EPL_Prediction_Tuning.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import scipy.stats as scipy\n",
    "import matplotlib.pyplot as plt\n",
    "import urllib #download from football data\n",
    "\n",
    "from __future__ import division\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Load the data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "('./Data/England/E0_16.csv', <httplib.HTTPMessage instance at 0x7f796fc5f950>)"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "testfile = urllib.URLopener()\n",
    "testfile.retrieve(\"http://www.football-data.co.uk/mmz4281/1617/E0.csv\", \"./Data/England/E0_16.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "df = pd.read_csv(\"./Data/England/E0_16.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index([u'Div', u'Date', u'HomeTeam', u'AwayTeam', u'FTHG', u'FTAG', u'FTR',\n",
       "       u'HTHG', u'HTAG', u'HTR', u'Referee', u'HS', u'AS', u'HST', u'AST',\n",
       "       u'HF', u'AF', u'HC', u'AC', u'HY', u'AY', u'HR', u'AR', u'B365H',\n",
       "       u'B365D', u'B365A', u'BWH', u'BWD', u'BWA', u'IWH', u'IWD', u'IWA',\n",
       "       u'LBH', u'LBD', u'LBA', u'PSH', u'PSD', u'PSA', u'WHH', u'WHD', u'WHA',\n",
       "       u'VCH', u'VCD', u'VCA', u'Bb1X2', u'BbMxH', u'BbAvH', u'BbMxD',\n",
       "       u'BbAvD', u'BbMxA', u'BbAvA', u'BbOU', u'BbMx>2.5', u'BbAv>2.5',\n",
       "       u'BbMx<2.5', u'BbAv<2.5', u'BbAH', u'BbAHh', u'BbMxAHH', u'BbAvAHH',\n",
       "       u'BbMxAHA', u'BbAvAHA', u'PSCH', u'PSCD', u'PSCA'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.columns"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Cleaning and subsetting required data\n",
    "\n",
    "Remove betting data as we don't need it for prediction. However, they could be good indicators of a team's current form - should try it out later."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "res_16 = df.ix[:,:23]\n",
    "res_16 = res_16.drop(['Div','Date'],axis=1)\n",
    "# res_14 = df_14.ix[:,:23]\n",
    "# res_14 = res_14.drop(['Div','Date'],axis=1)\n",
    "table_features = df.ix[:,:7]\n",
    "table_features = table_features.drop(['FTHG','FTAG','Div','Date'],axis=1)\n",
    "bet_16 = df.ix[:,23:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "res_16.head()\n",
    "feature_table = df.ix[:,:23]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>HomeTeam</th>\n",
       "      <th>AwayTeam</th>\n",
       "      <th>FTHG</th>\n",
       "      <th>FTAG</th>\n",
       "      <th>FTR</th>\n",
       "      <th>HTHG</th>\n",
       "      <th>HTAG</th>\n",
       "      <th>HTR</th>\n",
       "      <th>Referee</th>\n",
       "      <th>HS</th>\n",
       "      <th>...</th>\n",
       "      <th>HST</th>\n",
       "      <th>AST</th>\n",
       "      <th>HF</th>\n",
       "      <th>AF</th>\n",
       "      <th>HC</th>\n",
       "      <th>AC</th>\n",
       "      <th>HY</th>\n",
       "      <th>AY</th>\n",
       "      <th>HR</th>\n",
       "      <th>AR</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>288</th>\n",
       "      <td>Man United</td>\n",
       "      <td>West Brom</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>D</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>D</td>\n",
       "      <td>M Dean</td>\n",
       "      <td>18</td>\n",
       "      <td>...</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>15</td>\n",
       "      <td>13</td>\n",
       "      <td>6</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>289</th>\n",
       "      <td>Southampton</td>\n",
       "      <td>Bournemouth</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>D</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>D</td>\n",
       "      <td>J Moss</td>\n",
       "      <td>14</td>\n",
       "      <td>...</td>\n",
       "      <td>4</td>\n",
       "      <td>3</td>\n",
       "      <td>14</td>\n",
       "      <td>8</td>\n",
       "      <td>6</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>290</th>\n",
       "      <td>Watford</td>\n",
       "      <td>Sunderland</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>H</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>D</td>\n",
       "      <td>L Probert</td>\n",
       "      <td>23</td>\n",
       "      <td>...</td>\n",
       "      <td>10</td>\n",
       "      <td>3</td>\n",
       "      <td>12</td>\n",
       "      <td>8</td>\n",
       "      <td>13</td>\n",
       "      <td>9</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>291</th>\n",
       "      <td>Arsenal</td>\n",
       "      <td>Man City</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>D</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>A</td>\n",
       "      <td>A Marriner</td>\n",
       "      <td>8</td>\n",
       "      <td>...</td>\n",
       "      <td>3</td>\n",
       "      <td>5</td>\n",
       "      <td>8</td>\n",
       "      <td>15</td>\n",
       "      <td>3</td>\n",
       "      <td>8</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>292</th>\n",
       "      <td>Swansea</td>\n",
       "      <td>Middlesbrough</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>D</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>D</td>\n",
       "      <td>R Madley</td>\n",
       "      <td>15</td>\n",
       "      <td>...</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>9</td>\n",
       "      <td>16</td>\n",
       "      <td>11</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 21 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        HomeTeam       AwayTeam  FTHG  FTAG FTR  HTHG  HTAG HTR     Referee  \\\n",
       "288   Man United      West Brom     0     0   D     0     0   D      M Dean   \n",
       "289  Southampton    Bournemouth     0     0   D     0     0   D      J Moss   \n",
       "290      Watford     Sunderland     1     0   H     0     0   D   L Probert   \n",
       "291      Arsenal       Man City     2     2   D     1     2   A  A Marriner   \n",
       "292      Swansea  Middlesbrough     0     0   D     0     0   D    R Madley   \n",
       "\n",
       "     HS ...  HST  AST  HF  AF  HC  AC  HY  AY  HR  AR  \n",
       "288  18 ...    3    1  15  13   6   1   3   3   0   0  \n",
       "289  14 ...    4    3  14   8   6   3   1   0   0   0  \n",
       "290  23 ...   10    3  12   8  13   9   1   3   0   0  \n",
       "291   8 ...    3    5   8  15   3   8   3   2   0   0  \n",
       "292  15 ...    3    1   9  16  11   3   2   3   0   0  \n",
       "\n",
       "[5 rows x 21 columns]"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "res_16.tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "#Team, Home Goals Score, Away Goals Score, Attack Strength, Home Goals Conceded, Away Goals Conceded, Defensive Strength\n",
    "table_16 = pd.DataFrame(columns=('Team','HGS','AGS','HAS','AAS','HGC','AGC','HDS','ADS',))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "293"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "res_16.shape[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Average number of goals at home 1.60750853242\n",
      "Average number of goals away 1.20136518771\n",
      "Average number of goals conceded at home 1.20136518771\n",
      "Average number of goals conceded away 1.60750853242\n"
     ]
    }
   ],
   "source": [
    "avg_home_scored_16 = res_16.FTHG.sum()*1.0 / res_16.shape[0]\n",
    "avg_away_scored_16 = res_16.FTAG.sum()*1.0 / res_16.shape[0]\n",
    "avg_home_conceded_16 = avg_away_scored_16\n",
    "avg_away_conceded_16 = avg_home_scored_16\n",
    "print \"Average number of goals at home\",avg_home_scored_16\n",
    "print \"Average number of goals away\", avg_away_scored_16\n",
    "print \"Average number of goals conceded at home\",avg_home_conceded_16\n",
    "print \"Average number of goals conceded away\",avg_away_conceded_16"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "res_home = res_16.groupby('HomeTeam')\n",
    "res_away = res_16.groupby('AwayTeam')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# list(res_home)[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Team</th>\n",
       "      <th>HGS</th>\n",
       "      <th>AGS</th>\n",
       "      <th>HAS</th>\n",
       "      <th>AAS</th>\n",
       "      <th>HGC</th>\n",
       "      <th>AGC</th>\n",
       "      <th>HDS</th>\n",
       "      <th>ADS</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Arsenal</td>\n",
       "      <td>28</td>\n",
       "      <td>30</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>15</td>\n",
       "      <td>21</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Bournemouth</td>\n",
       "      <td>26</td>\n",
       "      <td>16</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>23</td>\n",
       "      <td>31</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Burnley</td>\n",
       "      <td>22</td>\n",
       "      <td>9</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>14</td>\n",
       "      <td>30</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Chelsea</td>\n",
       "      <td>37</td>\n",
       "      <td>23</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>10</td>\n",
       "      <td>13</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Crystal Palace</td>\n",
       "      <td>15</td>\n",
       "      <td>23</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>20</td>\n",
       "      <td>27</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Everton</td>\n",
       "      <td>34</td>\n",
       "      <td>18</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>10</td>\n",
       "      <td>23</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Hull</td>\n",
       "      <td>21</td>\n",
       "      <td>7</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>24</td>\n",
       "      <td>35</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Leicester</td>\n",
       "      <td>24</td>\n",
       "      <td>11</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>18</td>\n",
       "      <td>29</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Liverpool</td>\n",
       "      <td>39</td>\n",
       "      <td>25</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>14</td>\n",
       "      <td>23</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Man City</td>\n",
       "      <td>24</td>\n",
       "      <td>32</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>14</td>\n",
       "      <td>18</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>Man United</td>\n",
       "      <td>20</td>\n",
       "      <td>22</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>10</td>\n",
       "      <td>13</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>Middlesbrough</td>\n",
       "      <td>12</td>\n",
       "      <td>8</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>17</td>\n",
       "      <td>16</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>Southampton</td>\n",
       "      <td>14</td>\n",
       "      <td>19</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>14</td>\n",
       "      <td>22</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>Stoke</td>\n",
       "      <td>19</td>\n",
       "      <td>14</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>17</td>\n",
       "      <td>27</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>Sunderland</td>\n",
       "      <td>14</td>\n",
       "      <td>10</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>26</td>\n",
       "      <td>25</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>Swansea</td>\n",
       "      <td>21</td>\n",
       "      <td>15</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>30</td>\n",
       "      <td>33</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>Tottenham</td>\n",
       "      <td>35</td>\n",
       "      <td>22</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>8</td>\n",
       "      <td>13</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>Watford</td>\n",
       "      <td>22</td>\n",
       "      <td>12</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>23</td>\n",
       "      <td>25</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>West Brom</td>\n",
       "      <td>27</td>\n",
       "      <td>12</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>18</td>\n",
       "      <td>20</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>West Ham</td>\n",
       "      <td>17</td>\n",
       "      <td>24</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>27</td>\n",
       "      <td>27</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Team  HGS  AGS  HAS  AAS  HGC  AGC  HDS  ADS\n",
       "0          Arsenal   28   30  NaN  NaN   15   21  NaN  NaN\n",
       "1      Bournemouth   26   16  NaN  NaN   23   31  NaN  NaN\n",
       "2          Burnley   22    9  NaN  NaN   14   30  NaN  NaN\n",
       "3          Chelsea   37   23  NaN  NaN   10   13  NaN  NaN\n",
       "4   Crystal Palace   15   23  NaN  NaN   20   27  NaN  NaN\n",
       "5          Everton   34   18  NaN  NaN   10   23  NaN  NaN\n",
       "6             Hull   21    7  NaN  NaN   24   35  NaN  NaN\n",
       "7        Leicester   24   11  NaN  NaN   18   29  NaN  NaN\n",
       "8        Liverpool   39   25  NaN  NaN   14   23  NaN  NaN\n",
       "9         Man City   24   32  NaN  NaN   14   18  NaN  NaN\n",
       "10      Man United   20   22  NaN  NaN   10   13  NaN  NaN\n",
       "11   Middlesbrough   12    8  NaN  NaN   17   16  NaN  NaN\n",
       "12     Southampton   14   19  NaN  NaN   14   22  NaN  NaN\n",
       "13           Stoke   19   14  NaN  NaN   17   27  NaN  NaN\n",
       "14      Sunderland   14   10  NaN  NaN   26   25  NaN  NaN\n",
       "15         Swansea   21   15  NaN  NaN   30   33  NaN  NaN\n",
       "16       Tottenham   35   22  NaN  NaN    8   13  NaN  NaN\n",
       "17         Watford   22   12  NaN  NaN   23   25  NaN  NaN\n",
       "18       West Brom   27   12  NaN  NaN   18   20  NaN  NaN\n",
       "19        West Ham   17   24  NaN  NaN   27   27  NaN  NaN"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "table_16.Team = res_home.HomeTeam.all().values\n",
    "table_16.HGS = res_home.FTHG.sum().values\n",
    "table_16.HGC = res_home.FTAG.sum().values\n",
    "table_16.AGS = res_away.FTAG.sum().values\n",
    "table_16.AGC = res_away.FTHG.sum().values\n",
    "table_16"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#Assuming number of home games = number of away games\n",
    "num_games = res_16.shape[0]/20"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Team</th>\n",
       "      <th>HGS</th>\n",
       "      <th>AGS</th>\n",
       "      <th>HAS</th>\n",
       "      <th>AAS</th>\n",
       "      <th>HGC</th>\n",
       "      <th>AGC</th>\n",
       "      <th>HDS</th>\n",
       "      <th>ADS</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Arsenal</td>\n",
       "      <td>28</td>\n",
       "      <td>30</td>\n",
       "      <td>1.188960</td>\n",
       "      <td>1.704545</td>\n",
       "      <td>15</td>\n",
       "      <td>21</td>\n",
       "      <td>0.852273</td>\n",
       "      <td>0.891720</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Bournemouth</td>\n",
       "      <td>26</td>\n",
       "      <td>16</td>\n",
       "      <td>1.104034</td>\n",
       "      <td>0.909091</td>\n",
       "      <td>23</td>\n",
       "      <td>31</td>\n",
       "      <td>1.306818</td>\n",
       "      <td>1.316348</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Burnley</td>\n",
       "      <td>22</td>\n",
       "      <td>9</td>\n",
       "      <td>0.934183</td>\n",
       "      <td>0.511364</td>\n",
       "      <td>14</td>\n",
       "      <td>30</td>\n",
       "      <td>0.795455</td>\n",
       "      <td>1.273885</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Chelsea</td>\n",
       "      <td>37</td>\n",
       "      <td>23</td>\n",
       "      <td>1.571125</td>\n",
       "      <td>1.306818</td>\n",
       "      <td>10</td>\n",
       "      <td>13</td>\n",
       "      <td>0.568182</td>\n",
       "      <td>0.552017</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Crystal Palace</td>\n",
       "      <td>15</td>\n",
       "      <td>23</td>\n",
       "      <td>0.636943</td>\n",
       "      <td>1.306818</td>\n",
       "      <td>20</td>\n",
       "      <td>27</td>\n",
       "      <td>1.136364</td>\n",
       "      <td>1.146497</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Everton</td>\n",
       "      <td>34</td>\n",
       "      <td>18</td>\n",
       "      <td>1.443737</td>\n",
       "      <td>1.022727</td>\n",
       "      <td>10</td>\n",
       "      <td>23</td>\n",
       "      <td>0.568182</td>\n",
       "      <td>0.976645</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Hull</td>\n",
       "      <td>21</td>\n",
       "      <td>7</td>\n",
       "      <td>0.891720</td>\n",
       "      <td>0.397727</td>\n",
       "      <td>24</td>\n",
       "      <td>35</td>\n",
       "      <td>1.363636</td>\n",
       "      <td>1.486200</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Leicester</td>\n",
       "      <td>24</td>\n",
       "      <td>11</td>\n",
       "      <td>1.019108</td>\n",
       "      <td>0.625000</td>\n",
       "      <td>18</td>\n",
       "      <td>29</td>\n",
       "      <td>1.022727</td>\n",
       "      <td>1.231423</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Liverpool</td>\n",
       "      <td>39</td>\n",
       "      <td>25</td>\n",
       "      <td>1.656051</td>\n",
       "      <td>1.420455</td>\n",
       "      <td>14</td>\n",
       "      <td>23</td>\n",
       "      <td>0.795455</td>\n",
       "      <td>0.976645</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Man City</td>\n",
       "      <td>24</td>\n",
       "      <td>32</td>\n",
       "      <td>1.019108</td>\n",
       "      <td>1.818182</td>\n",
       "      <td>14</td>\n",
       "      <td>18</td>\n",
       "      <td>0.795455</td>\n",
       "      <td>0.764331</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>Man United</td>\n",
       "      <td>20</td>\n",
       "      <td>22</td>\n",
       "      <td>0.849257</td>\n",
       "      <td>1.250000</td>\n",
       "      <td>10</td>\n",
       "      <td>13</td>\n",
       "      <td>0.568182</td>\n",
       "      <td>0.552017</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>Middlesbrough</td>\n",
       "      <td>12</td>\n",
       "      <td>8</td>\n",
       "      <td>0.509554</td>\n",
       "      <td>0.454545</td>\n",
       "      <td>17</td>\n",
       "      <td>16</td>\n",
       "      <td>0.965909</td>\n",
       "      <td>0.679406</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>Southampton</td>\n",
       "      <td>14</td>\n",
       "      <td>19</td>\n",
       "      <td>0.594480</td>\n",
       "      <td>1.079545</td>\n",
       "      <td>14</td>\n",
       "      <td>22</td>\n",
       "      <td>0.795455</td>\n",
       "      <td>0.934183</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>Stoke</td>\n",
       "      <td>19</td>\n",
       "      <td>14</td>\n",
       "      <td>0.806794</td>\n",
       "      <td>0.795455</td>\n",
       "      <td>17</td>\n",
       "      <td>27</td>\n",
       "      <td>0.965909</td>\n",
       "      <td>1.146497</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>Sunderland</td>\n",
       "      <td>14</td>\n",
       "      <td>10</td>\n",
       "      <td>0.594480</td>\n",
       "      <td>0.568182</td>\n",
       "      <td>26</td>\n",
       "      <td>25</td>\n",
       "      <td>1.477273</td>\n",
       "      <td>1.061571</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>Swansea</td>\n",
       "      <td>21</td>\n",
       "      <td>15</td>\n",
       "      <td>0.891720</td>\n",
       "      <td>0.852273</td>\n",
       "      <td>30</td>\n",
       "      <td>33</td>\n",
       "      <td>1.704545</td>\n",
       "      <td>1.401274</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>Tottenham</td>\n",
       "      <td>35</td>\n",
       "      <td>22</td>\n",
       "      <td>1.486200</td>\n",
       "      <td>1.250000</td>\n",
       "      <td>8</td>\n",
       "      <td>13</td>\n",
       "      <td>0.454545</td>\n",
       "      <td>0.552017</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>Watford</td>\n",
       "      <td>22</td>\n",
       "      <td>12</td>\n",
       "      <td>0.934183</td>\n",
       "      <td>0.681818</td>\n",
       "      <td>23</td>\n",
       "      <td>25</td>\n",
       "      <td>1.306818</td>\n",
       "      <td>1.061571</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>West Brom</td>\n",
       "      <td>27</td>\n",
       "      <td>12</td>\n",
       "      <td>1.146497</td>\n",
       "      <td>0.681818</td>\n",
       "      <td>18</td>\n",
       "      <td>20</td>\n",
       "      <td>1.022727</td>\n",
       "      <td>0.849257</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>West Ham</td>\n",
       "      <td>17</td>\n",
       "      <td>24</td>\n",
       "      <td>0.721868</td>\n",
       "      <td>1.363636</td>\n",
       "      <td>27</td>\n",
       "      <td>27</td>\n",
       "      <td>1.534091</td>\n",
       "      <td>1.146497</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Team  HGS  AGS       HAS       AAS  HGC  AGC       HDS       ADS\n",
       "0          Arsenal   28   30  1.188960  1.704545   15   21  0.852273  0.891720\n",
       "1      Bournemouth   26   16  1.104034  0.909091   23   31  1.306818  1.316348\n",
       "2          Burnley   22    9  0.934183  0.511364   14   30  0.795455  1.273885\n",
       "3          Chelsea   37   23  1.571125  1.306818   10   13  0.568182  0.552017\n",
       "4   Crystal Palace   15   23  0.636943  1.306818   20   27  1.136364  1.146497\n",
       "5          Everton   34   18  1.443737  1.022727   10   23  0.568182  0.976645\n",
       "6             Hull   21    7  0.891720  0.397727   24   35  1.363636  1.486200\n",
       "7        Leicester   24   11  1.019108  0.625000   18   29  1.022727  1.231423\n",
       "8        Liverpool   39   25  1.656051  1.420455   14   23  0.795455  0.976645\n",
       "9         Man City   24   32  1.019108  1.818182   14   18  0.795455  0.764331\n",
       "10      Man United   20   22  0.849257  1.250000   10   13  0.568182  0.552017\n",
       "11   Middlesbrough   12    8  0.509554  0.454545   17   16  0.965909  0.679406\n",
       "12     Southampton   14   19  0.594480  1.079545   14   22  0.795455  0.934183\n",
       "13           Stoke   19   14  0.806794  0.795455   17   27  0.965909  1.146497\n",
       "14      Sunderland   14   10  0.594480  0.568182   26   25  1.477273  1.061571\n",
       "15         Swansea   21   15  0.891720  0.852273   30   33  1.704545  1.401274\n",
       "16       Tottenham   35   22  1.486200  1.250000    8   13  0.454545  0.552017\n",
       "17         Watford   22   12  0.934183  0.681818   23   25  1.306818  1.061571\n",
       "18       West Brom   27   12  1.146497  0.681818   18   20  1.022727  0.849257\n",
       "19        West Ham   17   24  0.721868  1.363636   27   27  1.534091  1.146497"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "table_16.HAS = (table_16.HGS / num_games) / avg_home_scored_16\n",
    "table_16.AAS = (table_16.AGS / num_games) / avg_away_scored_16\n",
    "table_16.HDS = (table_16.HGC / num_games) / avg_home_conceded_16\n",
    "table_16.ADS = (table_16.AGC / num_games) / avg_away_conceded_16\n",
    "table_16"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Team</th>\n",
       "      <th>HGS</th>\n",
       "      <th>AGS</th>\n",
       "      <th>HAS</th>\n",
       "      <th>AAS</th>\n",
       "      <th>HGC</th>\n",
       "      <th>AGC</th>\n",
       "      <th>HDS</th>\n",
       "      <th>ADS</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>Swansea</td>\n",
       "      <td>21</td>\n",
       "      <td>15</td>\n",
       "      <td>0.89172</td>\n",
       "      <td>0.852273</td>\n",
       "      <td>30</td>\n",
       "      <td>33</td>\n",
       "      <td>1.704545</td>\n",
       "      <td>1.401274</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       Team  HGS  AGS      HAS       AAS  HGC  AGC       HDS       ADS\n",
       "15  Swansea   21   15  0.89172  0.852273   30   33  1.704545  1.401274"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "table_16[table_16.Team == \"Swansea\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "''' feature_table contains all the fixtures in the current season.\n",
    "ftr = full time result\n",
    "hst = home shots on target\n",
    "ast = away shots on target\n",
    "'''\n",
    "\n",
    "feature_table = feature_table[['HomeTeam','AwayTeam','FTR','HST','AST']]\n",
    "f_HAS = []\n",
    "f_HDS = []\n",
    "f_AAS = []\n",
    "f_ADS = []\n",
    "for index,row in feature_table.iterrows():\n",
    "    f_HAS.append(table_16[table_16['Team'] == row['HomeTeam']]['HAS'].values[0])\n",
    "    f_HDS.append(table_16[table_16['Team'] == row['HomeTeam']]['HDS'].values[0])\n",
    "    f_AAS.append(table_16[table_16['Team'] == row['AwayTeam']]['AAS'].values[0])\n",
    "    f_ADS.append(table_16[table_16['Team'] == row['AwayTeam']]['ADS'].values[0])\n",
    "    \n",
    "feature_table['HAS'] = f_HAS\n",
    "feature_table['HDS'] = f_HDS\n",
    "feature_table['AAS'] = f_AAS\n",
    "feature_table['ADS'] = f_ADS"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>HomeTeam</th>\n",
       "      <th>AwayTeam</th>\n",
       "      <th>FTR</th>\n",
       "      <th>HST</th>\n",
       "      <th>AST</th>\n",
       "      <th>HAS</th>\n",
       "      <th>HDS</th>\n",
       "      <th>AAS</th>\n",
       "      <th>ADS</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Burnley</td>\n",
       "      <td>Swansea</td>\n",
       "      <td>A</td>\n",
       "      <td>3</td>\n",
       "      <td>9</td>\n",
       "      <td>0.934183</td>\n",
       "      <td>0.795455</td>\n",
       "      <td>0.852273</td>\n",
       "      <td>1.401274</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Crystal Palace</td>\n",
       "      <td>West Brom</td>\n",
       "      <td>A</td>\n",
       "      <td>4</td>\n",
       "      <td>3</td>\n",
       "      <td>0.636943</td>\n",
       "      <td>1.136364</td>\n",
       "      <td>0.681818</td>\n",
       "      <td>0.849257</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Everton</td>\n",
       "      <td>Tottenham</td>\n",
       "      <td>D</td>\n",
       "      <td>6</td>\n",
       "      <td>4</td>\n",
       "      <td>1.443737</td>\n",
       "      <td>0.568182</td>\n",
       "      <td>1.250000</td>\n",
       "      <td>0.552017</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Hull</td>\n",
       "      <td>Leicester</td>\n",
       "      <td>H</td>\n",
       "      <td>5</td>\n",
       "      <td>5</td>\n",
       "      <td>0.891720</td>\n",
       "      <td>1.363636</td>\n",
       "      <td>0.625000</td>\n",
       "      <td>1.231423</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Man City</td>\n",
       "      <td>Sunderland</td>\n",
       "      <td>H</td>\n",
       "      <td>4</td>\n",
       "      <td>3</td>\n",
       "      <td>1.019108</td>\n",
       "      <td>0.795455</td>\n",
       "      <td>0.568182</td>\n",
       "      <td>1.061571</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Middlesbrough</td>\n",
       "      <td>Stoke</td>\n",
       "      <td>D</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>0.509554</td>\n",
       "      <td>0.965909</td>\n",
       "      <td>0.795455</td>\n",
       "      <td>1.146497</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Southampton</td>\n",
       "      <td>Watford</td>\n",
       "      <td>D</td>\n",
       "      <td>6</td>\n",
       "      <td>1</td>\n",
       "      <td>0.594480</td>\n",
       "      <td>0.795455</td>\n",
       "      <td>0.681818</td>\n",
       "      <td>1.061571</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Arsenal</td>\n",
       "      <td>Liverpool</td>\n",
       "      <td>A</td>\n",
       "      <td>5</td>\n",
       "      <td>7</td>\n",
       "      <td>1.188960</td>\n",
       "      <td>0.852273</td>\n",
       "      <td>1.420455</td>\n",
       "      <td>0.976645</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Bournemouth</td>\n",
       "      <td>Man United</td>\n",
       "      <td>A</td>\n",
       "      <td>3</td>\n",
       "      <td>7</td>\n",
       "      <td>1.104034</td>\n",
       "      <td>1.306818</td>\n",
       "      <td>1.250000</td>\n",
       "      <td>0.552017</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Chelsea</td>\n",
       "      <td>West Ham</td>\n",
       "      <td>H</td>\n",
       "      <td>6</td>\n",
       "      <td>3</td>\n",
       "      <td>1.571125</td>\n",
       "      <td>0.568182</td>\n",
       "      <td>1.363636</td>\n",
       "      <td>1.146497</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>Man United</td>\n",
       "      <td>Southampton</td>\n",
       "      <td>H</td>\n",
       "      <td>5</td>\n",
       "      <td>1</td>\n",
       "      <td>0.849257</td>\n",
       "      <td>0.568182</td>\n",
       "      <td>1.079545</td>\n",
       "      <td>0.934183</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>Burnley</td>\n",
       "      <td>Liverpool</td>\n",
       "      <td>H</td>\n",
       "      <td>2</td>\n",
       "      <td>5</td>\n",
       "      <td>0.934183</td>\n",
       "      <td>0.795455</td>\n",
       "      <td>1.420455</td>\n",
       "      <td>0.976645</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>Leicester</td>\n",
       "      <td>Arsenal</td>\n",
       "      <td>D</td>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "      <td>1.019108</td>\n",
       "      <td>1.022727</td>\n",
       "      <td>1.704545</td>\n",
       "      <td>0.891720</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>Stoke</td>\n",
       "      <td>Man City</td>\n",
       "      <td>A</td>\n",
       "      <td>3</td>\n",
       "      <td>6</td>\n",
       "      <td>0.806794</td>\n",
       "      <td>0.965909</td>\n",
       "      <td>1.818182</td>\n",
       "      <td>0.764331</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>Swansea</td>\n",
       "      <td>Hull</td>\n",
       "      <td>A</td>\n",
       "      <td>3</td>\n",
       "      <td>4</td>\n",
       "      <td>0.891720</td>\n",
       "      <td>1.704545</td>\n",
       "      <td>0.397727</td>\n",
       "      <td>1.486200</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>Tottenham</td>\n",
       "      <td>Crystal Palace</td>\n",
       "      <td>H</td>\n",
       "      <td>5</td>\n",
       "      <td>2</td>\n",
       "      <td>1.486200</td>\n",
       "      <td>0.454545</td>\n",
       "      <td>1.306818</td>\n",
       "      <td>1.146497</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>Watford</td>\n",
       "      <td>Chelsea</td>\n",
       "      <td>A</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>0.934183</td>\n",
       "      <td>1.306818</td>\n",
       "      <td>1.306818</td>\n",
       "      <td>0.552017</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>West Brom</td>\n",
       "      <td>Everton</td>\n",
       "      <td>A</td>\n",
       "      <td>4</td>\n",
       "      <td>7</td>\n",
       "      <td>1.146497</td>\n",
       "      <td>1.022727</td>\n",
       "      <td>1.022727</td>\n",
       "      <td>0.976645</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>Sunderland</td>\n",
       "      <td>Middlesbrough</td>\n",
       "      <td>A</td>\n",
       "      <td>5</td>\n",
       "      <td>3</td>\n",
       "      <td>0.594480</td>\n",
       "      <td>1.477273</td>\n",
       "      <td>0.454545</td>\n",
       "      <td>0.679406</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>West Ham</td>\n",
       "      <td>Bournemouth</td>\n",
       "      <td>H</td>\n",
       "      <td>2</td>\n",
       "      <td>5</td>\n",
       "      <td>0.721868</td>\n",
       "      <td>1.534091</td>\n",
       "      <td>0.909091</td>\n",
       "      <td>1.316348</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          HomeTeam        AwayTeam FTR  HST  AST       HAS       HDS  \\\n",
       "0          Burnley         Swansea   A    3    9  0.934183  0.795455   \n",
       "1   Crystal Palace       West Brom   A    4    3  0.636943  1.136364   \n",
       "2          Everton       Tottenham   D    6    4  1.443737  0.568182   \n",
       "3             Hull       Leicester   H    5    5  0.891720  1.363636   \n",
       "4         Man City      Sunderland   H    4    3  1.019108  0.795455   \n",
       "5    Middlesbrough           Stoke   D    2    1  0.509554  0.965909   \n",
       "6      Southampton         Watford   D    6    1  0.594480  0.795455   \n",
       "7          Arsenal       Liverpool   A    5    7  1.188960  0.852273   \n",
       "8      Bournemouth      Man United   A    3    7  1.104034  1.306818   \n",
       "9          Chelsea        West Ham   H    6    3  1.571125  0.568182   \n",
       "10      Man United     Southampton   H    5    1  0.849257  0.568182   \n",
       "11         Burnley       Liverpool   H    2    5  0.934183  0.795455   \n",
       "12       Leicester         Arsenal   D    1    4  1.019108  1.022727   \n",
       "13           Stoke        Man City   A    3    6  0.806794  0.965909   \n",
       "14         Swansea            Hull   A    3    4  0.891720  1.704545   \n",
       "15       Tottenham  Crystal Palace   H    5    2  1.486200  0.454545   \n",
       "16         Watford         Chelsea   A    2    4  0.934183  1.306818   \n",
       "17       West Brom         Everton   A    4    7  1.146497  1.022727   \n",
       "18      Sunderland   Middlesbrough   A    5    3  0.594480  1.477273   \n",
       "19        West Ham     Bournemouth   H    2    5  0.721868  1.534091   \n",
       "\n",
       "         AAS       ADS  \n",
       "0   0.852273  1.401274  \n",
       "1   0.681818  0.849257  \n",
       "2   1.250000  0.552017  \n",
       "3   0.625000  1.231423  \n",
       "4   0.568182  1.061571  \n",
       "5   0.795455  1.146497  \n",
       "6   0.681818  1.061571  \n",
       "7   1.420455  0.976645  \n",
       "8   1.250000  0.552017  \n",
       "9   1.363636  1.146497  \n",
       "10  1.079545  0.934183  \n",
       "11  1.420455  0.976645  \n",
       "12  1.704545  0.891720  \n",
       "13  1.818182  0.764331  \n",
       "14  0.397727  1.486200  \n",
       "15  1.306818  1.146497  \n",
       "16  1.306818  0.552017  \n",
       "17  1.022727  0.976645  \n",
       "18  0.454545  0.679406  \n",
       "19  0.909091  1.316348  "
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "feature_table.head(20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def transformResult(row):\n",
    "    '''Converts results (H,A or D) into numeric values'''\n",
    "    if(row.FTR == 'H'):\n",
    "        return 1\n",
    "    elif(row.FTR == 'A'):\n",
    "        return -1\n",
    "    else:\n",
    "        return 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "feature_table[\"Result\"] = feature_table.apply(lambda row: transformResult(row),axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>HomeTeam</th>\n",
       "      <th>AwayTeam</th>\n",
       "      <th>FTR</th>\n",
       "      <th>HST</th>\n",
       "      <th>AST</th>\n",
       "      <th>HAS</th>\n",
       "      <th>HDS</th>\n",
       "      <th>AAS</th>\n",
       "      <th>ADS</th>\n",
       "      <th>Result</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>283</th>\n",
       "      <td>Burnley</td>\n",
       "      <td>Tottenham</td>\n",
       "      <td>A</td>\n",
       "      <td>2</td>\n",
       "      <td>7</td>\n",
       "      <td>0.934183</td>\n",
       "      <td>0.795455</td>\n",
       "      <td>1.250000</td>\n",
       "      <td>0.552017</td>\n",
       "      <td>-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>284</th>\n",
       "      <td>Chelsea</td>\n",
       "      <td>Crystal Palace</td>\n",
       "      <td>A</td>\n",
       "      <td>11</td>\n",
       "      <td>3</td>\n",
       "      <td>1.571125</td>\n",
       "      <td>0.568182</td>\n",
       "      <td>1.306818</td>\n",
       "      <td>1.146497</td>\n",
       "      <td>-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>285</th>\n",
       "      <td>Hull</td>\n",
       "      <td>West Ham</td>\n",
       "      <td>H</td>\n",
       "      <td>3</td>\n",
       "      <td>4</td>\n",
       "      <td>0.891720</td>\n",
       "      <td>1.363636</td>\n",
       "      <td>1.363636</td>\n",
       "      <td>1.146497</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>286</th>\n",
       "      <td>Leicester</td>\n",
       "      <td>Stoke</td>\n",
       "      <td>H</td>\n",
       "      <td>10</td>\n",
       "      <td>2</td>\n",
       "      <td>1.019108</td>\n",
       "      <td>1.022727</td>\n",
       "      <td>0.795455</td>\n",
       "      <td>1.146497</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>287</th>\n",
       "      <td>Liverpool</td>\n",
       "      <td>Everton</td>\n",
       "      <td>H</td>\n",
       "      <td>6</td>\n",
       "      <td>4</td>\n",
       "      <td>1.656051</td>\n",
       "      <td>0.795455</td>\n",
       "      <td>1.022727</td>\n",
       "      <td>0.976645</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>288</th>\n",
       "      <td>Man United</td>\n",
       "      <td>West Brom</td>\n",
       "      <td>D</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>0.849257</td>\n",
       "      <td>0.568182</td>\n",
       "      <td>0.681818</td>\n",
       "      <td>0.849257</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>289</th>\n",
       "      <td>Southampton</td>\n",
       "      <td>Bournemouth</td>\n",
       "      <td>D</td>\n",
       "      <td>4</td>\n",
       "      <td>3</td>\n",
       "      <td>0.594480</td>\n",
       "      <td>0.795455</td>\n",
       "      <td>0.909091</td>\n",
       "      <td>1.316348</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>290</th>\n",
       "      <td>Watford</td>\n",
       "      <td>Sunderland</td>\n",
       "      <td>H</td>\n",
       "      <td>10</td>\n",
       "      <td>3</td>\n",
       "      <td>0.934183</td>\n",
       "      <td>1.306818</td>\n",
       "      <td>0.568182</td>\n",
       "      <td>1.061571</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>291</th>\n",
       "      <td>Arsenal</td>\n",
       "      <td>Man City</td>\n",
       "      <td>D</td>\n",
       "      <td>3</td>\n",
       "      <td>5</td>\n",
       "      <td>1.188960</td>\n",
       "      <td>0.852273</td>\n",
       "      <td>1.818182</td>\n",
       "      <td>0.764331</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>292</th>\n",
       "      <td>Swansea</td>\n",
       "      <td>Middlesbrough</td>\n",
       "      <td>D</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>0.891720</td>\n",
       "      <td>1.704545</td>\n",
       "      <td>0.454545</td>\n",
       "      <td>0.679406</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        HomeTeam        AwayTeam FTR  HST  AST       HAS       HDS       AAS  \\\n",
       "283      Burnley       Tottenham   A    2    7  0.934183  0.795455  1.250000   \n",
       "284      Chelsea  Crystal Palace   A   11    3  1.571125  0.568182  1.306818   \n",
       "285         Hull        West Ham   H    3    4  0.891720  1.363636  1.363636   \n",
       "286    Leicester           Stoke   H   10    2  1.019108  1.022727  0.795455   \n",
       "287    Liverpool         Everton   H    6    4  1.656051  0.795455  1.022727   \n",
       "288   Man United       West Brom   D    3    1  0.849257  0.568182  0.681818   \n",
       "289  Southampton     Bournemouth   D    4    3  0.594480  0.795455  0.909091   \n",
       "290      Watford      Sunderland   H   10    3  0.934183  1.306818  0.568182   \n",
       "291      Arsenal        Man City   D    3    5  1.188960  0.852273  1.818182   \n",
       "292      Swansea   Middlesbrough   D    3    1  0.891720  1.704545  0.454545   \n",
       "\n",
       "          ADS  Result  \n",
       "283  0.552017      -1  \n",
       "284  1.146497      -1  \n",
       "285  1.146497       1  \n",
       "286  1.146497       1  \n",
       "287  0.976645       1  \n",
       "288  0.849257       0  \n",
       "289  1.316348       0  \n",
       "290  1.061571       1  \n",
       "291  0.764331       0  \n",
       "292  0.679406       0  "
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "feature_table.tail(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "X_train = feature_table[['HST','AST','HAS','HDS','AAS','ADS',]]\n",
    "y_train = feature_table['Result']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.naive_bayes import MultinomialNB\n",
    "from xgboost import XGBClassifier\n",
    "from sklearn.metrics import accuracy_score\n",
    "from sklearn.model_selection import cross_val_score\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn.ensemble import voting_classifier\n",
    "from sklearn.svm import LinearSVC"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "clf1 = RandomForestClassifier()\n",
    "clf2 = MultinomialNB()\n",
    "clf3 = XGBClassifier()\n",
    "clf4 = LinearSVC()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[ 0.48387097  0.63333333  0.6         0.63333333  0.44827586  0.62068966\n",
      "  0.55172414  0.55172414  0.78571429  0.67857143]\n",
      "0.59872371418\n"
     ]
    }
   ],
   "source": [
    "y_pred = clf4.fit(X_train,y_train).predict(X_train)\n",
    "accuracy_score(y_pred,y_train)\n",
    "scores = cross_val_score(clf1, X_train, y_train, cv=10)\n",
    "print scores\n",
    "print scores.mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[ 0.41935484  0.56666667  0.5         0.63333333  0.48275862  0.51724138\n",
      "  0.51724138  0.5862069   0.60714286  0.57142857]\n",
      "0.540137454314\n"
     ]
    }
   ],
   "source": [
    "y_pred = clf1.fit(X_train,y_train).predict(X_train)\n",
    "accuracy_score(y_pred,y_train)\n",
    "scores = cross_val_score(clf1, X_train, y_train, cv=10)\n",
    "print scores\n",
    "print scores.mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[ 0.58064516  0.66666667  0.53333333  0.63333333  0.62068966  0.75862069\n",
      "  0.55172414  0.62068966  0.67857143  0.64285714]\n",
      "0.628713120398\n"
     ]
    }
   ],
   "source": [
    "y_pred = clf2.fit(X_train,y_train).predict(X_train)\n",
    "accuracy_score(y_pred,y_train)\n",
    "scores = cross_val_score(clf2, X_train, y_train, cv=10)\n",
    "print scores\n",
    "print scores.mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[ 0.41935484  0.56666667  0.63333333  0.6         0.5862069   0.5862069\n",
      "  0.51724138  0.65517241  0.67857143  0.71428571]\n",
      "0.595703956777\n"
     ]
    }
   ],
   "source": [
    "y_pred = clf3.fit(X_train, y_train).predict(X_train)\n",
    "accuracy_score(y_pred,y_train)\n",
    "scores = cross_val_score(clf3, X_train, y_train, cv=10)\n",
    "print scores\n",
    "print scores.mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Adding Recent Performance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "#What is this for?\n",
    "ht = res_16.loc[40].HomeTeam\n",
    "at = res_16.loc[40].AwayTeam"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "feat_table = res_16.sort_index(ascending=False)\n",
    "feat_table = feat_table[['HomeTeam','AwayTeam','FTR','FTHG','FTAG','HS','AS','HC','AC']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Adding next week fixtures\n",
    "new_fixtures = pd.DataFrame( [['Chelsea','Man City','D',0,0,0,0,0,0],\n",
    "                             ['Liverpool','Bournemouth','D',0,0,0,0,0,0],\n",
    "                             ['Southampton','Crystal Palace','D',0,0,0,0,0,0],\n",
    "                             ['Hull','Middlesbrough','D',0,0,0,0,0,0],\n",
    "                             ['Arsenal','West Ham','D',0,0,0,0,0,0],\n",
    "                             ['Swansea','Tottenham','D',0,0,0,0,0,0],\n",
    "                             ['Man United','Everton','D',0,0,0,0,0,0],\n",
    "                             ['Leicester','Sunderland','D',0,0,0,0,0,0],\n",
    "                             ['Burnley','Stoke','D',0,0,0,0,0,0],\n",
    "                             ['Watford','West Brom','D',0,0,0,0,0,0]],columns=feat_table.columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(303, 9)"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "new_feat_table = new_fixtures.append(feat_table,ignore_index=True)\n",
    "new_feat_table = new_feat_table.sort_index(ascending=False)\n",
    "new_feat_table = new_feat_table.reset_index().drop(['index'], axis=1)\n",
    "new_feat_table = new_feat_table.sort_index(ascending=False)\n",
    "# feat_table = n\n",
    "feat_table = new_feat_table\n",
    "feat_table.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#Adding k recent performance measures\n",
    "feat_table[\"pastHS\"] = 0.0\n",
    "feat_table[\"pastHC\"] = 0.0\n",
    "feat_table[\"pastAS\"] = 0.0\n",
    "feat_table[\"pastAC\"] = 0.0\n",
    "feat_table[\"pastHG\"] = 0.0\n",
    "feat_table[\"pastAG\"] = 0.0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>HomeTeam</th>\n",
       "      <th>AwayTeam</th>\n",
       "      <th>FTR</th>\n",
       "      <th>FTHG</th>\n",
       "      <th>FTAG</th>\n",
       "      <th>HS</th>\n",
       "      <th>AS</th>\n",
       "      <th>HC</th>\n",
       "      <th>AC</th>\n",
       "      <th>pastHS</th>\n",
       "      <th>pastHC</th>\n",
       "      <th>pastAS</th>\n",
       "      <th>pastAC</th>\n",
       "      <th>pastHG</th>\n",
       "      <th>pastAG</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>302</th>\n",
       "      <td>Chelsea</td>\n",
       "      <td>Man City</td>\n",
       "      <td>D</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>301</th>\n",
       "      <td>Liverpool</td>\n",
       "      <td>Bournemouth</td>\n",
       "      <td>D</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>300</th>\n",
       "      <td>Southampton</td>\n",
       "      <td>Crystal Palace</td>\n",
       "      <td>D</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>299</th>\n",
       "      <td>Hull</td>\n",
       "      <td>Middlesbrough</td>\n",
       "      <td>D</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>298</th>\n",
       "      <td>Arsenal</td>\n",
       "      <td>West Ham</td>\n",
       "      <td>D</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>297</th>\n",
       "      <td>Swansea</td>\n",
       "      <td>Tottenham</td>\n",
       "      <td>D</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>296</th>\n",
       "      <td>Man United</td>\n",
       "      <td>Everton</td>\n",
       "      <td>D</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>295</th>\n",
       "      <td>Leicester</td>\n",
       "      <td>Sunderland</td>\n",
       "      <td>D</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>294</th>\n",
       "      <td>Burnley</td>\n",
       "      <td>Stoke</td>\n",
       "      <td>D</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>293</th>\n",
       "      <td>Watford</td>\n",
       "      <td>West Brom</td>\n",
       "      <td>D</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>292</th>\n",
       "      <td>Swansea</td>\n",
       "      <td>Middlesbrough</td>\n",
       "      <td>D</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>15</td>\n",
       "      <td>8</td>\n",
       "      <td>11</td>\n",
       "      <td>3</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>291</th>\n",
       "      <td>Arsenal</td>\n",
       "      <td>Man City</td>\n",
       "      <td>D</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>14</td>\n",
       "      <td>3</td>\n",
       "      <td>8</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        HomeTeam        AwayTeam FTR  FTHG  FTAG  HS  AS  HC  AC  pastHS  \\\n",
       "302      Chelsea        Man City   D     0     0   0   0   0   0     0.0   \n",
       "301    Liverpool     Bournemouth   D     0     0   0   0   0   0     0.0   \n",
       "300  Southampton  Crystal Palace   D     0     0   0   0   0   0     0.0   \n",
       "299         Hull   Middlesbrough   D     0     0   0   0   0   0     0.0   \n",
       "298      Arsenal        West Ham   D     0     0   0   0   0   0     0.0   \n",
       "297      Swansea       Tottenham   D     0     0   0   0   0   0     0.0   \n",
       "296   Man United         Everton   D     0     0   0   0   0   0     0.0   \n",
       "295    Leicester      Sunderland   D     0     0   0   0   0   0     0.0   \n",
       "294      Burnley           Stoke   D     0     0   0   0   0   0     0.0   \n",
       "293      Watford       West Brom   D     0     0   0   0   0   0     0.0   \n",
       "292      Swansea   Middlesbrough   D     0     0  15   8  11   3     0.0   \n",
       "291      Arsenal        Man City   D     2     2   8  14   3   8     0.0   \n",
       "\n",
       "     pastHC  pastAS  pastAC  pastHG  pastAG  \n",
       "302     0.0     0.0     0.0     0.0     0.0  \n",
       "301     0.0     0.0     0.0     0.0     0.0  \n",
       "300     0.0     0.0     0.0     0.0     0.0  \n",
       "299     0.0     0.0     0.0     0.0     0.0  \n",
       "298     0.0     0.0     0.0     0.0     0.0  \n",
       "297     0.0     0.0     0.0     0.0     0.0  \n",
       "296     0.0     0.0     0.0     0.0     0.0  \n",
       "295     0.0     0.0     0.0     0.0     0.0  \n",
       "294     0.0     0.0     0.0     0.0     0.0  \n",
       "293     0.0     0.0     0.0     0.0     0.0  \n",
       "292     0.0     0.0     0.0     0.0     0.0  \n",
       "291     0.0     0.0     0.0     0.0     0.0  "
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "feat_table.head(12)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# for index,row in feat_table.iterrows():\n",
    "#     #print row['HomeTeam'], table_16[table_16['Team'] == row['HomeTeam']]['HAS'].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/sibi/acad/prog_tools/anaconda2/lib/python2.7/site-packages/ipykernel/__main__.py:7: UserWarning: Boolean Series key will be reindexed to match DataFrame index.\n",
      "/home/sibi/acad/prog_tools/anaconda2/lib/python2.7/site-packages/ipykernel/__main__.py:8: UserWarning: Boolean Series key will be reindexed to match DataFrame index.\n"
     ]
    }
   ],
   "source": [
    "# Adding k recent performance metrics. Change value of k.\n",
    "k = 4\n",
    "for i in range(feat_table.shape[0]-1,-1,-1):\n",
    "    row = feat_table.loc[i]\n",
    "    ht = row.HomeTeam\n",
    "    at = row.AwayTeam\n",
    "    ht_stats = feat_table.loc[i-1:-1][(feat_table.HomeTeam == ht) | (feat_table.AwayTeam == ht)].head(k)\n",
    "    at_stats = feat_table.loc[i-1:-1][(feat_table.HomeTeam == at) | (feat_table.AwayTeam == at)].head(k)\n",
    "\n",
    "    feat_table.set_value(i, 'pastHC', (ht_stats[ht_stats[\"AwayTeam\"] == ht].sum().HC + ht_stats[ht_stats[\"HomeTeam\"] == ht].sum().HC)/k)\n",
    "    feat_table.set_value(i, 'pastAC', (at_stats[at_stats[\"AwayTeam\"] == at].sum().HC + at_stats[at_stats[\"HomeTeam\"] == at].sum().HC)/k)\n",
    "    feat_table.set_value(i, 'pastHS', (ht_stats[ht_stats[\"AwayTeam\"] == ht].sum().HS + ht_stats[ht_stats[\"HomeTeam\"] == ht].sum().AS)/k)\n",
    "    feat_table.set_value(i, 'pastAS', (at_stats[at_stats[\"AwayTeam\"] == at].sum().HS + at_stats[at_stats[\"HomeTeam\"] == at].sum().AS)/k)\n",
    "    feat_table.set_value(i, 'pastHG', (ht_stats[ht_stats[\"AwayTeam\"] == ht].sum().FTAG + ht_stats[ht_stats[\"HomeTeam\"] == ht].sum().FTHG)/k)\n",
    "    feat_table.set_value(i, 'pastAG', (at_stats[at_stats[\"AwayTeam\"] == at].sum().FTAG + at_stats[at_stats[\"HomeTeam\"] == at].sum().FTHG)/k)\n",
    "\n",
    "f_HAS = []\n",
    "f_HDS = []\n",
    "f_AAS = []\n",
    "f_ADS = []\n",
    "for index,row in feat_table.iterrows():\n",
    "    #print row\n",
    "    f_HAS.append(table_16[table_16['Team'] == row['HomeTeam']]['HAS'].values[0])\n",
    "    f_HDS.append(table_16[table_16['Team'] == row['HomeTeam']]['HDS'].values[0])\n",
    "    f_AAS.append(table_16[table_16['Team'] == row['HomeTeam']]['AAS'].values[0])\n",
    "    f_ADS.append(table_16[table_16['Team'] == row['HomeTeam']]['ADS'].values[0])\n",
    "    \n",
    "feat_table['HAS'] = f_HAS\n",
    "feat_table['HDS'] = f_HDS\n",
    "feat_table['AAS'] = f_AAS\n",
    "feat_table['ADS'] = f_ADS"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "test_table = feat_table.drop(['FTHG','FTAG','HS','AS','HC','AC'],axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "test_table[\"Result\"] = test_table.apply(lambda row: transformResult(row),axis=1)\n",
    "test_table.sort_index(inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(303, 19)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "293"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# num_games decides the train-test split\n",
    "print feat_table.shape\n",
    "num_games = feat_table.shape[0]-10\n",
    "num_games"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "X_train = test_table[['pastHS','pastHC','pastAS','pastAC','pastHG','pastAG','HAS','HDS','AAS','ADS']].loc[0:num_games]\n",
    "y_train = test_table['Result'].loc[0:num_games]\n",
    "X_test = test_table[['pastHS','pastHC','pastAS','pastAC','pastHG','pastAG','HAS','HDS','AAS','ADS']].loc[num_games:]\n",
    "y_test = test_table['Result'].loc[num_games:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "### Try predictions here too."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "# Adding home advantage"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'test_table' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-40-1088e6108626>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m()\u001b[0m\n\u001b[1;32m----> 1\u001b[1;33m \u001b[0mtest_table\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m\"pastCornerDiff\"\u001b[0m\u001b[1;33m]\u001b[0m \u001b[1;33m=\u001b[0m \u001b[1;33m(\u001b[0m\u001b[0mtest_table\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m\"pastHC\"\u001b[0m\u001b[1;33m]\u001b[0m \u001b[1;33m-\u001b[0m \u001b[0mtest_table\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m\"pastAC\"\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m/\u001b[0m\u001b[0mk\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m      2\u001b[0m \u001b[0mtest_table\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m\"pastGoalDiff\"\u001b[0m\u001b[1;33m]\u001b[0m \u001b[1;33m=\u001b[0m \u001b[1;33m(\u001b[0m\u001b[0mtest_table\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m\"pastHG\"\u001b[0m\u001b[1;33m]\u001b[0m \u001b[1;33m-\u001b[0m \u001b[0mtest_table\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m\"pastAG\"\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m/\u001b[0m\u001b[0mk\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      3\u001b[0m \u001b[0mtest_table\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m\"pastShotsDiff\"\u001b[0m\u001b[1;33m]\u001b[0m \u001b[1;33m=\u001b[0m \u001b[1;33m(\u001b[0m\u001b[0mtest_table\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m\"pastHS\"\u001b[0m\u001b[1;33m]\u001b[0m \u001b[1;33m-\u001b[0m \u001b[0mtest_table\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m\"pastAG\"\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m/\u001b[0m\u001b[0mk\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;31mNameError\u001b[0m: name 'test_table' is not defined"
     ]
    }
   ],
   "source": [
    "test_table[\"pastCornerDiff\"] = (test_table[\"pastHC\"] - test_table[\"pastAC\"])/k\n",
    "test_table[\"pastGoalDiff\"] = (test_table[\"pastHG\"] - test_table[\"pastAG\"])/k\n",
    "test_table[\"pastShotsDiff\"] = (test_table[\"pastHS\"] - test_table[\"pastAG\"])/k"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "293\n"
     ]
    }
   ],
   "source": [
    "''' number of games to exclude in the training set for validation\n",
    "For example, if 240 games have been played, test_table has 250 fixtures - the last 10 being the ones that haven't\n",
    "been played. So, we set aside 20 fixtures from the training set(240 fixtures) for validation.''' \n",
    "num_games = feat_table.shape[0]-10\n",
    "print num_games\n",
    "v_split = 20\n",
    "n_games = num_games - v_split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'test_table' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-38-1ca51974c653>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m()\u001b[0m\n\u001b[1;32m----> 1\u001b[1;33m \u001b[0mtest_table\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mdrop\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m'pastHC'\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;34m'pastAS'\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;34m'pastAC'\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;34m'pastHG'\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;34m'pastAG'\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m,\u001b[0m\u001b[0maxis\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m1\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m      2\u001b[0m \u001b[0mX_train\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mtest_table\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m'pastCornerDiff'\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;34m'pastGoalDiff'\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;34m'pastShotsDiff'\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;34m'HAS'\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;34m'HDS'\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;34m'AAS'\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;34m'ADS'\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mloc\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;36m0\u001b[0m\u001b[1;33m:\u001b[0m\u001b[0mn_games\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      3\u001b[0m \u001b[0my_train\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mtest_table\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m'Result'\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mloc\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;36m0\u001b[0m\u001b[1;33m:\u001b[0m\u001b[0mn_games\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      4\u001b[0m \u001b[0mX_test\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mtest_table\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m'pastCornerDiff'\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;34m'pastGoalDiff'\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;34m'pastShotsDiff'\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;34m'HAS'\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;34m'HDS'\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;34m'AAS'\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;34m'ADS'\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mloc\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0mn_games\u001b[0m\u001b[1;33m:\u001b[0m\u001b[0mnum_games\u001b[0m\u001b[1;33m-\u001b[0m\u001b[1;36m1\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      5\u001b[0m \u001b[0my_test\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mtest_table\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m'Result'\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mloc\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0mn_games\u001b[0m\u001b[1;33m:\u001b[0m\u001b[0mnum_games\u001b[0m\u001b[1;33m-\u001b[0m\u001b[1;36m1\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;31mNameError\u001b[0m: name 'test_table' is not defined"
     ]
    }
   ],
   "source": [
    "test_table.drop(['pastHC','pastAS','pastAC','pastHG','pastAG'],axis=1)\n",
    "X_train = test_table[['pastCornerDiff','pastGoalDiff','pastShotsDiff','HAS','HDS','AAS','ADS']].loc[0:n_games]\n",
    "y_train = test_table['Result'].loc[0:n_games]\n",
    "X_test = test_table[['pastCornerDiff','pastGoalDiff','pastShotsDiff','HAS','HDS','AAS','ADS']].loc[n_games:num_games-1]\n",
    "y_test = test_table['Result'].loc[n_games:num_games-1]\n",
    "X_predict = test_table[['pastCornerDiff','pastGoalDiff','pastShotsDiff','HAS','HDS','AAS','ADS']].loc[num_games:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'X_predict' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-39-e75393e09975>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m()\u001b[0m\n\u001b[1;32m----> 1\u001b[1;33m \u001b[0mX_predict\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m: name 'X_predict' is not defined"
     ]
    }
   ],
   "source": [
    "X_predict"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# clf1.fit(X_train,y_train)\n",
    "# #clf2.fit(X_train,y_train)\n",
    "# clf3.fit(X_train,y_train)\n",
    "# clf4.fit(X_train,y_train)\n",
    "# clf5.fit(X_train,y_train)\n",
    "# clf6.fit(X_train,y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "#KNN\n",
    "plot_scores_knn = []\n",
    "for b in xrange(1,50):\n",
    "    clf_knn = KNeighborsClassifier(n_neighbors=b)\n",
    "    clf_knn.fit(X_train,y_train)\n",
    "    scores = accuracy_score(y_test,clf_knn.predict(X_test))\n",
    "    plot_scores_knn.append(scores)\n",
    "\n",
    "#XGBClassifier\n",
    "plot_scores_XGB = []\n",
    "for i in xrange(1,100):\n",
    "    clf_XGB = XGBClassifier(n_estimators=i,max_depth=100)\n",
    "    clf_XGB.fit(X_train, y_train)\n",
    "    scores = accuracy_score(y_test,clf_XGB.predict(X_test))\n",
    "    plot_scores_XGB.append(scores)\n",
    "    \n",
    "#Logistic Regression\n",
    "plot_scores_logreg= []\n",
    "cs = [0.01,0.02,0.1,0.5,1,3,4,5,10]\n",
    "for c in cs:\n",
    "    clf_logreg = LogisticRegression(C=c,solver='lbfgs',multi_class='ovr')\n",
    "    clf_logreg.fit(X_train, y_train)\n",
    "    scores = accuracy_score(y_test,clf_logreg.predict(X_test))\n",
    "    plot_scores_logreg.append(scores)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABHgAAAHyCAYAAACQ4XlaAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fda39dcfb10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = plt.figure(figsize = (16,7))\n",
    "# plt.style.use('ggplot')\n",
    "\n",
    "ax1 = fig.add_subplot(1,3,1)\n",
    "ax1.plot(xrange(1,50),plot_scores_knn);\n",
    "ax1.set_title(\"KNN - Accuracy vs N\")\n",
    "ax1.set_xticks(xrange(1,50,5));\n",
    "\n",
    "ax2 = fig.add_subplot(1,3,2)\n",
    "ax2.plot(xrange(1,100),plot_scores_XGB);\n",
    "ax2.set_xticks(xrange(1,100,6));\n",
    "ax2.set_title(\"XGB - Accuracy vs n_estimators\")\n",
    "\n",
    "ax3 = fig.add_subplot(1,3,3)\n",
    "ax3.plot(xrange(1,10),plot_scores_logreg);\n",
    "ax3.set_xticks(xrange(1,10));\n",
    "ax3.set_title(\"Logistic Regression - Accuracy vs C\")\n",
    "\n",
    "fig.tight_layout()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.6 12\n",
      "0.5 1\n",
      "0.5 3\n"
     ]
    }
   ],
   "source": [
    "max_knn_n = max(plot_scores_knn)\n",
    "max_knn_ind = plot_scores_knn.index(max_knn_n)\n",
    "\n",
    "max_XGB_e = max(plot_scores_XGB)\n",
    "max_XGB_ind = plot_scores_XGB.index(max_XGB_e)\n",
    "\n",
    "max_logreg_c = max(plot_scores_logreg)\n",
    "max_logreg_ind = plot_scores_logreg.index(max_logreg_c)\n",
    "\n",
    "print max_knn_n, max_knn_ind\n",
    "print max_XGB_e, max_XGB_ind\n",
    "print max_logreg_c, max_logreg_ind\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "clf_knn = KNeighborsClassifier(n_neighbors=max_knn_ind).fit(X_train,y_train)\n",
    "clf_XGB = XGBClassifier(n_estimators=max_XGB_ind).fit(X_train,y_train)\n",
    "clf_logreg = LogisticRegression(C=max_logreg_ind,solver='lbfgs',multi_class='ovr').fit(X_train,y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "y_pred_knn = clf_knn.predict(X_predict)\n",
    "y_pred_XGB = clf_XGB.predict(X_predict)\n",
    "y_pred_logreg = clf_logreg.predict(X_predict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "this_week = test_table[['HomeTeam','AwayTeam']].loc[num_games:]\n",
    "this_week['Result_knn']=y_pred_knn\n",
    "this_week['Result_XGB']=y_pred_XGB\n",
    "this_week['Result_logreg']=y_pred_logreg"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>HomeTeam</th>\n",
       "      <th>AwayTeam</th>\n",
       "      <th>Result_knn</th>\n",
       "      <th>Result_XGB</th>\n",
       "      <th>Result_logreg</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>293</th>\n",
       "      <td>Watford</td>\n",
       "      <td>West Brom</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>294</th>\n",
       "      <td>Burnley</td>\n",
       "      <td>Stoke</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>295</th>\n",
       "      <td>Leicester</td>\n",
       "      <td>Sunderland</td>\n",
       "      <td>-1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>296</th>\n",
       "      <td>Man United</td>\n",
       "      <td>Everton</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>297</th>\n",
       "      <td>Swansea</td>\n",
       "      <td>Tottenham</td>\n",
       "      <td>-1</td>\n",
       "      <td>0</td>\n",
       "      <td>-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>298</th>\n",
       "      <td>Arsenal</td>\n",
       "      <td>West Ham</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>299</th>\n",
       "      <td>Hull</td>\n",
       "      <td>Middlesbrough</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>300</th>\n",
       "      <td>Southampton</td>\n",
       "      <td>Crystal Palace</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>301</th>\n",
       "      <td>Liverpool</td>\n",
       "      <td>Bournemouth</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>302</th>\n",
       "      <td>Chelsea</td>\n",
       "      <td>Man City</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        HomeTeam        AwayTeam  Result_knn  Result_XGB  Result_logreg\n",
       "293      Watford       West Brom           1           1              1\n",
       "294      Burnley           Stoke           1           1              1\n",
       "295    Leicester      Sunderland          -1           1              1\n",
       "296   Man United         Everton           0           1              1\n",
       "297      Swansea       Tottenham          -1           0             -1\n",
       "298      Arsenal        West Ham           1           1              1\n",
       "299         Hull   Middlesbrough           0           0              1\n",
       "300  Southampton  Crystal Palace          -1          -1              1\n",
       "301    Liverpool     Bournemouth           1           1              1\n",
       "302      Chelsea        Man City           1           1              1"
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "this_week"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def transformResultBack(row,col_name):\n",
    "    if(row[col_name] == 1):\n",
    "        return 'H'\n",
    "    elif(row[col_name] == -1):\n",
    "        return 'A'\n",
    "    else:\n",
    "        return 'D'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "this_week[\"Res_knn\"] = this_week.apply(lambda row: transformResultBack(row,\"Result_knn\"),axis=1)\n",
    "this_week[\"Res_XGB\"] = this_week.apply(lambda row: transformResultBack(row,\"Result_XGB\"),axis=1)\n",
    "this_week[\"Res_logreg\"] = this_week.apply(lambda row: transformResultBack(row,\"Result_logreg\"),axis=1)\n",
    "\n",
    "this_week.drop([\"Result_knn\", \"Result_XGB\",\"Result_logreg\"],axis=1,inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.6 0.5 0.5\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>HomeTeam</th>\n",
       "      <th>AwayTeam</th>\n",
       "      <th>Res_knn</th>\n",
       "      <th>Res_XGB</th>\n",
       "      <th>Res_logreg</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>293</th>\n",
       "      <td>Watford</td>\n",
       "      <td>West Brom</td>\n",
       "      <td>H</td>\n",
       "      <td>H</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>294</th>\n",
       "      <td>Burnley</td>\n",
       "      <td>Stoke</td>\n",
       "      <td>H</td>\n",
       "      <td>H</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>295</th>\n",
       "      <td>Leicester</td>\n",
       "      <td>Sunderland</td>\n",
       "      <td>A</td>\n",
       "      <td>H</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>296</th>\n",
       "      <td>Man United</td>\n",
       "      <td>Everton</td>\n",
       "      <td>D</td>\n",
       "      <td>H</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>297</th>\n",
       "      <td>Swansea</td>\n",
       "      <td>Tottenham</td>\n",
       "      <td>A</td>\n",
       "      <td>D</td>\n",
       "      <td>A</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>298</th>\n",
       "      <td>Arsenal</td>\n",
       "      <td>West Ham</td>\n",
       "      <td>H</td>\n",
       "      <td>H</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>299</th>\n",
       "      <td>Hull</td>\n",
       "      <td>Middlesbrough</td>\n",
       "      <td>D</td>\n",
       "      <td>D</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>300</th>\n",
       "      <td>Southampton</td>\n",
       "      <td>Crystal Palace</td>\n",
       "      <td>A</td>\n",
       "      <td>A</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>301</th>\n",
       "      <td>Liverpool</td>\n",
       "      <td>Bournemouth</td>\n",
       "      <td>H</td>\n",
       "      <td>H</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>302</th>\n",
       "      <td>Chelsea</td>\n",
       "      <td>Man City</td>\n",
       "      <td>H</td>\n",
       "      <td>H</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        HomeTeam        AwayTeam Res_knn Res_XGB Res_logreg\n",
       "293      Watford       West Brom       H       H          H\n",
       "294      Burnley           Stoke       H       H          H\n",
       "295    Leicester      Sunderland       A       H          H\n",
       "296   Man United         Everton       D       H          H\n",
       "297      Swansea       Tottenham       A       D          A\n",
       "298      Arsenal        West Ham       H       H          H\n",
       "299         Hull   Middlesbrough       D       D          H\n",
       "300  Southampton  Crystal Palace       A       A          H\n",
       "301    Liverpool     Bournemouth       H       H          H\n",
       "302      Chelsea        Man City       H       H          H"
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print max_knn_n, max_XGB_e, max_logreg_c\n",
    "this_week"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python [Root]",
   "language": "python",
   "name": "Python [Root]"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}