypes: float64(14)\n",
      "memory usage: 656.3 KB\n"
     ]
    }
   ],
   "source": [
    "X_train.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 173,
   "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>0</th>\n",
       "      <th>1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.1157</td>\n",
       "      <td>net.clearance</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.0776</td>\n",
       "      <td>player.distance.travelled</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.0769</td>\n",
       "      <td>opponent.depth</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.0757</td>\n",
       "      <td>true.distance.from.baseline</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.0751</td>\n",
       "      <td>speed</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.0720</td>\n",
       "      <td>player.depth</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.0710</td>\n",
       "      <td>previous.distance.from.sideline</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.0704</td>\n",
       "      <td>previous.depth</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.0670</td>\n",
       "      <td>true.distance.from.sideline</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0.0666</td>\n",
       "      <td>previous.net.clearance</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0.0624</td>\n",
       "      <td>previous.speed</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0.0589</td>\n",
       "      <td>player.impact.depth</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>0.0579</td>\n",
       "      <td>speed.difference</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>0.0529</td>\n",
       "      <td>distance.from.sideline</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         0                                1\n",
       "0   0.1157                    net.clearance\n",
       "1   0.0776        player.distance.travelled\n",
       "2   0.0769                   opponent.depth\n",
       "3   0.0757      true.distance.from.baseline\n",
       "4   0.0751                            speed\n",
       "5   0.0720                     player.depth\n",
       "6   0.0710  previous.distance.from.sideline\n",
       "7   0.0704                   previous.depth\n",
       "8   0.0670      true.distance.from.sideline\n",
       "9   0.0666           previous.net.clearance\n",
       "10  0.0624                   previous.speed\n",
       "11  0.0589              player.impact.depth\n",
       "12  0.0579                 speed.difference\n",
       "13  0.0529           distance.from.sideline"
      ]
     },
     "execution_count": 173,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "impf=sorted(zip(map(lambda x: round(x, 4), ada_model.feature_importances_), X_train.columns), \n",
    "             reverse=True)\n",
    "impf = pd.DataFrame(impf, index=None)\n",
    "impf"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 185,
   "metadata": {},
   "outputs": [],
   "source": [
    "impf1 = impf[1]\n",
    "impf1=impf1[0:13]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 186,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0                       net.clearance\n",
       "1           player.distance.travelled\n",
       "2                      opponent.depth\n",
       "3         true.distance.from.baseline\n",
       "4                               speed\n",
       "5                        player.depth\n",
       "6     previous.distance.from.sideline\n",
       "7                      previous.depth\n",
       "8         true.distance.from.sideline\n",
       "9              previous.net.clearance\n",
       "10                     previous.speed\n",
       "11                player.impact.depth\n",
       "12                   speed.difference\n",
       "Name: 1, dtype: object"
      ]
     },
     "execution_count": 186,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "impf1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 207,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "AdaBoostClassifier(algorithm='SAMME.R',\n",
       "          base_estimator=DecisionTreeClassifier(class_weight=None, criterion='gini', max_depth=6,\n",
       "            max_features=8, max_leaf_nodes=10, min_impurity_decrease=0.0,\n",
       "            min_impurity_split=None, min_samples_leaf=2,\n",
       "            min_samples_split=2, min_weight_fraction_leaf=0.1,\n",
       "            presort=False, random_state=121212, splitter='best'),\n",
       "          learning_rate=0.5, n_estimators=1000, random_state=None)"
      ]
     },
     "execution_count": 207,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Adaboost_model1 = AdaBoostClassifier(\n",
    "    DecisionTreeClassifier(max_depth=6,max_features=8, max_leaf_nodes=10,\n",
    "            min_samples_leaf=2, min_samples_split=2,\n",
    "            min_weight_fraction_leaf=0.1, presort=False, random_state=121212,\n",
    "            splitter='best'),\n",
    "    n_estimators=1000,\n",
    "    learning_rate=0.5)\n",
    "\n",
    "\n",
    "Adaboost_model1.fit(X_train[impf1], y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 208,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Confusion matrix [[1114  237    7]\n",
      " [ 262 2324   36]\n",
      " [   0   16 2004]]\n",
      "Test Confusion matrix [[341 101  18]\n",
      " [125 727  27]\n",
      " [ 15  33 614]]\n",
      "Train Accuracy Score 0.907\n",
      "Test Accuracy Score 0.8405797101449275\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "y_pred=Adaboost_model1.predict(X_train[impf1])\n",
    "y_pred1=Adaboost_model1.predict(X_test[impf1])\n",
    "\n",
    "print(\"Train Confusion matrix\",confusion_matrix(y_train,y_pred))\n",
    "print(\"Test Confusion matrix\",confusion_matrix(y_test,y_pred1))\n",
    "\n",
    "print(\"Train Accuracy Score\",accuracy_score(y_train,y_pred))\n",
    "print(\"Test Accuracy Score\",accuracy_score(y_test,y_pred1))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Multinomial regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 226,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.738\n",
      "Accuracy: 0.7146426786606697\n"
     ]
    }
   ],
   "source": [
    "from sklearn.linear_model import LogisticRegression, LogisticRegressionCV\n",
    "mul_lr = LogisticRegression(multi_class='multinomial', solver='newton-cg', warm_start=True, random_state=12345, n_jobs=100).fit(X_train, y_train)\n",
    "predicted_mlr = mul_lr.predict(X_train)\n",
    "preds_val_mlr=mul_lr.predict(X_test)\n",
    "print ('Accuracy:', accuracy_score(y_train, predicted_mlr))\n",
    "print ('Accuracy:', accuracy_score(y_test, preds_val_mlr))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## SVM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 228,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.7448333333333333\n",
      "Accuracy: 0.7326336831584208\n"
     ]
    }
   ],
   "source": [
    "from sklearn.svm import SVC # \"Support Vector Classifier\" \n",
    "clf_svm = SVC(kernel='linear', gamma=5) \n",
    "\n",
    "# fitting x samples and y classes \n",
    "clf_svm.fit(X_train, y_train)\n",
    "predicted_svm = clf_svm.predict(X_train)\n",
    "preds_val_svm=clf_svm.predict(X_test)\n",
    "print ('Accuracy:', accuracy_score(y_train, predicted_svm))\n",
    "print ('Accuracy:', accuracy_score(y_test, preds_val_svm))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Build KNN Classifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 252,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.821\n",
      "0.7361319340329835\n"
     ]
    }
   ],
   "source": [
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "knn_model= KNeighborsClassifier(algorithm = 'kd_tree',n_neighbors=7, weights='uniform', n_jobs=100 , p=2)\n",
    "knn_model.fit(X_train,y_train)\n",
    "\n",
    "#predictions\n",
    "y_pred_train_knn = knn_model.predict(X_train)\n",
    "y_pred_test_knn = knn_model.predict(X_test)\n",
    "\n",
    "#Accuracy check\n",
    "print(accuracy_score(y_train,y_pred_train_knn))\n",
    "print(accuracy_score(y_test,y_pred_test_knn))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# GridSearch Cross validation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 248,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\cross_validation.py:41: DeprecationWarning: This module was deprecated in version 0.18 in favor of the model_selection module into which all the refactored classes and functions are moved. Also note that the interface of the new CV iterators are different from that of this module. This module will be removed in 0.20.\n",
      "  \"This module will be removed in 0.20.\", DeprecationWarning)\n",
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\grid_search.py:42: DeprecationWarning: This module was deprecated in version 0.18 in favor of the model_selection module into which all the refactored classes and functions are moved. This module will be removed in 0.20.\n",
      "  DeprecationWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 3 folds for each of 12 candidates, totalling 36 fits\n",
      "0.7563333333333333 {'n_neighbors': 12}\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=4)]: Done  36 out of  36 | elapsed:   12.8s finished\n"
     ]
    }
   ],
   "source": [
    "#from sklearn.model_selection import GridSearchCV\n",
    "from sklearn.grid_search import GridSearchCV\n",
    "parameters = {'n_neighbors':list(range(3,15))}\n",
    "clf = GridSearchCV(KNeighborsClassifier(), parameters, n_jobs=4,verbose=2)\n",
    "clf.fit(X=X_train, y=y_train)\n",
    "knn_model = clf.best_estimator_\n",
    "print (clf.best_score_, clf.best_params_) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 251,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.8075\n",
      "0.7526236881559221\n"
     ]
    }
   ],
   "source": [
    "y_pred_train_knn_cv=knn_model.predict(X_train)\n",
    "y_pred_test_knn_cv=knn_model.predict(X_test)\n",
    "\n",
    "print(accuracy_score(y_train,y_pred_train_knn_cv))\n",
    "print(accuracy_score(y_test,y_pred_test_knn_cv))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Stacking"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 368,
   "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>Gridcvknn</th>\n",
       "      <th>AdaBoost_cv</th>\n",
       "      <th>XGBoost</th>\n",
       "      <th>Multinom</th>\n",
       "      <th>svm</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>UE</td>\n",
       "      <td>FE</td>\n",
       "      <td>FE</td>\n",
       "      <td>UE</td>\n",
       "      <td>UE</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>UE</td>\n",
       "      <td>UE</td>\n",
       "      <td>UE</td>\n",
       "      <td>UE</td>\n",
       "      <td>UE</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>UE</td>\n",
       "      <td>UE</td>\n",
       "      <td>UE</td>\n",
       "      <td>UE</td>\n",
       "      <td>UE</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>UE</td>\n",
       "      <td>FE</td>\n",
       "      <td>FE</td>\n",
       "      <td>FE</td>\n",
       "      <td>FE</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>W</td>\n",
       "      <td>W</td>\n",
       "      <td>W</td>\n",
       "      <td>UE</td>\n",
       "      <td>UE</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  Gridcvknn AdaBoost_cv XGBoost Multinom svm\n",
       "0        UE          FE      FE       UE  UE\n",
       "1        UE          UE      UE       UE  UE\n",
       "2        UE          UE      UE       UE  UE\n",
       "3        UE          FE      FE       FE  FE\n",
       "4         W           W       W       UE  UE"
      ]
     },
     "execution_count": 368,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "base_predictions_train = pd.DataFrame( {'Gridcvknn': y_pred_train_knn_cv,\n",
    "     'AdaBoost_cv': y_pred_cv,\n",
    "      'XGBoost': predicted_xgb,\n",
    "     'Multinom': predicted_mlr,\n",
    "      'svm': predicted_svm,\n",
    "                                        \n",
    "    })\n",
    "base_predictions_train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 340,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(6000,)"
      ]
     },
     "execution_count": 340,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "temp= pd.DataFrame(y_pred_train_knn_cv)\n",
    "target.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 369,
   "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>Gridcvknn</th>\n",
       "      <th>AdaBoost_cv</th>\n",
       "      <th>XGBoost</th>\n",
       "      <th>Multinom</th>\n",
       "      <th>svm</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>UE</td>\n",
       "      <td>UE</td>\n",
       "      <td>UE</td>\n",
       "      <td>UE</td>\n",
       "      <td>UE</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>UE</td>\n",
       "      <td>UE</td>\n",
       "      <td>UE</td>\n",
       "      <td>UE</td>\n",
       "      <td>UE</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>W</td>\n",
       "      <td>W</td>\n",
       "      <td>W</td>\n",
       "      <td>UE</td>\n",
       "      <td>W</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>W</td>\n",
       "      <td>W</td>\n",
       "      <td>W</td>\n",
       "      <td>UE</td>\n",
       "      <td>UE</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>UE</td>\n",
       "      <td>UE</td>\n",
       "      <td>FE</td>\n",
       "      <td>FE</td>\n",
       "      <td>FE</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  Gridcvknn AdaBoost_cv XGBoost Multinom svm\n",
       "0        UE          UE      UE       UE  UE\n",
       "1        UE          UE      UE       UE  UE\n",
       "2         W           W       W       UE   W\n",
       "3         W           W       W       UE  UE\n",
       "4        UE          UE      FE       FE  FE"
      ]
     },
     "execution_count": 369,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "base_predictions_test = pd.DataFrame( {'Gridcvknn': y_pred_test_knn_cv,\n",
    "     'AdaBoost_cv': y_pred_test_cv,\n",
    "     'XGBoost': preds_val_xgb,\n",
    "     'Multinom': preds_val_mlr,\n",
    "     'svm': preds_val_svm,\n",
    "                                        \n",
    "    })\n",
    "base_predictions_test.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 370,
   "metadata": {},
   "outputs": [],
   "source": [
    "base_predictions_train[\"Gridcvknn\"] = base_predictions_train[\"Gridcvknn\"].replace({\"FE\":1,\"UE\":2, \"W\":3}) \n",
    "base_predictions_train[\"AdaBoost_cv\"] = base_predictions_train[\"AdaBoost_cv\"].replace({\"FE\":1,\"UE\":2, \"W\":3}) \n",
    "base_predictions_train[\"XGBoost\"] = base_predictions_train[\"XGBoost\"].replace({\"FE\":1,\"UE\":2, \"W\":3}) \n",
    "base_predictions_train[\"Multinom\"] = base_predictions_train[\"Multinom\"].replace({\"FE\":1,\"UE\":2, \"W\":3}) \n",
    "base_predictions_train[\"svm\"] = base_predictions_train[\"svm\"].replace({\"FE\":1,\"UE\":2, \"W\":3}) \n",
    "target = y_train.map({\"FE\":1,\"UE\":2, \"W\":3})#using gradient boosting method called Catboost which will accept catagorical variables\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 371,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "#base_predictions_test[\"Gridcvknn\",\"knn\",\"AdaBoost_cv\",\"XGBoost\",\"Multinom\",\"svm\"] =base_predictions_test[\"Gridcvknn\",\"knn\",\"AdaBoost_cv\",\"XGBoost\",\"Multinom\",\"svm\"].map({\"FE\":1,\"UE\":2, \"W\":3})\n",
    "base_predictions_test[\"Gridcvknn\"] = base_predictions_test[\"Gridcvknn\"].replace({\"FE\":1,\"UE\":2, \"W\":3}) \n",
    "base_predictions_test[\"AdaBoost_cv\"] = base_predictions_test[\"AdaBoost_cv\"].replace({\"FE\":1,\"UE\":2, \"W\":3}) \n",
    "base_predictions_test[\"XGBoost\"] = base_predictions_test[\"XGBoost\"].replace({\"FE\":1,\"UE\":2, \"W\":3}) \n",
    "base_predictions_test[\"Multinom\"] = base_predictions_test[\"Multinom\"].replace({\"FE\":1,\"UE\":2, \"W\":3}) \n",
    "base_predictions_test[\"svm\"] = base_predictions_test[\"svm\"].replace({\"FE\":1,\"UE\":2, \"W\":3}) \n",
    "test_target= y_test.map({\"FE\":1,\"UE\":2, \"W\":3})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 329,
   "metadata": {},
   "outputs": [],
   "source": [
    "cat_cols= [\"Gridcvknn\",\"AdaBoost_cv\",\"XGBoost\",\"Multinom\",\"svm\"]\n",
    "base_predictions_train[cat_cols]= base_predictions_train[cat_cols].astype('int64')\n",
    "base_predictions_test[cat_cols]= base_predictions_test[cat_cols].astype('int64')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Decision Tree"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 372,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.tree import DecisionTreeClassifier \n",
    "clf_entropy = DecisionTreeClassifier( \n",
    "            criterion = \"entropy\", random_state = 4444, \n",
    "            max_depth = 6, min_samples_leaf = 4) \n",
    "clf_gini = DecisionTreeClassifier(criterion = \"gini\", \n",
    "            random_state = 4444,max_depth=6, min_samples_leaf=4) \n",
    "  \n",
    "# Performing training \n",
    "clf_entropy.fit(base_predictions_train, target)\n",
    "clf_gini.fit(base_predictions_train, target)\n",
    "\n",
    "predicted_DT_gini_stack = clf_gini.predict(base_predictions_train)\n",
    "predicted_DT_entropy_stack = clf_entropy.predict(base_predictions_train)\n",
    "\n",
    "preds1_gini_stack=clf_gini.predict(base_predictions_test)\n",
    "preds1_entropy_stack=clf_entropy.predict(base_predictions_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 373,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy train gini: 0.994\n",
      "Accuracy test gini: 0.8530734632683659\n",
      "Accuracy train entropy: 0.994\n",
      "Accuracy test entropy: 0.8530734632683659\n"
     ]
    }
   ],
   "source": [
    "print ('Accuracy train gini:', accuracy_score(target, predicted_DT_gini_stack))\n",
    "print ('Accuracy test gini:', accuracy_score(test_target, preds1_gini_stack))\n",
    "\n",
    "print ('Accuracy train entropy:', accuracy_score(target, predicted_DT_entropy_stack))\n",
    "print ('Accuracy test entropy:', accuracy_score(test_target, preds1_entropy_stack))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Logistic Regression on Stacked DF"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 389,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.994\n",
      "Accuracy: 0.8530734632683659\n"
     ]
    }
   ],
   "source": [
    "lr = LogisticRegression(penalty='l2', max_iter=100, solver='newton-cg')\n",
    "lr.fit(base_predictions_train, target)\n",
    "predicted_lr = lr.predict(base_predictions_train)\n",
    "preds_val_lr=lr.predict(base_predictions_test)\n",
    "print ('Accuracy:', accuracy_score(target, predicted_lr))\n",
    "print ('Accuracy:', accuracy_score(test_target, preds_val_lr))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}