2019 NCAA march madness competition.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# March madness tournament predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Import libraries needed\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.linear_model import LinearRegression\n",
    "from sklearn.svm import SVR\n",
    "from sklearn.svm import SVC\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "from sklearn.utils import shuffle\n",
    "import os"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Feature selection \n",
    "## stats.NBA.com emphasizes Four Factors as predictors for winning a NBA game   <br> [Why are the Four factos important?](https://stats.nba.com/help/faq/)\n",
    " (introduced by Oliver Dean)\n",
    "- __eFG%:__ effective field goal percentage\n",
    "- __TOVR:__ turnover ratio\n",
    "- __OR%:__ offensive rebound percentage\n",
    "- __FTAR:__ free throw attempt rate\n",
    "\n",
    "We have included 3 more factors:\n",
    "<font color='red'>+</font> __DR%:__ defensive rebound percentage\n",
    "<font color='red'>+</font> __TDR%:__ total defensive rebound percentage\n",
    "<font color='red'>+</font> __PIE:__ player impact estimate\n",
    "\n",
    "“High PIE% is highly correlated to winning. In fact, a team’s PIE rating and a team’s winning percentage correlate at an R square of .908 which indicates a \"strong\" correlation.” – [stats.nba.com](https://stats.nba.com/)\n",
    "\n",
    "\n",
    "#### Calculating the PIE\n",
    "What is PIE?\n",
    "It is a simple metric that gives an excellent indication of performance at both the team and player level. It’s a major improvement to our EFF Rating. Notably 2 things changed: (1) We included Personal Fouls, (2) We added a denominator. We feel the key here is the denominator because it acts as an \"automatic equalizer\". Using the denominator, we find there is no need to consider the \"PACE\" of the statistics that are being analyzed. In its simplest terms, PIE shows what % of game events did that player or team achieve. The stats being analyzed are your traditional basketball statistics (PTS, REB, AST, TOV, etc..) A team that achieves more than 50% is likely to be a winning team. A player that achieves more than 10% is likely to be better than the average player. A high PIE % is highly correlated to winning. In fact, a team’s PIE rating and a team’s winning percentage correlate at an R square of .908 which indicates a \"strong\" correlation. We’ve introduced this statistic because we feel it incorporates a bit of defense into the equation. When a team misses a shot, all 5 players on the other team’s PIE rating goes up. The formula for PIE can be found here in the Glossary.\n",
    "Source: [What is PIE?](https://stats.nba.com/help/faq/)\n",
    "<br>\n",
    "Formula: (PTS + FGM + FTM - FGA - FTA + DREB + (.5 * OREB) + AST + STL + (.5 * BLK) - PF - TO) / (GmPTS + GmFGM + GmFTM - GmFGA - GmFTA + GmDREB + (.5 * GmOREB) + GmAST + GmSTL + (.5 * GmBLK) - GmPF - GmTO)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Season</th>\n",
       "      <th>DayNum</th>\n",
       "      <th>WTeamID</th>\n",
       "      <th>WScore</th>\n",
       "      <th>LTeamID</th>\n",
       "      <th>LScore</th>\n",
       "      <th>WLoc</th>\n",
       "      <th>NumOT</th>\n",
       "      <th>WFGM</th>\n",
       "      <th>WFGA</th>\n",
       "      <th>...</th>\n",
       "      <th>DDR%</th>\n",
       "      <th>WTDR%</th>\n",
       "      <th>LTDR%</th>\n",
       "      <th>DTDR%</th>\n",
       "      <th>WTOR%</th>\n",
       "      <th>LTOR%</th>\n",
       "      <th>DTOR%</th>\n",
       "      <th>WPIE</th>\n",
       "      <th>LPIE</th>\n",
       "      <th>DPIE</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2016</td>\n",
       "      <td>11</td>\n",
       "      <td>1104</td>\n",
       "      <td>77</td>\n",
       "      <td>1244</td>\n",
       "      <td>64</td>\n",
       "      <td>H</td>\n",
       "      <td>0</td>\n",
       "      <td>29</td>\n",
       "      <td>57</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.029730</td>\n",
       "      <td>0.509091</td>\n",
       "      <td>0.490909</td>\n",
       "      <td>0.018182</td>\n",
       "      <td>0.454545</td>\n",
       "      <td>0.545455</td>\n",
       "      <td>-0.090909</td>\n",
       "      <td>0.429530</td>\n",
       "      <td>0.187919</td>\n",
       "      <td>0.241611</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2016</td>\n",
       "      <td>11</td>\n",
       "      <td>1105</td>\n",
       "      <td>68</td>\n",
       "      <td>1408</td>\n",
       "      <td>67</td>\n",
       "      <td>A</td>\n",
       "      <td>1</td>\n",
       "      <td>25</td>\n",
       "      <td>64</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.120843</td>\n",
       "      <td>0.433962</td>\n",
       "      <td>0.566038</td>\n",
       "      <td>-0.132075</td>\n",
       "      <td>0.437500</td>\n",
       "      <td>0.562500</td>\n",
       "      <td>-0.125000</td>\n",
       "      <td>0.280488</td>\n",
       "      <td>0.337398</td>\n",
       "      <td>-0.056911</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2016</td>\n",
       "      <td>11</td>\n",
       "      <td>1112</td>\n",
       "      <td>79</td>\n",
       "      <td>1334</td>\n",
       "      <td>61</td>\n",
       "      <td>H</td>\n",
       "      <td>0</td>\n",
       "      <td>24</td>\n",
       "      <td>61</td>\n",
       "      <td>...</td>\n",
       "      <td>0.296795</td>\n",
       "      <td>0.596491</td>\n",
       "      <td>0.403509</td>\n",
       "      <td>0.192982</td>\n",
       "      <td>0.772727</td>\n",
       "      <td>0.227273</td>\n",
       "      <td>0.545455</td>\n",
       "      <td>0.471264</td>\n",
       "      <td>0.176245</td>\n",
       "      <td>0.295019</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2016</td>\n",
       "      <td>11</td>\n",
       "      <td>1115</td>\n",
       "      <td>58</td>\n",
       "      <td>1370</td>\n",
       "      <td>56</td>\n",
       "      <td>A</td>\n",
       "      <td>0</td>\n",
       "      <td>20</td>\n",
       "      <td>55</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.051282</td>\n",
       "      <td>0.483333</td>\n",
       "      <td>0.516667</td>\n",
       "      <td>-0.033333</td>\n",
       "      <td>0.444444</td>\n",
       "      <td>0.555556</td>\n",
       "      <td>-0.111111</td>\n",
       "      <td>0.327731</td>\n",
       "      <td>0.277311</td>\n",
       "      <td>0.050420</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2016</td>\n",
       "      <td>11</td>\n",
       "      <td>1116</td>\n",
       "      <td>86</td>\n",
       "      <td>1380</td>\n",
       "      <td>68</td>\n",
       "      <td>H</td>\n",
       "      <td>0</td>\n",
       "      <td>32</td>\n",
       "      <td>66</td>\n",
       "      <td>...</td>\n",
       "      <td>0.235247</td>\n",
       "      <td>0.553191</td>\n",
       "      <td>0.446809</td>\n",
       "      <td>0.106383</td>\n",
       "      <td>0.708333</td>\n",
       "      <td>0.291667</td>\n",
       "      <td>0.416667</td>\n",
       "      <td>0.436533</td>\n",
       "      <td>0.210526</td>\n",
       "      <td>0.226006</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 62 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   Season  DayNum  WTeamID  WScore  LTeamID  LScore WLoc  NumOT  WFGM  WFGA  \\\n",
       "0    2016      11     1104      77     1244      64    H      0    29    57   \n",
       "1    2016      11     1105      68     1408      67    A      1    25    64   \n",
       "2    2016      11     1112      79     1334      61    H      0    24    61   \n",
       "3    2016      11     1115      58     1370      56    A      0    20    55   \n",
       "4    2016      11     1116      86     1380      68    H      0    32    66   \n",
       "\n",
       "     ...         DDR%     WTDR%     LTDR%     DTDR%     WTOR%     LTOR%  \\\n",
       "0    ...    -0.029730  0.509091  0.490909  0.018182  0.454545  0.545455   \n",
       "1    ...    -0.120843  0.433962  0.566038 -0.132075  0.437500  0.562500   \n",
       "2    ...     0.296795  0.596491  0.403509  0.192982  0.772727  0.227273   \n",
       "3    ...    -0.051282  0.483333  0.516667 -0.033333  0.444444  0.555556   \n",
       "4    ...     0.235247  0.553191  0.446809  0.106383  0.708333  0.291667   \n",
       "\n",
       "      DTOR%      WPIE      LPIE      DPIE  \n",
       "0 -0.090909  0.429530  0.187919  0.241611  \n",
       "1 -0.125000  0.280488  0.337398 -0.056911  \n",
       "2  0.545455  0.471264  0.176245  0.295019  \n",
       "3 -0.111111  0.327731  0.277311  0.050420  \n",
       "4  0.416667  0.436533  0.210526  0.226006  \n",
       "\n",
       "[5 rows x 62 columns]"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def load_and_construct_features_in_df(year):\n",
    "    \"\"\"\n",
    "    Input: year: Season year, e.g. 2017\n",
    "    Output: df: DataFrame with constructed features that can be used for training a ML model.\n",
    "    \"\"\"\n",
    "    # Read in data\n",
    "    #directory = 'G:/My Drive/Research Work/7 Kaggle/2019 NCAA/mens-machine-learning-competition-2019/DataFiles/'\n",
    "    directory = os.getcwd().replace('\\\\','/')\n",
    "    df_teams_seeds = pd.read_csv(directory + '/DataFiles/NCAATourneySeeds.csv').query('Season =='+str(year))\n",
    "\n",
    "    # Strip letters in seed column an return an integer\n",
    "    df_teams_seeds['Seed'] = df_teams_seeds['Seed'].apply(lambda x: int(x[1:3]))\n",
    "    #print df_teams_seeds.head()\n",
    "\n",
    "    df_tournament_results = pd.read_csv(directory + '/DataFiles/RegularSeasonDetailedResults.csv').query('Season=='+str(year))\n",
    "    #print df_tournament_results.head()\n",
    "\n",
    "    #Add 'WSeed','LSeed' feature to Detailed Tournament list by merging with renamed lists\n",
    "    # and calculate SeedDifference and add as a feature to df_final\n",
    "    df_renamed_win = df_teams_seeds.rename(columns={'TeamID':'WTeamID','Seed':'WSeed'})\n",
    "    df_renamed_loss = df_teams_seeds.rename(columns={'TeamID':'LTeamID','Seed':'LSeed'})\n",
    "    df_mixed = df_tournament_results.merge(df_renamed_win, how='left', on=['Season','WTeamID'])\n",
    "    df_final = df_mixed.merge(df_renamed_loss, how='left', on=['Season','LTeamID'])\n",
    "    \n",
    "    ### FEATURE CONSTRUCTION STARTS HERE ###\n",
    "    df_final['SeedDifference'] = df_final['WSeed'] - df_final['LSeed'] \n",
    "    df_final['ScoreDifference'] = df_final['WScore'] - df_final['LScore'] \n",
    "    #df_final.head()\n",
    "\n",
    "    # DTOVR (Difference in 'Turnover Ratio'): 100 * TOV / (FGA + (0.44 * FTA) + AST + TOV)\n",
    "    df_final['WTOVR'] = 100* df_final['WTO']/(df_final['WFGA'] + (0.44*df_final['WFTA']) + df_final['WAst'] + df_final['WTO'])\n",
    "    df_final['LTOVR'] = 100* df_final['LTO']/(df_final['LFGA'] + (0.44*df_final['LFTA']) + df_final['LAst'] + df_final['LTO'])\n",
    "    df_final['DTOVR'] = df_final['WTOVR'] - df_final['LTOVR']\n",
    "    # DeFG% (Difference in 'effective FieldGoal percentage'):  (FGM + 0.5 * 3PM) / FGA.\n",
    "    df_final['WeFG%'] = (df_final['WFGM'] + 0.5*df_final['WFGM3'])/(df_final['WFGA'] + df_final['WFGA3'])\n",
    "    df_final['LeFG%'] = (df_final['LFGM'] + 0.5*df_final['LFGM3'])/(df_final['LFGA'] + df_final['LFGA3'])\n",
    "    df_final['DeFG%'] = df_final['WeFG%'] - df_final['LeFG%']\n",
    "\n",
    "    #df_final['DeFG%'].describe()\n",
    "    # DFTAR (Difference in 'FreeThrow Attempt rate'):  FTA / FGA.\n",
    "    df_final['WFTAR'] = df_final['WFTA']/df_final['WFGA']\n",
    "    df_final['LFTAR'] = df_final['LFTA']/df_final['LFGA']\n",
    "    df_final['DFTAR'] = df_final['WFTAR'] - df_final['LFTAR']\n",
    "    # DOR% (Difference in 'Offensive rebound %'):  WOR / LDR+WOR. Notice that we divide by the \n",
    "    # a) number of OffensiveRebounds + number of OpponentsDefensiveRebounds (We look at one side of the board).\n",
    "    df_final['WOR%'] = df_final['WOR']/(df_final['LDR'] + df_final['WOR'])\n",
    "    df_final['LOR%'] = df_final['LOR']/(df_final['LOR'] + df_final['WDR'])\n",
    "    df_final['DOR%'] = df_final['WOR%'] - df_final['LOR%']\n",
    "    # b) DDR% (Difference in 'Defensive rebound %'):  WDR / LOR+WDR.\n",
    "    df_final['WDR%'] = df_final['WDR']/(df_final['LOR'] + df_final['WDR'])\n",
    "    df_final['LDR%'] = df_final['LDR']/(df_final['LDR'] + df_final['WOR'])\n",
    "    df_final['DDR%'] = df_final['WDR%'] - df_final['LDR%']\n",
    "    # c) DTDR% (Difference in 'Total Defensive rebound %'):  WDR / LDR+WDR.\n",
    "    df_final['WTDR%'] = df_final['WDR']/(df_final['LDR'] + df_final['WDR'])\n",
    "    df_final['LTDR%'] = df_final['LDR']/(df_final['LDR'] + df_final['WDR'])\n",
    "    df_final['DTDR%'] = df_final['WTDR%'] - df_final['LTDR%']\n",
    "    # d)\n",
    "    df_final['WTOR%'] = df_final['WOR']/(df_final['LOR'] + df_final['WOR'])\n",
    "    df_final['LTOR%'] = df_final['LOR']/(df_final['LOR'] + df_final['WOR'])\n",
    "    df_final['DTOR%'] = df_final['WTOR%'] - df_final['LTOR%']\n",
    "    # Adding PIE value to df_final\n",
    "    df_final['WPIE'] = (df_final['WScore'] + df_final['WFGM'] + df_final['WFTM'] - df_final['WFGA'] \n",
    "                        - df_final['WFTA'] + df_final['WDR'] + (0.5*df_final['WOR']) + df_final['WAst'] \n",
    "                        + df_final['WStl'] + (0.5*df_final['WBlk']) - df_final['WPF'] - df_final['WTO']\n",
    "                        )/(\n",
    "                        df_final['WScore'] + df_final['LScore'] + df_final['WFGM'] + df_final['LFGM'] \n",
    "                        + df_final['WFTM'] + df_final['LFTM'] - df_final['WFGA'] - df_final['LFGA'] \n",
    "                        + df_final['WDR'] + df_final['LDR'] + (0.5*df_final['WOR'] + 0.5*df_final['LOR']) \n",
    "                        + df_final['WAst'] + df_final['LAst'] + df_final['WStl'] + df_final['LStl'] \n",
    "                        + (0.5*df_final['WBlk'] + 0.5*df_final['LBlk']) - df_final['WPF'] - df_final['LPF']\n",
    "                        - df_final['WTO'] - df_final['WTO']\n",
    "                        )\n",
    "\n",
    "    df_final['LPIE'] = (df_final['LScore'] + df_final['LFGM'] + df_final['LFTM'] - df_final['LFGA']\n",
    "                     - df_final['LFTA'] + df_final['LDR'] + (0.5*df_final['LOR']) + df_final['LAst']\n",
    "                     + df_final['LStl'] + (0.5*df_final['LBlk']) - df_final['LPF'] - df_final['LTO']\n",
    "                     )/(\n",
    "                     df_final['WScore'] + df_final['LScore'] + df_final['WFGM'] + df_final['LFGM'] \n",
    "                     + df_final['WFTM'] + df_final['LFTM'] - df_final['WFGA'] - df_final['LFGA'] \n",
    "                     + df_final['WDR'] + df_final['LDR'] + (0.5*df_final['WOR'] + 0.5*df_final['LOR']) \n",
    "                     + df_final['WAst'] + df_final['LAst'] + df_final['WStl'] + df_final['LStl'] \n",
    "                     + (0.5*df_final['WBlk'] + 0.5*df_final['LBlk'])  - df_final['WPF'] - df_final['LPF']\n",
    "                     - df_final['WTO'] - df_final['WTO']\n",
    "                     )\n",
    "    df_final['DPIE'] = df_final['WPIE'] - df_final['LPIE']\n",
    "    return df_final\n",
    "df_final = load_and_construct_features_in_df(year=2016)\n",
    "df_final.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 7 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot Correlations of features\n",
    "import matplotlib.gridspec as gridspec\n",
    "features = ['DPIE','DeFG%','DTDR%','DFTAR','DTOVR','DOR%','DDR%']\n",
    "plt.figure(figsize=(10,10))\n",
    "gs = gridspec.GridSpec(3, 3,\n",
    "                       width_ratios=[1,1,1],\n",
    "                       height_ratios=[1,1,1],\n",
    "                       hspace=0.32, wspace=0.32)\n",
    "counter = 0\n",
    "plt.suptitle('Feature correlation (2016 season)', y=0.92, size= 18)\n",
    "for feature in features:\n",
    "    ax = plt.subplot(gs[counter])\n",
    "    plt.plot(df_final['ScoreDifference'], df_final[feature], marker='.', markerfacecolor = None, markeredgewidth = 0, linestyle='none', zorder= 1, label = feature[1:], alpha = 0.15)\n",
    "    df_grouped = df_final.groupby(['ScoreDifference']).describe().reset_index()\n",
    "    plt.plot(df_grouped['ScoreDifference'], df_grouped[feature]['mean'], marker='.', color='C1', linestyle='none', zorder=2, label='')\n",
    "    if counter in [4,5,6]:\n",
    "        plt.xlabel('Score Difference', size=13)\n",
    "    if counter in [0,3,6]:\n",
    "        plt.ylabel('Feature Difference', size=13)\n",
    "    plt.legend(fontsize=18)\n",
    "    plt.tick_params(labelsize=13)\n",
    "    counter += 1\n",
    "plt.savefig('FeatureCorrelation.png')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Check for collinearity"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[2.87110212 0.04421489 0.10289838 1.19634912 0.98090457 0.80453092]\n"
     ]
    }
   ],
   "source": [
    "X_train, Y_train = get_training_data(df_final, columns=['DPIE','DeFG%', 'DOR%','DTDR%','DFTAR','DTOVR'])\n",
    "corr = np.corrcoef(X_train, rowvar=0)  # correlation matrix\n",
    "w, v = np.linalg.eig(corr)        # eigen values & eigen vectors\n",
    "print w\n",
    "#print v\n",
    "# 'DDR%','DOR%' are collinear. 'DTOR%' and 'DTDR%' are almost collinear.\n",
    "# 'DPIE','DeFG%', 'DOR%','DTDR%','DFTAR','DTOVR'"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Result: __Defensive rebound %__ was found to be collinear with the __offensive rebound %__."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Get training data and setup ML algorithm\n",
    "What will be the input for my classifier? The PIE value shows the strongest correlation, hence associating the difference of PIE value with 0 or 1 could be one way or the values themself."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Extracting training data from the final dataframe of the whole 2017 season\n",
    "def get_training_data(df_final, columns):\n",
    "    \"\"\"\n",
    "    Input:  df_final: Dataframe with all features that can be used for estimating game results.\n",
    "            Columns: List of column labels. Column labels must be present in df_final.\n",
    "    Output: X_train, Y_train datasets, which can be directly fed into the LogReg model\n",
    "    \"\"\"\n",
    "    for feature in columns:\n",
    "        if feature not in df_final.columns:\n",
    "            raise Exception('The feature {}'.format(feature)+' is not in the DataFrame. All features must be labels of a column of the dataframe.')\n",
    "    \n",
    "    df_wins = pd.DataFrame()\n",
    "    df_losses = pd.DataFrame()\n",
    "    for feature in columns:\n",
    "        df_wins[str(feature)] = df_final[str(feature)]\n",
    "        df_losses[str(feature)] = -df_final[str(feature)]\n",
    "    df_wins['Result'] = 1\n",
    "    df_losses['Result'] = 0\n",
    "\n",
    "    training_data = pd.concat([df_losses, df_wins], axis=0)\n",
    "\n",
    "    if len(columns) == 1:\n",
    "        X_train = training_data[columns].values.reshape(-1,1)\n",
    "    else:\n",
    "        X_train = training_data[columns].values\n",
    "    Y_train = training_data.Result.values\n",
    "    return (X_train, Y_train)\n",
    "X_train, Y_train = get_training_data(df_final, columns=['DOR%', 'DPIE', 'DTOVR', 'DeFG%'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best log_loss: -0.1094, with best C: 108.436596869\n"
     ]
    }
   ],
   "source": [
    "def train_ML_model(X_train, Y_train, cv, model):\n",
    "    '''\n",
    "    Input:  X_train, Y_train: X and Y training data\n",
    "            cv: Cross-validation integer value\n",
    "            model: A string of either: 'LinReg','LogReg','DecTree','SVC'.\n",
    "    Output: GridSearchedObject, ClassifierObject, LogLoss, Regularization Parameter C\n",
    "    '''\n",
    "    if model not in ['LinReg','LogReg','DecTree','SVC']:\n",
    "        raise Exception('Model needs to be LinReg, LogReg, DecTree or SVC')\n",
    "\n",
    "    X_train, Y_train = shuffle(X_train, Y_train)\n",
    "    if model == 'LinReg':\n",
    "        linreg =LinearRegression()\n",
    "        paramsLinReg = {'fit_intercept':[True,False], 'normalize':[True,False], 'copy_X':[True, False]}\n",
    "        clf = GridSearchCV(linreg, paramsLinReg,cv=cv)\n",
    "        clf.fit(X_train, Y_train)\n",
    "        print('Best score: {:.4}'.format(clf.best_score_ ))\n",
    "        C_dummy = 0\n",
    "        return clf, linreg, clf.best_score_ , C_dummy\n",
    "    if model == 'LogReg':\n",
    "        logreg = LogisticRegression()\n",
    "        params = {'C': np.logspace(start=-15, stop=15, num=200)} #Parmeter of ML model\n",
    "        clf = GridSearchCV(logreg, params, scoring='neg_log_loss', refit=True, cv=cv)\n",
    "        clf.fit(X_train, Y_train)\n",
    "        print('Best log_loss: {:.4}, with best C: {}'.format(clf.best_score_, clf.best_params_['C']))\n",
    "        return clf, logreg, clf.best_score_, clf.best_params_['C']\n",
    "    if model == 'SVC':\n",
    "        SVCreg = SVC(probability=True)\n",
    "        params = {'C': np.logspace(start=-2, stop=3, num=10), 'kernel':['rbf']}\n",
    "        clf = GridSearchCV(SVCreg, params, scoring='neg_log_loss', refit=True, cv=cv)\n",
    "        clf.fit(X_train, Y_train)\n",
    "        print('Best log_loss: {:.4}, with best C: {}'.format(clf.best_score_, clf.best_params_['C']))\n",
    "        return clf, SVCreg, clf.best_score_, clf.best_params_['C']\n",
    "    if model == 'DecTree':\n",
    "        Tree = DecisionTreeClassifier()\n",
    "        params = {'max_depth': range(1,8),'min_samples_split': np.arange(2,7),\n",
    "                  'min_samples_leaf':np.arange(1,5), 'max_features': np.arange(1,X_train.shape[1]+1)} #Parmeter of ML model\n",
    "        clf = GridSearchCV(Tree, params, scoring='neg_log_loss', refit=True, cv=cv)\n",
    "        clf.fit(X_train, Y_train)\n",
    "        print('Best log_loss: {:.4}, with best max_depth: {}, best min_samples_split: {}, min_samples_leaf:{}, max_features_split:{}'.format(clf.best_score_, clf.best_params_['max_depth'], clf.best_params_['min_samples_split'], clf.best_params_['min_samples_leaf'], clf.best_params_['max_features']))\n",
    "        return clf, Tree, clf.best_score_, clf.best_params_['max_depth']\n",
    "clf, reg, score, C = train_ML_model(X_train, Y_train, cv=3, model='LogReg')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Make a prediction\n",
    "#### So now in order to make a prediction I have to take in e.g. the Difference of the Player Impact Estimate (DPIE) of the two teams playing:\n",
    "There are different options to do this:\n",
    "- Take the average over every season game (pretournament)\n",
    "- Take the average over the past 8 games\n",
    "- Just include previous games of these two team"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ID</th>\n",
       "      <th>OR%</th>\n",
       "      <th>PIE</th>\n",
       "      <th>TOVR</th>\n",
       "      <th>eFG%</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1112</td>\n",
       "      <td>0.345147</td>\n",
       "      <td>0.450142</td>\n",
       "      <td>13.2258</td>\n",
       "      <td>0.420496</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1114</td>\n",
       "      <td>0.267387</td>\n",
       "      <td>0.427308</td>\n",
       "      <td>12.0852</td>\n",
       "      <td>0.392038</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1122</td>\n",
       "      <td>0.313898</td>\n",
       "      <td>0.343985</td>\n",
       "      <td>15.4346</td>\n",
       "      <td>0.395195</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1124</td>\n",
       "      <td>0.397768</td>\n",
       "      <td>0.393503</td>\n",
       "      <td>13.2398</td>\n",
       "      <td>0.404594</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1138</td>\n",
       "      <td>0.317453</td>\n",
       "      <td>0.347124</td>\n",
       "      <td>14.3933</td>\n",
       "      <td>0.365046</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     ID       OR%       PIE     TOVR      eFG%\n",
       "0  1112  0.345147  0.450142  13.2258  0.420496\n",
       "1  1114  0.267387  0.427308  12.0852  0.392038\n",
       "2  1122  0.313898  0.343985  15.4346  0.395195\n",
       "3  1124  0.397768  0.393503  13.2398  0.404594\n",
       "4  1138  0.317453  0.347124  14.3933  0.365046"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# At first we will take the average value over every season game\n",
    "# Load the list that made it to the tournament\n",
    "def generate_tournament_team_list(year):\n",
    "    \"\"\"\n",
    "    List of teams will be generated that made it into the Tournament.\n",
    "    Output: List and copy of list.\n",
    "    \"\"\"\n",
    "    directory = os.getcwd().replace('\\\\','/')\n",
    "    df_teams = pd.read_csv(directory + '/DataFiles/NCAATourneyCompactResults.csv').query('Season=='+str(year))\n",
    "    df_teams = df_teams.sort_values(by=['WTeamID','LTeamID'])\n",
    "    df_teams\n",
    "\n",
    "    a = df_teams.drop_duplicates(subset=['WTeamID'])['WTeamID'] #List of winning teams\n",
    "    b = df_teams.drop_duplicates(subset=['LTeamID'])['LTeamID'] # List of loosing Teams\n",
    "    Team_IDs = a.append(b,ignore_index=True).drop_duplicates().values.tolist() # Add both lists and drop duplicates\n",
    "    Team_IDs_copy = a.append(b,ignore_index=True).drop_duplicates().values.tolist()# List of TeamIDs\n",
    "    Team_IDs_copy = sorted(Team_IDs_copy)\n",
    "    Team_IDs = sorted(Team_IDs)\n",
    "    #print 'Number of teams playing in '+str(year)+' Tournnament: '+ str(len(Team_IDs))+'/68 (64 + The first four).'\n",
    "    return Team_IDs, Team_IDs_copy\n",
    "Team_IDs, Team_IDs_copy = generate_tournament_team_list(year=2017)\n",
    "\n",
    "def generate_df_for_prediction(df_final, columns, year):\n",
    "    '''\n",
    "    Input:  df_dinal: DF that contains features which will be used as estimators. Require 'WTeamID' and 'LTeamID' features.\n",
    "            columns: List of features that will be averaged. Columns must be contained in df_final\n",
    "    Output: df with TeamID and averaged feature over the past season.\n",
    "    '''\n",
    "    for feature in columns:\n",
    "        if 'W'+feature not in df_final.columns or 'L'+feature not in df_final.columns:\n",
    "            raise Exception('The feature {}'.format(feature)+' is not in the DataFrame. All features must be labels of a column of the dataframe.')\n",
    "\n",
    "    Team_IDs, Team_IDs_copy = generate_tournament_team_list(year)\n",
    "            \n",
    "    df_for_prediction = pd.DataFrame(columns=['ID'] + columns)\n",
    "    counter = 0\n",
    "    for ID in Team_IDs:\n",
    "        df_for_prediction.at[counter, 'ID'] = str(ID)\n",
    "        for feature in columns:\n",
    "            a = df_final[df_final['WTeamID']==ID]['W'+str(feature)] # Gives a colum of a specific team with the feature called 'feature'.\n",
    "            b = df_final[df_final['LTeamID']==ID]['L'+str(feature)]\n",
    "            AVERAGE = a.append(b,ignore_index=True).mean()\n",
    "            df_for_prediction.at[counter, feature] = AVERAGE\n",
    "            #df_for_prediction = df_for_prediction.append({'ID':str(ID), str(feature):AVERAGE}, ignore_index=True)\n",
    "            del a\n",
    "            del b\n",
    "            del AVERAGE\n",
    "        counter += 1\n",
    "    return df_for_prediction\n",
    "df_for_prediction = generate_df_for_prediction(df_final, columns=['OR%', 'PIE', 'TOVR', 'eFG%'], year=2016)\n",
    "df_for_prediction.head()\n",
    "#df_for_prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ID</th>\n",
       "      <th>Season</th>\n",
       "      <th>Team1</th>\n",
       "      <th>Team2</th>\n",
       "      <th>eFG%1</th>\n",
       "      <th>eFG%2</th>\n",
       "      <th>DeFG%</th>\n",
       "      <th>OR%1</th>\n",
       "      <th>OR%2</th>\n",
       "      <th>DOR%</th>\n",
       "      <th>PIE1</th>\n",
       "      <th>PIE2</th>\n",
       "      <th>DPIE</th>\n",
       "      <th>TOVR1</th>\n",
       "      <th>TOVR2</th>\n",
       "      <th>DTOVR</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2016_1112_1114</td>\n",
       "      <td>2016</td>\n",
       "      <td>1112</td>\n",
       "      <td>1114</td>\n",
       "      <td>0.420496</td>\n",
       "      <td>0.392038</td>\n",
       "      <td>0.028458</td>\n",
       "      <td>0.345147</td>\n",
       "      <td>0.267387</td>\n",
       "      <td>0.077760</td>\n",
       "      <td>0.450142</td>\n",
       "      <td>0.427308</td>\n",
       "      <td>0.022834</td>\n",
       "      <td>13.225821</td>\n",
       "      <td>12.085212</td>\n",
       "      <td>1.140609</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2016_1112_1122</td>\n",
       "      <td>2016</td>\n",
       "      <td>1112</td>\n",
       "      <td>1122</td>\n",
       "      <td>0.420496</td>\n",
       "      <td>0.395195</td>\n",
       "      <td>0.025301</td>\n",
       "      <td>0.345147</td>\n",
       "      <td>0.313898</td>\n",
       "      <td>0.031249</td>\n",
       "      <td>0.450142</td>\n",
       "      <td>0.343985</td>\n",
       "      <td>0.106157</td>\n",
       "      <td>13.225821</td>\n",
       "      <td>15.434563</td>\n",
       "      <td>-2.208742</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2016_1112_1124</td>\n",
       "      <td>2016</td>\n",
       "      <td>1112</td>\n",
       "      <td>1124</td>\n",
       "      <td>0.420496</td>\n",
       "      <td>0.404594</td>\n",
       "      <td>0.015901</td>\n",
       "      <td>0.345147</td>\n",
       "      <td>0.397768</td>\n",
       "      <td>-0.052622</td>\n",
       "      <td>0.450142</td>\n",
       "      <td>0.393503</td>\n",
       "      <td>0.056639</td>\n",
       "      <td>13.225821</td>\n",
       "      <td>13.239782</td>\n",
       "      <td>-0.013961</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2016_1112_1138</td>\n",
       "      <td>2016</td>\n",
       "      <td>1112</td>\n",
       "      <td>1138</td>\n",
       "      <td>0.420496</td>\n",
       "      <td>0.365046</td>\n",
       "      <td>0.055450</td>\n",
       "      <td>0.345147</td>\n",
       "      <td>0.317453</td>\n",
       "      <td>0.027694</td>\n",
       "      <td>0.450142</td>\n",
       "      <td>0.347124</td>\n",
       "      <td>0.103018</td>\n",
       "      <td>13.225821</td>\n",
       "      <td>14.393276</td>\n",
       "      <td>-1.167455</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2016_1112_1139</td>\n",
       "      <td>2016</td>\n",
       "      <td>1112</td>\n",
       "      <td>1139</td>\n",
       "      <td>0.420496</td>\n",
       "      <td>0.403263</td>\n",
       "      <td>0.017233</td>\n",
       "      <td>0.345147</td>\n",
       "      <td>0.327521</td>\n",
       "      <td>0.017625</td>\n",
       "      <td>0.450142</td>\n",
       "      <td>0.406119</td>\n",
       "      <td>0.044023</td>\n",
       "      <td>13.225821</td>\n",
       "      <td>10.857235</td>\n",
       "      <td>2.368585</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               ID Season Team1 Team2     eFG%1     eFG%2     DeFG%      OR%1  \\\n",
       "0  2016_1112_1114   2016  1112  1114  0.420496  0.392038  0.028458  0.345147   \n",
       "1  2016_1112_1122   2016  1112  1122  0.420496  0.395195  0.025301  0.345147   \n",
       "2  2016_1112_1124   2016  1112  1124  0.420496  0.404594  0.015901  0.345147   \n",
       "3  2016_1112_1138   2016  1112  1138  0.420496  0.365046  0.055450  0.345147   \n",
       "4  2016_1112_1139   2016  1112  1139  0.420496  0.403263  0.017233  0.345147   \n",
       "\n",
       "       OR%2      DOR%      PIE1      PIE2      DPIE      TOVR1      TOVR2  \\\n",
       "0  0.267387  0.077760  0.450142  0.427308  0.022834  13.225821  12.085212   \n",
       "1  0.313898  0.031249  0.450142  0.343985  0.106157  13.225821  15.434563   \n",
       "2  0.397768 -0.052622  0.450142  0.393503  0.056639  13.225821  13.239782   \n",
       "3  0.317453  0.027694  0.450142  0.347124  0.103018  13.225821  14.393276   \n",
       "4  0.327521  0.017625  0.450142  0.406119  0.044023  13.225821  10.857235   \n",
       "\n",
       "      DTOVR  \n",
       "0  1.140609  \n",
       "1 -2.208742  \n",
       "2 -0.013961  \n",
       "3 -1.167455  \n",
       "4  2.368585  "
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Make a DF which provides Game_ID and the difference of an indicator/feature, e.g. 'DeFG%'\n",
    "def generate_prediction_features(df_for_prediction, columns, year):\n",
    "    \"\"\"\n",
    "    Input:  df_for_prediction: Obtained from generate_df_for_prediction(df_final, columns).\n",
    "            columns: Features containes in df_for_prediction e.g. ['eFG%','PIE']\n",
    "            year: Seaon year\n",
    "    Outout: X_Prediction columns\n",
    "    \"\"\"\n",
    "    for feature in columns:\n",
    "        if feature not in df_for_prediction.columns:\n",
    "            raise Exception('The feature {}'.format(feature)+' is not in the DataFrame. All features must be labels of a column of the dataframe.')\n",
    "    # ['eFG%','PIE']\n",
    "    col = ['ID','Season','Team1','Team2']\n",
    "    for feature in columns:\n",
    "        cols = col + [feature+'1'] + [feature+'2'] + ['D'+feature]\n",
    "    #cols = ['ID','Season','Team1','Team2','eFG%1','eFG%2','DeFG%']\n",
    "    df_prediction = pd.DataFrame(columns=cols)\n",
    "    #Tournament list need to be regenerate every time generater predictions is called because the teams are removed from the list\n",
    "    Team_IDs, Team_IDs_copy = generate_tournament_team_list(year=year)\n",
    "    print 'Number of teams playing in '+str(year)+' Tournnament: '+ str(len(Team_IDs))+'/68 (64 + The first four).'\n",
    "    dummy_a = Team_IDs\n",
    "    dummy_b = Team_IDs_copy\n",
    "    counter = 0\n",
    "    for ID1 in dummy_a:\n",
    "        #print ID\n",
    "        dummy_b.remove(ID1)\n",
    "        for ID2 in dummy_b:\n",
    "            Game_ID = str(year)+'_'+str(ID1)+'_'+str(ID2)\n",
    "            df_prediction.at[counter, 'ID'] = Game_ID # Add Game_ID value to row = counter and cell='ID' \n",
    "            df_prediction.at[counter, 'Season'] = year\n",
    "            df_prediction.at[counter, 'Team1'] = ID1\n",
    "            df_prediction.at[counter, 'Team2'] = ID2\n",
    "            for feature in columns:\n",
    "                dummy_var_a = df_for_prediction[df_for_prediction['ID'] == str(ID1)][str(feature)].values[0]\n",
    "                dummy_var_b = df_for_prediction[df_for_prediction['ID'] == str(ID2)][str(feature)].values[0]\n",
    "                df_prediction.at[counter, feature+'1'] = dummy_var_a\n",
    "                df_prediction.at[counter, feature+'2'] = dummy_var_b\n",
    "                df_prediction.at[counter, 'D'+feature] = dummy_var_a - dummy_var_b\n",
    "            counter += 1\n",
    "    for features in col: # convert values in features column to float values\n",
    "        df_prediction['D'+feature] = df_prediction['D'+feature].apply(lambda x: float(x))\n",
    "    #print len(df_prediction)\n",
    "    return df_prediction\n",
    "df_pred = generate_prediction_features(df_for_prediction, columns=['OR%', 'PIE', 'TOVR', 'eFG%'], year=2016)\n",
    "df_pred.head()\n",
    "#df_pred['DeFG%'].describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ID</th>\n",
       "      <th>Season</th>\n",
       "      <th>Team1</th>\n",
       "      <th>Team2</th>\n",
       "      <th>eFG%1</th>\n",
       "      <th>eFG%2</th>\n",
       "      <th>DeFG%</th>\n",
       "      <th>OR%1</th>\n",
       "      <th>OR%2</th>\n",
       "      <th>DOR%</th>\n",
       "      <th>PIE1</th>\n",
       "      <th>PIE2</th>\n",
       "      <th>DPIE</th>\n",
       "      <th>TOVR1</th>\n",
       "      <th>TOVR2</th>\n",
       "      <th>DTOVR</th>\n",
       "      <th>Pred</th>\n",
       "      <th>PredWL</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2016_1112_1114</td>\n",
       "      <td>2016</td>\n",
       "      <td>1112</td>\n",
       "      <td>1114</td>\n",
       "      <td>0.420496</td>\n",
       "      <td>0.392038</td>\n",
       "      <td>0.028458</td>\n",
       "      <td>0.345147</td>\n",
       "      <td>0.267387</td>\n",
       "      <td>0.077760</td>\n",
       "      <td>0.450142</td>\n",
       "      <td>0.427308</td>\n",
       "      <td>0.022834</td>\n",
       "      <td>13.225821</td>\n",
       "      <td>12.085212</td>\n",
       "      <td>1.140609</td>\n",
       "      <td>0.785366</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2016_1112_1122</td>\n",
       "      <td>2016</td>\n",
       "      <td>1112</td>\n",
       "      <td>1122</td>\n",
       "      <td>0.420496</td>\n",
       "      <td>0.395195</td>\n",
       "      <td>0.025301</td>\n",
       "      <td>0.345147</td>\n",
       "      <td>0.313898</td>\n",
       "      <td>0.031249</td>\n",
       "      <td>0.450142</td>\n",
       "      <td>0.343985</td>\n",
       "      <td>0.106157</td>\n",
       "      <td>13.225821</td>\n",
       "      <td>15.434563</td>\n",
       "      <td>-2.208742</td>\n",
       "      <td>0.990263</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2016_1112_1124</td>\n",
       "      <td>2016</td>\n",
       "      <td>1112</td>\n",
       "      <td>1124</td>\n",
       "      <td>0.420496</td>\n",
       "      <td>0.404594</td>\n",
       "      <td>0.015901</td>\n",
       "      <td>0.345147</td>\n",
       "      <td>0.397768</td>\n",
       "      <td>-0.052622</td>\n",
       "      <td>0.450142</td>\n",
       "      <td>0.393503</td>\n",
       "      <td>0.056639</td>\n",
       "      <td>13.225821</td>\n",
       "      <td>13.239782</td>\n",
       "      <td>-0.013961</td>\n",
       "      <td>0.860822</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2016_1112_1138</td>\n",
       "      <td>2016</td>\n",
       "      <td>1112</td>\n",
       "      <td>1138</td>\n",
       "      <td>0.420496</td>\n",
       "      <td>0.365046</td>\n",
       "      <td>0.055450</td>\n",
       "      <td>0.345147</td>\n",
       "      <td>0.317453</td>\n",
       "      <td>0.027694</td>\n",
       "      <td>0.450142</td>\n",
       "      <td>0.347124</td>\n",
       "      <td>0.103018</td>\n",
       "      <td>13.225821</td>\n",
       "      <td>14.393276</td>\n",
       "      <td>-1.167455</td>\n",
       "      <td>0.989895</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2016_1112_1139</td>\n",
       "      <td>2016</td>\n",
       "      <td>1112</td>\n",
       "      <td>1139</td>\n",
       "      <td>0.420496</td>\n",
       "      <td>0.403263</td>\n",
       "      <td>0.017233</td>\n",
       "      <td>0.345147</td>\n",
       "      <td>0.327521</td>\n",
       "      <td>0.017625</td>\n",
       "      <td>0.450142</td>\n",
       "      <td>0.406119</td>\n",
       "      <td>0.044023</td>\n",
       "      <td>13.225821</td>\n",
       "      <td>10.857235</td>\n",
       "      <td>2.368585</td>\n",
       "      <td>0.781976</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               ID Season Team1 Team2     eFG%1     eFG%2     DeFG%      OR%1  \\\n",
       "0  2016_1112_1114   2016  1112  1114  0.420496  0.392038  0.028458  0.345147   \n",
       "1  2016_1112_1122   2016  1112  1122  0.420496  0.395195  0.025301  0.345147   \n",
       "2  2016_1112_1124   2016  1112  1124  0.420496  0.404594  0.015901  0.345147   \n",
       "3  2016_1112_1138   2016  1112  1138  0.420496  0.365046  0.055450  0.345147   \n",
       "4  2016_1112_1139   2016  1112  1139  0.420496  0.403263  0.017233  0.345147   \n",
       "\n",
       "       OR%2      DOR%      PIE1      PIE2      DPIE      TOVR1      TOVR2  \\\n",
       "0  0.267387  0.077760  0.450142  0.427308  0.022834  13.225821  12.085212   \n",
       "1  0.313898  0.031249  0.450142  0.343985  0.106157  13.225821  15.434563   \n",
       "2  0.397768 -0.052622  0.450142  0.393503  0.056639  13.225821  13.239782   \n",
       "3  0.317453  0.027694  0.450142  0.347124  0.103018  13.225821  14.393276   \n",
       "4  0.327521  0.017625  0.450142  0.406119  0.044023  13.225821  10.857235   \n",
       "\n",
       "      DTOVR      Pred  PredWL  \n",
       "0  1.140609  0.785366       1  \n",
       "1 -2.208742  0.990263       1  \n",
       "2 -0.013961  0.860822       1  \n",
       "3 -1.167455  0.989895       1  \n",
       "4  2.368585  0.781976       1  "
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def get_predcitions(df_pred, columns, classifier):\n",
    "    '''Input:   df_pred: DF which containes predictors, e.g. 'DeFG%','DPIE' \n",
    "                columns: Colums with labels of predictors, e.g. ['DeFG%','DPIE']\n",
    "                classifier: A classifier from the machine learning section\n",
    "       Output: Dataframe with updated predictions based of ML classifier model.\n",
    "    '''\n",
    "    for feature in columns:\n",
    "        if feature not in df_pred.columns:\n",
    "            raise Exception('The feature {}'.format(feature)+' is not in the DataFrame. All features must be labels of a column of the dataframe.')\n",
    "    if len(columns) == 1:\n",
    "        X_pred_columns = df_pred[columns].values.reshape(-1,1)\n",
    "    elif len(columns) > 1:\n",
    "        X_pred_columns = df_pred[columns].values\n",
    "    elif len(columns) == 0:\n",
    "        return 'Columns should be a list of feature and at least contain 1 feature.'\n",
    "    \n",
    "    if type(reg) == type(LinearRegression()):\n",
    "        X_pred = classifier.predict(X_pred_columns)\n",
    "    else:\n",
    "        X_pred = classifier.predict_proba(X_pred_columns)[:,1]\n",
    "    #print X_pred\n",
    "    df_pred['Pred'] = X_pred\n",
    "    df_pred['PredWL'] = df_pred['Pred'].apply(lambda x: 0 if x <= 0.5 else 1)\n",
    "    return df_pred\n",
    "df_predicted_values = get_predcitions(df_pred, columns= ['DOR%', 'DPIE', 'DTOVR', 'DeFG%'], classifier=clf)\n",
    "df_predicted_values.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Validation: Check how many predictions are correct"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2016 Tournament games predicted correctly: 50 out of 67.0. (74.6%)\n"
     ]
    }
   ],
   "source": [
    "def check_correct_answers(df_predicted_values ,columns, year, print_df_head = False):\n",
    "    '''\n",
    "    Input: df_predicted_values: DataFrame with ID (year_team1_team2) and predictions (0 to 1) or predictions as win or loss (0 or 1)\n",
    "           columns: List of features, e.g. ['eFG%','PIE'].\n",
    "           year: Season year\n",
    "    Output: Printed number of games predicted correctly.\n",
    "    '''\n",
    "    directory = os.getcwd().replace('\\\\','/')\n",
    "    # Load in results\n",
    "    df_results = pd.read_csv(directory + '/DataFiles/NCAATourneyCompactResults.csv').query('Season=='+str(year))\n",
    "    df_results['Resu'] = 1\n",
    "    df_results.head()\n",
    "    df_results = df_results.sort_values(by=['WTeamID','LTeamID'])\n",
    "    df_results['ID'] = df_results['Season'].apply(lambda x: str(x))+'_'+df_results['WTeamID'].apply(lambda x: str(x))+'_'+df_results['LTeamID'].apply(lambda x: str(x))\n",
    "   \n",
    "    #iterate over df_results['ID'] and swap if team1<team2 (Game_ID = Year_Team1_Team2)\n",
    "    for index, row in df_results.iterrows():\n",
    "        team1, team2 = row['ID'].split('_')[1], row['ID'].split('_')[2]\n",
    "        if float(team1) > float(team2): #This is the case that gives NAN right now\n",
    "            #print row['ID'].split('_')\n",
    "            swapped_ID = row['ID'].split('_')[0]+'_'+team2+'_'+team1\n",
    "            df_results.at[index,'ID'] = swapped_ID\n",
    "            df_results.at[index,'Resu'] = 0\n",
    "    #df_results.head()\n",
    "\n",
    "    # Merge df_results with df_predicted_values\n",
    "    df_pred_result = df_results.merge(df_predicted_values, how='left', on=['ID'])\n",
    "    cols = ['Season_x','ID','WTeamID','Team1','WScore','LTeamID','LScore','Team2','Resu','PredWL','Pred']\n",
    "    for feature in columns:\n",
    "        cols = cols + [feature+'1', feature+'2', 'D'+feature] #'eFG%1','eFG%2','DeFG%','PIE1', 'PIE2', 'DPIE'\n",
    "    df_pred_result = df_pred_result[cols]\n",
    "    if print_df_head == True:\n",
    "        print df_pred_result.head()\n",
    "\n",
    "    # Number of correctly estimated games y\n",
    "    numOfWins = (df_pred_result['PredWL'] == df_pred_result['Resu']).sum()\n",
    "    gamesPlayed = float(len((df_pred_result['PredWL'])))\n",
    "    numOfWinsPerc = round(100*(numOfWins/gamesPlayed),1)\n",
    "    print str(year)+' Tournament games predicted correctly: '+ str(numOfWins)+' out of '+ str(gamesPlayed)+'. ('+str(numOfWinsPerc)+'%)'\n",
    "    return numOfWins, gamesPlayed, numOfWinsPerc\n",
    "    \n",
    "numOfWins, gamesPlayed, numOfWinsPerc = check_correct_answers(df_predicted_values, columns = ['eFG%','PIE'], year = 2016, print_df_head = False)\n",
    "#print numOfWins, gamesPlayed, numOfWinsPerc"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Wrapping all functions until now"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best log_loss: -0.109, with best C: 217.111794569\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 50 out of 67.0. (74.6%)\n"
     ]
    }
   ],
   "source": [
    "# Summarize all functions\n",
    "year=2016\n",
    "df_final = load_and_construct_features_in_df(year)\n",
    "#print df_final.columns\n",
    "\n",
    "def main(df_final, columns, year, model):\n",
    "    \"\"\"Input: Columns: List in the form of ['eFG%','PIE'] instead of ['DeFG%','DPIE']\n",
    "              model: model: A string of either: 'LinReg','LogReg','DecTree','SVC'.\n",
    "    \"\"\"\n",
    "    columns1 = columns # e.g., ['eFG%','PIE']\n",
    "    columns2 = map(lambda x: 'D'+x, columns) # e.g., ['DeFG%','DPIE']\n",
    "    \n",
    "    X_train, Y_train = get_training_data(df_final, columns = columns2)\n",
    "    clf, reg, score, C = train_ML_model(X_train, Y_train, cv=3, model=model)\n",
    "\n",
    "    df_for_prediction = generate_df_for_prediction(df_final, columns= columns1, year=year)\n",
    "    df_pred = generate_prediction_features(df_for_prediction, columns= columns1, year=year)\n",
    "    df_predicted_values = get_predcitions(df_pred, columns= columns2, classifier=clf)\n",
    "\n",
    "    numOfWins, gamesPlayed, numOfWinsPerc = check_correct_answers(df_predicted_values, columns = columns1, year = year, print_df_head = False)\n",
    "    return year, columns, numOfWins, gamesPlayed, numOfWinsPerc, score, C, clf\n",
    "columns=['PIE','OR%', 'TOVR']\n",
    "columns=['PIE']\n",
    "columns=['OR%', 'PIE', 'TOVR', 'eFG%']\n",
    "year, columns, numOfWins, gamesPlayed, numOfWinsPerc, score, C, clf = main(df_final, columns=columns, year=year, model='LogReg')\n",
    "# Note: Before 2011 only 65 Teams played in the NBA."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##  Train and evaluate different models for different features (fixed year)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 262,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best score: 0.6538\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 47 out of 67.0. (70.1%)\n",
      "Best score: 0.6237\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 41 out of 67.0. (61.2%)\n",
      "Best score: 0.6541\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 46 out of 67.0. (68.7%)\n",
      "Best score: 0.6536\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 47 out of 67.0. (70.1%)\n",
      "Best score: 0.637\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 48 out of 67.0. (71.6%)\n",
      "Best score: 0.6507\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 51 out of 67.0. (76.1%)\n",
      "Best score: 0.648\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 49 out of 67.0. (73.1%)\n",
      "Best score: 0.5988\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 36 out of 67.0. (53.7%)\n",
      "Best score: 0.6236\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 39 out of 67.0. (58.2%)\n",
      "Best score: 0.6522\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 49 out of 67.0. (73.1%)\n",
      "Best score: 0.5899\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 46 out of 67.0. (68.7%)\n",
      "Best score: 0.6362\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 51 out of 67.0. (76.1%)\n",
      "Best score: 0.6373\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 48 out of 67.0. (71.6%)\n",
      "Best score: 0.6082\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 45 out of 67.0. (67.2%)\n",
      "Best score: 0.6489\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 51 out of 67.0. (76.1%)\n",
      "Best score: 0.6507\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 51 out of 67.0. (76.1%)\n",
      "Best score: 0.6368\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 48 out of 67.0. (71.6%)\n",
      "Best score: 0.5119\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 34 out of 67.0. (50.7%)\n",
      "Best score: 0.6433\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 49 out of 67.0. (73.1%)\n",
      "Best score: 0.6445\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 49 out of 67.0. (73.1%)\n",
      "Best score: 0.6327\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 48 out of 67.0. (71.6%)\n",
      "Best score: 0.6442\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 49 out of 67.0. (73.1%)\n",
      "Best score: 0.5689\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 39 out of 67.0. (58.2%)\n",
      "Best score: 0.5284\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 39 out of 67.0. (58.2%)\n",
      "Best score: 0.5854\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 45 out of 67.0. (67.2%)\n",
      "Best score: 0.6358\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 50 out of 67.0. (74.6%)\n",
      "Best score: 0.296\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 29 out of 67.0. (43.3%)\n",
      "Best score: 0.5941\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 38 out of 67.0. (56.7%)\n",
      "Best score: 0.6485\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 50 out of 67.0. (74.6%)\n",
      "Best score: 0.5834\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 47 out of 67.0. (70.1%)\n",
      "Best score: 0.6351\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 49 out of 67.0. (73.1%)\n",
      "Best score: 0.6367\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 49 out of 67.0. (73.1%)\n",
      "Best score: 0.5087\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 35 out of 67.0. (52.2%)\n",
      "Best score: 0.4798\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 38 out of 67.0. (56.7%)\n",
      "Best score: 0.6431\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 49 out of 67.0. (73.1%)\n",
      "Best score: 0.452\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 41 out of 67.0. (61.2%)\n",
      "Best score: 0.632\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 46 out of 67.0. (68.7%)\n",
      "Best score: 0.6328\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 46 out of 67.0. (68.7%)\n",
      "Best score: 0.3779\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 38 out of 67.0. (56.7%)\n",
      "Best score: 0.6429\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 49 out of 67.0. (73.1%)\n",
      "Best score: 0.6435\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 50 out of 67.0. (74.6%)\n",
      "Best score: 0.6325\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 45 out of 67.0. (67.2%)\n",
      "Best score: 0.4966\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 41 out of 67.0. (61.2%)\n",
      "Best score: 0.2352\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 29 out of 67.0. (43.3%)\n",
      "Best score: 0.1511\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 37 out of 67.0. (55.2%)\n",
      "Best score: 0.5614\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 44 out of 67.0. (65.7%)\n",
      "Best score: 0.635\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 47 out of 67.0. (70.1%)\n",
      "Best score: 0.4734\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 36 out of 67.0. (53.7%)\n",
      "Best score: 0.4518\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 40 out of 67.0. (59.7%)\n",
      "Best score: 0.4376\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 45 out of 67.0. (67.2%)\n",
      "Best score: 0.6319\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 45 out of 67.0. (67.2%)\n",
      "Best score: 0.2002\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2016 Tournament games predicted correctly: 25 out of 67.0. (37.3%)\n",
      "Best score: 0.378\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 38 out of 67.0. (56.7%)\n",
      "Best score: 0.6426\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 50 out of 67.0. (74.6%)\n",
      "Best score: 0.3471\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 43 out of 67.0. (64.2%)\n",
      "Best score: 0.6323\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 45 out of 67.0. (67.2%)\n",
      "Best score: 0.6323\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 45 out of 67.0. (67.2%)\n",
      "Best score: 0.08258\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 41 out of 67.0. (61.2%)\n",
      "Best score: 0.4142\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 39 out of 67.0. (58.2%)\n",
      "Best score: 0.1319\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 23 out of 67.0. (34.3%)\n",
      "Best score: 0.07574\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 30 out of 67.0. (44.8%)\n",
      "Best score: 0.3454\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 43 out of 67.0. (64.2%)\n",
      "Best score: 0.6324\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 45 out of 67.0. (67.2%)\n"
     ]
    }
   ],
   "source": [
    "#Get model parameters for one year but different features\n",
    "import itertools\n",
    "\n",
    "df_stats = pd.DataFrame(columns=['Season','Columns','WinsPred','GamesPlayed','WinsPred%','Score','C','Coef weight'])\n",
    "counter = 0\n",
    "for year in range(2016,2017): # for one year\n",
    "    df_final = load_and_construct_features_in_df(year)\n",
    "    model = 'LinReg' # model: A string of either: 'LinReg','LogReg','DecTree','SVC'\n",
    "    columns_main = ['TOVR','eFG%','FTAR','OR%','TDR%','PIE']\n",
    "    # The combination of column elements is done here, with 7,6,5,4 elements in the column vector.\n",
    "    for k in [6,5,4,3,2,1]: #[6,5,4,3,2,1]\n",
    "        list_of_feature_combinations = list(itertools.combinations(columns_main,k))\n",
    "        #print len(list_of_feature_combinations)\n",
    "        #print list_of_feature_combinations\n",
    "        for element in list_of_feature_combinations:\n",
    "            col = sorted(list(element))\n",
    "            year, columns, numOfWins, gamesPlayed, numOfWinsPerc, score, C, clf= main(df_final, columns=col, year=year, model=model)\n",
    "            df_stats.at[counter,'Season'] = year\n",
    "            df_stats.at[counter,'Columns'] = columns\n",
    "            df_stats.at[counter,'WinsPred'] = numOfWins\n",
    "            df_stats.at[counter,'GamesPlayed'] = gamesPlayed\n",
    "            df_stats.at[counter,'WinsPred%'] = numOfWinsPerc\n",
    "            df_stats.at[counter,'Score'] = score\n",
    "            df_stats.at[counter,'C'] = C\n",
    "            if model == 'SVC':\n",
    "                df_stats.at[counter,'Coef weight'] = 0\n",
    "            elif model == 'DecTree':\n",
    "                df_stats.at[counter,'Coef weight'] = clf.best_estimator_.feature_importances_ \n",
    "            elif model == 'LogReg':\n",
    "                df_stats.at[counter,'Coef weight'] = clf.best_estimator_.coef_[0]\n",
    "            elif model == 'LinReg':\n",
    "                df_stats.at[counter,'Coef weight'] = clf.best_estimator_.coef_\n",
    "            counter += 1\n",
    "df_stats\n",
    "#df_stats.to_csv('ModelFeatureAnalysis_of_LinReg_1-6feat_2016.csv', index=False)\n",
    "#df_stats.to_csv('ModelFeatureAnalysis_of_LogReg_1-6feat_2016.csv', index=False)\n",
    "#df_stats.to_csv('ModelFeatureAnalysis_of_SVC_1-6feat_2016.csv', index=False)\n",
    "#df_stats.to_csv('ModelFeatureAnalysis_of_Tree_1-6feat_2016.csv', index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Plot feature number vs % of wins"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "LinReg_2016_FeatureAnalysis = pd.read_csv('ModelFeatureAnalysis_of_LinReg_1-6feat_2016.csv')\n",
    "LinReg_2016_FeatureAnalysis['#Features'] = LinReg_2016_FeatureAnalysis['Columns'].apply(lambda x: len(x.strip('[]').split(',')))\n",
    "LogReg_2016_FeatureAnalysis = pd.read_csv('ModelFeatureAnalysis_of_LogReg_1-6feat_2016.csv')\n",
    "LogReg_2016_FeatureAnalysis['#Features'] = LogReg_2016_FeatureAnalysis['Columns'].apply(lambda x: len(x.strip('[]').split(',')))\n",
    "Tree_2016_FeatureAnalysis = pd.read_csv('ModelFeatureAnalysis_of_Tree_1-6feat_2016.csv')\n",
    "Tree_2016_FeatureAnalysis['#Features'] = Tree_2016_FeatureAnalysis['Columns'].apply(lambda x: len(x.strip('[]').split(',')))\n",
    "SVC_2016_FeatureAnalysis = pd.read_csv('ModelFeatureAnalysis_of_SVC_1-6feat_2016.csv')\n",
    "SVC_2016_FeatureAnalysis['#Features'] = SVC_2016_FeatureAnalysis['Columns'].apply(lambda x: len(x.strip('[]').split(',')))\n",
    "\n",
    "# Number of features vs WinsPred%\n",
    "plt.scatter(LinReg_2016_FeatureAnalysis['#Features']+0.15, LinReg_2016_FeatureAnalysis['WinsPred%'], color='C0', alpha=0.35, label='LinReg', zorder=2)\n",
    "plt.plot(LinReg_2016_FeatureAnalysis.groupby(by=['#Features']).mean().reset_index()['#Features']+0.15, \n",
    "            LinReg_2016_FeatureAnalysis.groupby(by=['#Features']).mean().reset_index()['WinsPred%'], 'o-',color='C0', label='AVG')\n",
    "\n",
    "plt.scatter(LogReg_2016_FeatureAnalysis['#Features']-0.15, LogReg_2016_FeatureAnalysis['WinsPred%'], color='C1', alpha=0.35, label='LogReg')\n",
    "plt.plot(LogReg_2016_FeatureAnalysis.groupby(by=['#Features']).mean().reset_index()['#Features']-0.15, \n",
    "            LogReg_2016_FeatureAnalysis.groupby(by=['#Features']).mean().reset_index()['WinsPred%'], 'o-',color='C1', label='AVG', zorder=4)\n",
    "\n",
    "plt.scatter(SVC_2016_FeatureAnalysis['#Features']+0.05, SVC_2016_FeatureAnalysis['WinsPred%'], color='C2', alpha=0.35, label='SVC')\n",
    "plt.plot(SVC_2016_FeatureAnalysis.groupby(by=['#Features']).mean().reset_index()['#Features']+0.05, \n",
    "            SVC_2016_FeatureAnalysis.groupby(by=['#Features']).mean().reset_index()['WinsPred%'], 'o-',color='C2', label='AVG', zorder=3)\n",
    "\n",
    "plt.scatter(Tree_2016_FeatureAnalysis['#Features']-0.05, Tree_2016_FeatureAnalysis['WinsPred%'], color='C3', alpha=0.35, label='Tree')\n",
    "plt.plot(Tree_2016_FeatureAnalysis.groupby(by=['#Features']).mean().reset_index()['#Features']-0.05, \n",
    "            Tree_2016_FeatureAnalysis.groupby(by=['#Features']).mean().reset_index()['WinsPred%'], 'o-',color='C3', label='AVG', zorder=1)\n",
    "\n",
    "\n",
    "plt.legend(bbox_to_anchor=(1.05, 1), loc=2, borderaxespad=0.,  ncol=2)\n",
    "plt.xlabel('Number of features', size=13)\n",
    "plt.ylabel('% of correct predictions', size=13)\n",
    "plt.tick_params(labelsize=13)\n",
    "plt.savefig('NumberOfFeaturesvsWins.png', dpi=300, bbox_inches='tight')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### We can conclude that just simply adding features does not yield the best result. Some feature do worse then throwing a coin."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0,0.5,'% of correct predictions')"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "dummy = SVC_2016_FeatureAnalysis[SVC_2016_FeatureAnalysis['#Features'] == 1]\n",
    "plt.scatter(dummy['Columns'], dummy['WinsPred%'])\n",
    "plt.ylabel('% of correct predictions')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### We see that the Free Throw attempt rate 'FTAR' performs the poorest, follow by total defensive rebounds 'TDR%' and defensive rebounds 'OR%'."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#plt.figure(figsize=(10,10))\n",
    "Tree_6_Feature = Tree_2016_FeatureAnalysis[Tree_2016_FeatureAnalysis['#Features'] == 6]\n",
    "\n",
    "def convert_strList_to_floats(df):\n",
    "    if str(df) == str(Tree_6_Feature):\n",
    "        dummy = map(lambda x: x[:5], df['Coef weight'][0][1:-1].split(' ')[0:-1])\n",
    "        return np.array(map(lambda x: float(x), [d for d in dummy if d != '']))\n",
    "    else:\n",
    "        dummy = map(lambda x: x[:5], df['Coef weight'][0][1:-1].split(' ')[1:])\n",
    "        return np.array(map(lambda x: float(x), [d for d in dummy if d != '']))\n",
    "\n",
    "LogReg_6_Feature = LogReg_2016_FeatureAnalysis[LogReg_2016_FeatureAnalysis['#Features'] == 6]\n",
    "yLog = convert_strList_to_floats(LogReg_6_Feature)/abs(convert_strList_to_floats(LogReg_6_Feature)).sum()\n",
    "LinReg_6_Feature = LinReg_2016_FeatureAnalysis[LinReg_2016_FeatureAnalysis['#Features'] == 6]\n",
    "yLin = convert_strList_to_floats(LinReg_6_Feature)/abs(convert_strList_to_floats(LinReg_6_Feature)).sum()\n",
    "Tree_6_Feature = Tree_2016_FeatureAnalysis[Tree_2016_FeatureAnalysis['#Features'] == 6]\n",
    "yTree= convert_strList_to_floats(Tree_6_Feature)/abs(convert_strList_to_floats(Tree_6_Feature)).sum()\n",
    "plt.bar(np.arange(0, len(yLog),1)-0.2, abs(yLog), width=0.2, label='LogReg')\n",
    "plt.bar(np.arange(0, len(yLin),1)+0.0, abs(yLin), width=0.2, label='LinReg')\n",
    "plt.bar(np.arange(0, len(yTree),1)+0.2, abs(yTree), width=0.2, label='Tree')\n",
    "plt.xlabel('feature name', size=13)\n",
    "plt.ylabel('$\\mid$ coefficient weight $\\mid$', size=13)\n",
    "plt.tick_params(labelsize=13)\n",
    "featureList= map(lambda x: x[1:-1] ,LogReg_2016_FeatureAnalysis['Columns'][0][1:-1].split(', '))\n",
    "plt.xticks(range(0, len(featureList)),featureList)  # Set locations and labels\n",
    "plt.legend()\n",
    "plt.savefig('CoefficientWeight.png', dpa='1000')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Season</th>\n",
       "      <th>Columns</th>\n",
       "      <th>WinsPred</th>\n",
       "      <th>GamesPlayed</th>\n",
       "      <th>WinsPred%</th>\n",
       "      <th>Score</th>\n",
       "      <th>C</th>\n",
       "      <th>Coef weight</th>\n",
       "      <th>#Features</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>42</th>\n",
       "      <td>2016</td>\n",
       "      <td>['TOVR', 'eFG%']</td>\n",
       "      <td>41</td>\n",
       "      <td>67</td>\n",
       "      <td>61.2</td>\n",
       "      <td>-0.332756</td>\n",
       "      <td>3.072113e+02</td>\n",
       "      <td>[-0.2427595  31.58883838]</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>43</th>\n",
       "      <td>2016</td>\n",
       "      <td>['FTAR', 'TOVR']</td>\n",
       "      <td>29</td>\n",
       "      <td>67</td>\n",
       "      <td>43.3</td>\n",
       "      <td>-0.558251</td>\n",
       "      <td>9.547716e+00</td>\n",
       "      <td>[ 4.78002918 -0.14724575]</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44</th>\n",
       "      <td>2016</td>\n",
       "      <td>['OR%', 'TOVR']</td>\n",
       "      <td>37</td>\n",
       "      <td>67</td>\n",
       "      <td>55.2</td>\n",
       "      <td>-0.610370</td>\n",
       "      <td>8.504489e+10</td>\n",
       "      <td>[ 4.76441456 -0.11507029]</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>45</th>\n",
       "      <td>2016</td>\n",
       "      <td>['TDR%', 'TOVR']</td>\n",
       "      <td>44</td>\n",
       "      <td>67</td>\n",
       "      <td>65.7</td>\n",
       "      <td>-0.247553</td>\n",
       "      <td>5.415871e+01</td>\n",
       "      <td>[30.58861216 -0.54572012]</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>46</th>\n",
       "      <td>2016</td>\n",
       "      <td>['PIE', 'TOVR']</td>\n",
       "      <td>48</td>\n",
       "      <td>67</td>\n",
       "      <td>71.6</td>\n",
       "      <td>-0.117023</td>\n",
       "      <td>3.872040e+12</td>\n",
       "      <td>[36.86396915 -0.1492847 ]</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>47</th>\n",
       "      <td>2016</td>\n",
       "      <td>['FTAR', 'eFG%']</td>\n",
       "      <td>37</td>\n",
       "      <td>67</td>\n",
       "      <td>55.2</td>\n",
       "      <td>-0.373604</td>\n",
       "      <td>3.072113e+02</td>\n",
       "      <td>[ 4.28960659 24.83199391]</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>48</th>\n",
       "      <td>2016</td>\n",
       "      <td>['OR%', 'eFG%']</td>\n",
       "      <td>40</td>\n",
       "      <td>67</td>\n",
       "      <td>59.7</td>\n",
       "      <td>-0.382747</td>\n",
       "      <td>3.072113e+02</td>\n",
       "      <td>[ 6.2167734  26.25766119]</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>49</th>\n",
       "      <td>2016</td>\n",
       "      <td>['TDR%', 'eFG%']</td>\n",
       "      <td>45</td>\n",
       "      <td>67</td>\n",
       "      <td>67.2</td>\n",
       "      <td>-0.401510</td>\n",
       "      <td>7.663411e+01</td>\n",
       "      <td>[ 5.82137328 18.81806253]</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50</th>\n",
       "      <td>2016</td>\n",
       "      <td>['PIE', 'eFG%']</td>\n",
       "      <td>46</td>\n",
       "      <td>67</td>\n",
       "      <td>68.7</td>\n",
       "      <td>-0.127408</td>\n",
       "      <td>3.072113e+02</td>\n",
       "      <td>[35.79084454 -1.95998907]</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51</th>\n",
       "      <td>2016</td>\n",
       "      <td>['FTAR', 'OR%']</td>\n",
       "      <td>25</td>\n",
       "      <td>67</td>\n",
       "      <td>37.3</td>\n",
       "      <td>-0.581219</td>\n",
       "      <td>9.547716e+00</td>\n",
       "      <td>[4.10369707 4.91485102]</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>52</th>\n",
       "      <td>2016</td>\n",
       "      <td>['FTAR', 'TDR%']</td>\n",
       "      <td>38</td>\n",
       "      <td>67</td>\n",
       "      <td>56.7</td>\n",
       "      <td>-0.457338</td>\n",
       "      <td>1.911644e+01</td>\n",
       "      <td>[ 2.74653326 11.63569161]</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>53</th>\n",
       "      <td>2016</td>\n",
       "      <td>['FTAR', 'PIE']</td>\n",
       "      <td>45</td>\n",
       "      <td>67</td>\n",
       "      <td>67.2</td>\n",
       "      <td>-0.127321</td>\n",
       "      <td>2.171118e+02</td>\n",
       "      <td>[-0.90891855 36.05055433]</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>54</th>\n",
       "      <td>2016</td>\n",
       "      <td>['OR%', 'TDR%']</td>\n",
       "      <td>43</td>\n",
       "      <td>67</td>\n",
       "      <td>64.2</td>\n",
       "      <td>-0.479169</td>\n",
       "      <td>1.084366e+02</td>\n",
       "      <td>[-0.87647866 12.97250835]</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>55</th>\n",
       "      <td>2016</td>\n",
       "      <td>['OR%', 'PIE']</td>\n",
       "      <td>48</td>\n",
       "      <td>67</td>\n",
       "      <td>71.6</td>\n",
       "      <td>-0.125426</td>\n",
       "      <td>1.534368e+02</td>\n",
       "      <td>[ 2.95664837 35.4583398 ]</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>56</th>\n",
       "      <td>2016</td>\n",
       "      <td>['PIE', 'TDR%']</td>\n",
       "      <td>46</td>\n",
       "      <td>67</td>\n",
       "      <td>68.7</td>\n",
       "      <td>-0.121886</td>\n",
       "      <td>3.489101e+03</td>\n",
       "      <td>[39.20505026 -5.8394251 ]</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    Season           Columns  WinsPred  GamesPlayed  WinsPred%     Score  \\\n",
       "42    2016  ['TOVR', 'eFG%']        41           67       61.2 -0.332756   \n",
       "43    2016  ['FTAR', 'TOVR']        29           67       43.3 -0.558251   \n",
       "44    2016   ['OR%', 'TOVR']        37           67       55.2 -0.610370   \n",
       "45    2016  ['TDR%', 'TOVR']        44           67       65.7 -0.247553   \n",
       "46    2016   ['PIE', 'TOVR']        48           67       71.6 -0.117023   \n",
       "47    2016  ['FTAR', 'eFG%']        37           67       55.2 -0.373604   \n",
       "48    2016   ['OR%', 'eFG%']        40           67       59.7 -0.382747   \n",
       "49    2016  ['TDR%', 'eFG%']        45           67       67.2 -0.401510   \n",
       "50    2016   ['PIE', 'eFG%']        46           67       68.7 -0.127408   \n",
       "51    2016   ['FTAR', 'OR%']        25           67       37.3 -0.581219   \n",
       "52    2016  ['FTAR', 'TDR%']        38           67       56.7 -0.457338   \n",
       "53    2016   ['FTAR', 'PIE']        45           67       67.2 -0.127321   \n",
       "54    2016   ['OR%', 'TDR%']        43           67       64.2 -0.479169   \n",
       "55    2016    ['OR%', 'PIE']        48           67       71.6 -0.125426   \n",
       "56    2016   ['PIE', 'TDR%']        46           67       68.7 -0.121886   \n",
       "\n",
       "               C                Coef weight  #Features  \n",
       "42  3.072113e+02  [-0.2427595  31.58883838]          2  \n",
       "43  9.547716e+00  [ 4.78002918 -0.14724575]          2  \n",
       "44  8.504489e+10  [ 4.76441456 -0.11507029]          2  \n",
       "45  5.415871e+01  [30.58861216 -0.54572012]          2  \n",
       "46  3.872040e+12  [36.86396915 -0.1492847 ]          2  \n",
       "47  3.072113e+02  [ 4.28960659 24.83199391]          2  \n",
       "48  3.072113e+02  [ 6.2167734  26.25766119]          2  \n",
       "49  7.663411e+01  [ 5.82137328 18.81806253]          2  \n",
       "50  3.072113e+02  [35.79084454 -1.95998907]          2  \n",
       "51  9.547716e+00    [4.10369707 4.91485102]          2  \n",
       "52  1.911644e+01  [ 2.74653326 11.63569161]          2  \n",
       "53  2.171118e+02  [-0.90891855 36.05055433]          2  \n",
       "54  1.084366e+02  [-0.87647866 12.97250835]          2  \n",
       "55  1.534368e+02  [ 2.95664837 35.4583398 ]          2  \n",
       "56  3.489101e+03  [39.20505026 -5.8394251 ]          2  "
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "LogReg_2016_FeatureAnalysis[LogReg_2016_FeatureAnalysis['#Features'] == 2]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We find again that __['FTAR', 'OR%']__ have the lowest prediction percentage (just __37.3%!!__). 'TDR' sometimes performs well. Let's check what the best result has."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Season</th>\n",
       "      <th>Columns</th>\n",
       "      <th>WinsPred</th>\n",
       "      <th>GamesPlayed</th>\n",
       "      <th>WinsPred%</th>\n",
       "      <th>Score</th>\n",
       "      <th>C</th>\n",
       "      <th>Coef weight</th>\n",
       "      <th>#Features</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2016</td>\n",
       "      <td>['FTAR', 'OR%', 'TDR%', 'TOVR', 'eFG%']</td>\n",
       "      <td>42</td>\n",
       "      <td>67</td>\n",
       "      <td>62.7</td>\n",
       "      <td>-0.185736</td>\n",
       "      <td>2.171118e+02</td>\n",
       "      <td>[ 4.71178163  0.83372529 21.7891919  -0.574443...</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2016</td>\n",
       "      <td>['FTAR', 'PIE', 'TDR%', 'TOVR', 'eFG%']</td>\n",
       "      <td>49</td>\n",
       "      <td>67</td>\n",
       "      <td>73.1</td>\n",
       "      <td>-0.115745</td>\n",
       "      <td>2.171118e+02</td>\n",
       "      <td>[ 0.23297294 33.33915275  2.83271426 -0.220802...</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>2016</td>\n",
       "      <td>['FTAR', 'OR%', 'PIE', 'TDR%', 'TOVR']</td>\n",
       "      <td>49</td>\n",
       "      <td>67</td>\n",
       "      <td>73.1</td>\n",
       "      <td>-0.112282</td>\n",
       "      <td>1.499268e+10</td>\n",
       "      <td>[-0.42975027  5.65791047 41.24620472 -5.512861...</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2016</td>\n",
       "      <td>['FTAR', 'OR%', 'PIE', 'TOVR', 'eFG%']</td>\n",
       "      <td>50</td>\n",
       "      <td>67</td>\n",
       "      <td>74.6</td>\n",
       "      <td>-0.109464</td>\n",
       "      <td>1.084366e+02</td>\n",
       "      <td>[ 0.75455947  6.02760272 33.52950415 -0.233520...</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>2016</td>\n",
       "      <td>['FTAR', 'OR%', 'PIE', 'TDR%', 'eFG%']</td>\n",
       "      <td>51</td>\n",
       "      <td>67</td>\n",
       "      <td>76.1</td>\n",
       "      <td>-0.106301</td>\n",
       "      <td>2.171118e+02</td>\n",
       "      <td>[  0.63044746  11.34080265  42.05708932 -16.75...</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2016</td>\n",
       "      <td>['OR%', 'PIE', 'TDR%', 'TOVR', 'eFG%']</td>\n",
       "      <td>52</td>\n",
       "      <td>67</td>\n",
       "      <td>77.6</td>\n",
       "      <td>-0.105512</td>\n",
       "      <td>2.171118e+02</td>\n",
       "      <td>[  9.99455931  40.4079143  -12.11825485  -0.07...</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Season                                  Columns  WinsPred  GamesPlayed  \\\n",
       "1    2016  ['FTAR', 'OR%', 'TDR%', 'TOVR', 'eFG%']        42           67   \n",
       "3    2016  ['FTAR', 'PIE', 'TDR%', 'TOVR', 'eFG%']        49           67   \n",
       "5    2016   ['FTAR', 'OR%', 'PIE', 'TDR%', 'TOVR']        49           67   \n",
       "2    2016   ['FTAR', 'OR%', 'PIE', 'TOVR', 'eFG%']        50           67   \n",
       "6    2016   ['FTAR', 'OR%', 'PIE', 'TDR%', 'eFG%']        51           67   \n",
       "4    2016   ['OR%', 'PIE', 'TDR%', 'TOVR', 'eFG%']        52           67   \n",
       "\n",
       "   WinsPred%     Score             C  \\\n",
       "1       62.7 -0.185736  2.171118e+02   \n",
       "3       73.1 -0.115745  2.171118e+02   \n",
       "5       73.1 -0.112282  1.499268e+10   \n",
       "2       74.6 -0.109464  1.084366e+02   \n",
       "6       76.1 -0.106301  2.171118e+02   \n",
       "4       77.6 -0.105512  2.171118e+02   \n",
       "\n",
       "                                         Coef weight  #Features  \n",
       "1  [ 4.71178163  0.83372529 21.7891919  -0.574443...          5  \n",
       "3  [ 0.23297294 33.33915275  2.83271426 -0.220802...          5  \n",
       "5  [-0.42975027  5.65791047 41.24620472 -5.512861...          5  \n",
       "2  [ 0.75455947  6.02760272 33.52950415 -0.233520...          5  \n",
       "6  [  0.63044746  11.34080265  42.05708932 -16.75...          5  \n",
       "4  [  9.99455931  40.4079143  -12.11825485  -0.07...          5  "
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "LogReg_2016_FeatureAnalysis[LogReg_2016_FeatureAnalysis['#Features'] == 5].sort_values(by=['WinsPred%'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    " __'OR%', 'PIE','eFG%', 'TOVR__' are always at the top __'OR%','TDR%' and 'FTAR'__ seem to be at the bottom."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##  Train and evaluate different models for different years (fixed features)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 405,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LinReg\n",
      "Best score: 0.6585\n",
      "Number of teams playing in 2003 Tournnament: 65/68 (64 + The first four).\n",
      "2003 Tournament games predicted correctly: 47 out of 64.0. (73.4%)\n",
      "LinReg\n",
      "Best score: 0.6391\n",
      "Number of teams playing in 2004 Tournnament: 65/68 (64 + The first four).\n",
      "2004 Tournament games predicted correctly: 41 out of 64.0. (64.1%)\n",
      "LinReg\n",
      "Best score: 0.6396\n",
      "Number of teams playing in 2005 Tournnament: 65/68 (64 + The first four).\n",
      "2005 Tournament games predicted correctly: 44 out of 64.0. (68.8%)\n",
      "LinReg\n",
      "Best score: 0.6415\n",
      "Number of teams playing in 2006 Tournnament: 65/68 (64 + The first four).\n",
      "2006 Tournament games predicted correctly: 46 out of 64.0. (71.9%)\n",
      "LinReg\n",
      "Best score: 0.6321\n",
      "Number of teams playing in 2007 Tournnament: 65/68 (64 + The first four).\n",
      "2007 Tournament games predicted correctly: 43 out of 64.0. (67.2%)\n",
      "LinReg\n",
      "Best score: 0.6361\n",
      "Number of teams playing in 2008 Tournnament: 65/68 (64 + The first four).\n",
      "2008 Tournament games predicted correctly: 46 out of 64.0. (71.9%)\n",
      "LinReg\n",
      "Best score: 0.6438\n",
      "Number of teams playing in 2009 Tournnament: 65/68 (64 + The first four).\n",
      "2009 Tournament games predicted correctly: 45 out of 64.0. (70.3%)\n",
      "LinReg\n",
      "Best score: 0.646\n",
      "Number of teams playing in 2010 Tournnament: 65/68 (64 + The first four).\n",
      "2010 Tournament games predicted correctly: 38 out of 64.0. (59.4%)\n",
      "LinReg\n",
      "Best score: 0.64\n",
      "Number of teams playing in 2011 Tournnament: 68/68 (64 + The first four).\n",
      "2011 Tournament games predicted correctly: 45 out of 67.0. (67.2%)\n",
      "LinReg\n",
      "Best score: 0.6475\n",
      "Number of teams playing in 2012 Tournnament: 68/68 (64 + The first four).\n",
      "2012 Tournament games predicted correctly: 42 out of 67.0. (62.7%)\n",
      "LinReg\n",
      "Best score: 0.646\n",
      "Number of teams playing in 2013 Tournnament: 68/68 (64 + The first four).\n",
      "2013 Tournament games predicted correctly: 40 out of 67.0. (59.7%)\n",
      "LinReg\n",
      "Best score: 0.6444\n",
      "Number of teams playing in 2014 Tournnament: 68/68 (64 + The first four).\n",
      "2014 Tournament games predicted correctly: 41 out of 67.0. (61.2%)\n",
      "LinReg\n",
      "Best score: 0.6372\n",
      "Number of teams playing in 2015 Tournnament: 68/68 (64 + The first four).\n",
      "2015 Tournament games predicted correctly: 44 out of 67.0. (65.7%)\n",
      "LinReg\n",
      "Best score: 0.6363\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 51 out of 67.0. (76.1%)\n",
      "LinReg\n",
      "Best score: 0.6326\n",
      "Number of teams playing in 2017 Tournnament: 68/68 (64 + The first four).\n",
      "2017 Tournament games predicted correctly: 43 out of 67.0. (64.2%)\n",
      "LinReg\n",
      "Best score: 0.6398\n",
      "Number of teams playing in 2018 Tournnament: 68/68 (64 + The first four).\n",
      "2018 Tournament games predicted correctly: 44 out of 67.0. (65.7%)\n",
      "   Season                 Columns WinsPred GamesPlayed WinsPred%  \\\n",
      "0    2003  [OR%, PIE, TOVR, eFG%]       47          64      73.4   \n",
      "1    2004  [OR%, PIE, TOVR, eFG%]       41          64      64.1   \n",
      "2    2005  [OR%, PIE, TOVR, eFG%]       44          64      68.8   \n",
      "3    2006  [OR%, PIE, TOVR, eFG%]       46          64      71.9   \n",
      "4    2007  [OR%, PIE, TOVR, eFG%]       43          64      67.2   \n",
      "5    2008  [OR%, PIE, TOVR, eFG%]       46          64      71.9   \n",
      "6    2009  [OR%, PIE, TOVR, eFG%]       45          64      70.3   \n",
      "7    2010  [OR%, PIE, TOVR, eFG%]       38          64      59.4   \n",
      "8    2011  [OR%, PIE, TOVR, eFG%]       45          67      67.2   \n",
      "9    2012  [OR%, PIE, TOVR, eFG%]       42          67      62.7   \n",
      "10   2013  [OR%, PIE, TOVR, eFG%]       40          67      59.7   \n",
      "11   2014  [OR%, PIE, TOVR, eFG%]       41          67      61.2   \n",
      "12   2015  [OR%, PIE, TOVR, eFG%]       44          67      65.7   \n",
      "13   2016  [OR%, PIE, TOVR, eFG%]       51          67      76.1   \n",
      "14   2017  [OR%, PIE, TOVR, eFG%]       43          67      64.2   \n",
      "15   2018  [OR%, PIE, TOVR, eFG%]       44          67      65.7   \n",
      "\n",
      "                                          Coef weight     Score  C  \n",
      "0   [0.03466936124285434, 1.4354557555941243, -0.0...  0.658501  0  \n",
      "1   [0.12219967021070889, 1.3606204743554302, -0.0...  0.639092  0  \n",
      "2   [0.08406515913581077, 1.3418785685430108, -0.0...  0.639597  0  \n",
      "3   [0.08244929443574932, 1.352757838108566, -0.00...  0.641458  0  \n",
      "4   [0.025855082333502345, 1.3376082621539194, -0....  0.632101  0  \n",
      "5   [0.07731006702944146, 1.2668050496280392, -0.0...  0.636058  0  \n",
      "6   [0.09087074513799502, 1.2895191788631284, -0.0...  0.643815  0  \n",
      "7   [0.031648087994844004, 1.3633416565302798, -0....  0.645978  0  \n",
      "8   [0.1132529485316191, 1.3986289690355302, -0.00...  0.640001  0  \n",
      "9   [0.09131534951258945, 1.3690792002189165, -0.0...  0.647489  0  \n",
      "10  [0.024290385990444543, 1.3682328195548878, -0....  0.646005  0  \n",
      "11  [0.1366510662755733, 1.476078666090125, -0.006...  0.644412  0  \n",
      "12  [-0.0008506304225442497, 1.3826706562907363, -...  0.637187  0  \n",
      "13  [0.08191310838587891, 1.4595214513079566, -0.0...  0.636314  0  \n",
      "14  [0.04187462365912158, 1.4742987757776098, -0.0...  0.632575  0  \n",
      "15  [0.031986342450708714, 1.4903176261480042, -0....   0.63982  0  \n",
      "LogReg\n",
      "Best log_loss: -0.1032, with best C: 217.111794569\n",
      "Number of teams playing in 2003 Tournnament: 65/68 (64 + The first four).\n",
      "2003 Tournament games predicted correctly: 47 out of 64.0. (73.4%)\n",
      "LogReg\n",
      "Best log_loss: -0.1165, with best C: 108.436596869\n",
      "Number of teams playing in 2004 Tournnament: 65/68 (64 + The first four).\n",
      "2004 Tournament games predicted correctly: 42 out of 64.0. (65.6%)\n",
      "LogReg\n",
      "Best log_loss: -0.1121, with best C: 108.436596869\n",
      "Number of teams playing in 2005 Tournnament: 65/68 (64 + The first four).\n",
      "2005 Tournament games predicted correctly: 44 out of 64.0. (68.8%)\n",
      "LogReg\n",
      "Best log_loss: -0.1101, with best C: 307.211299886\n",
      "Number of teams playing in 2006 Tournnament: 65/68 (64 + The first four).\n",
      "2006 Tournament games predicted correctly: 47 out of 64.0. (73.4%)\n",
      "LogReg\n",
      "Best log_loss: -0.1131, with best C: 307.211299886\n",
      "Number of teams playing in 2007 Tournnament: 65/68 (64 + The first four).\n",
      "2007 Tournament games predicted correctly: 43 out of 64.0. (67.2%)\n",
      "LogReg\n",
      "Best log_loss: -0.1059, with best C: 153.436840893\n",
      "Number of teams playing in 2008 Tournnament: 65/68 (64 + The first four).\n",
      "2008 Tournament games predicted correctly: 49 out of 64.0. (76.6%)\n",
      "LogReg\n",
      "Best log_loss: -0.1092, with best C: 54.1587137808\n",
      "Number of teams playing in 2009 Tournnament: 65/68 (64 + The first four).\n",
      "2009 Tournament games predicted correctly: 46 out of 64.0. (71.9%)\n",
      "LogReg\n",
      "Best log_loss: -0.1041, with best C: 76.6341086801\n",
      "Number of teams playing in 2010 Tournnament: 65/68 (64 + The first four).\n",
      "2010 Tournament games predicted correctly: 40 out of 64.0. (62.5%)\n",
      "LogReg\n",
      "Best log_loss: -0.11, with best C: 76.6341086801\n",
      "Number of teams playing in 2011 Tournnament: 68/68 (64 + The first four).\n",
      "2011 Tournament games predicted correctly: 45 out of 67.0. (67.2%)\n",
      "LogReg\n",
      "Best log_loss: -0.1083, with best C: 54.1587137808\n",
      "Number of teams playing in 2012 Tournnament: 68/68 (64 + The first four).\n",
      "2012 Tournament games predicted correctly: 44 out of 67.0. (65.7%)\n",
      "LogReg\n",
      "Best log_loss: -0.1024, with best C: 870.359136149\n",
      "Number of teams playing in 2013 Tournnament: 68/68 (64 + The first four).\n",
      "2013 Tournament games predicted correctly: 39 out of 67.0. (58.2%)\n",
      "LogReg\n",
      "Best log_loss: -0.1087, with best C: 153.436840893\n",
      "Number of teams playing in 2014 Tournnament: 68/68 (64 + The first four).\n",
      "2014 Tournament games predicted correctly: 42 out of 67.0. (62.7%)\n",
      "LogReg\n",
      "Best log_loss: -0.1043, with best C: 153.436840893\n",
      "Number of teams playing in 2015 Tournnament: 68/68 (64 + The first four).\n",
      "2015 Tournament games predicted correctly: 45 out of 67.0. (67.2%)\n",
      "LogReg\n",
      "Best log_loss: -0.1089, with best C: 434.701315813\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 50 out of 67.0. (74.6%)\n",
      "LogReg\n",
      "Best log_loss: -0.1058, with best C: 615.098578858\n",
      "Number of teams playing in 2017 Tournnament: 68/68 (64 + The first four).\n",
      "2017 Tournament games predicted correctly: 44 out of 67.0. (65.7%)\n",
      "LogReg\n",
      "Best log_loss: -0.1088, with best C: 153.436840893\n",
      "Number of teams playing in 2018 Tournnament: 68/68 (64 + The first four).\n",
      "2018 Tournament games predicted correctly: 45 out of 67.0. (67.2%)\n",
      "   Season                 Columns WinsPred GamesPlayed WinsPred%  \\\n",
      "0    2003  [OR%, PIE, TOVR, eFG%]       47          64      73.4   \n",
      "1    2004  [OR%, PIE, TOVR, eFG%]       42          64      65.6   \n",
      "2    2005  [OR%, PIE, TOVR, eFG%]       44          64      68.8   \n",
      "3    2006  [OR%, PIE, TOVR, eFG%]       47          64      73.4   \n",
      "4    2007  [OR%, PIE, TOVR, eFG%]       43          64      67.2   \n",
      "5    2008  [OR%, PIE, TOVR, eFG%]       49          64      76.6   \n",
      "6    2009  [OR%, PIE, TOVR, eFG%]       46          64      71.9   \n",
      "7    2010  [OR%, PIE, TOVR, eFG%]       40          64      62.5   \n",
      "8    2011  [OR%, PIE, TOVR, eFG%]       45          67      67.2   \n",
      "9    2012  [OR%, PIE, TOVR, eFG%]       44          67      65.7   \n",
      "10   2013  [OR%, PIE, TOVR, eFG%]       39          67      58.2   \n",
      "11   2014  [OR%, PIE, TOVR, eFG%]       42          67      62.7   \n",
      "12   2015  [OR%, PIE, TOVR, eFG%]       45          67      67.2   \n",
      "13   2016  [OR%, PIE, TOVR, eFG%]       50          67      74.6   \n",
      "14   2017  [OR%, PIE, TOVR, eFG%]       44          67      65.7   \n",
      "15   2018  [OR%, PIE, TOVR, eFG%]       45          67      67.2   \n",
      "\n",
      "                                          Coef weight     Score        C  \n",
      "0   [4.030477366917761, 33.49086652877295, -0.1844... -0.103228  217.112  \n",
      "1   [5.542392176186442, 29.87072302536203, -0.2155... -0.116509  108.437  \n",
      "2   [5.480545588561088, 31.00628332363461, -0.2147... -0.112061  108.437  \n",
      "3   [5.40357328214483, 32.20589012204449, -0.24636... -0.110092  307.211  \n",
      "4   [3.809490421161616, 31.96764036833921, -0.2095... -0.113058  307.211  \n",
      "5   [6.468363303684296, 30.647463837656186, -0.277... -0.105939  153.437  \n",
      "6   [5.4739559415154275, 31.279394398710206, -0.22... -0.109179  54.1587  \n",
      "7   [4.741570368241458, 33.14398055779688, -0.2163... -0.104107  76.6341  \n",
      "8   [6.286461298223976, 31.994410739842913, -0.218... -0.109957  76.6341  \n",
      "9   [5.446057347243828, 31.649338388214634, -0.227... -0.108262  54.1587  \n",
      "10  [5.5196658245688255, 33.72972944691241, -0.252...  -0.10238  870.359  \n",
      "11  [5.995254753376865, 34.82470060742012, -0.2122... -0.108724  153.437  \n",
      "12  [4.867479136361889, 34.27286048649802, -0.2268... -0.104317  153.437  \n",
      "13  [5.927042813552175, 34.78977485649933, -0.2266... -0.108936  434.701  \n",
      "14  [6.559320057558486, 36.727820321339834, -0.210... -0.105846  615.099  \n",
      "15  [5.643203481433902, 35.27752766350507, -0.2277... -0.108752  153.437  \n",
      "DecTree\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best log_loss: -0.1224, with best max_depth: 3, best min_samples_split: 2, min_samples_leaf:2, max_features_split:4\n",
      "Number of teams playing in 2003 Tournnament: 65/68 (64 + The first four).\n",
      "2003 Tournament games predicted correctly: 44 out of 64.0. (68.8%)\n",
      "DecTree\n",
      "Best log_loss: -0.135, with best max_depth: 4, best min_samples_split: 3, min_samples_leaf:2, max_features_split:3\n",
      "Number of teams playing in 2004 Tournnament: 65/68 (64 + The first four).\n",
      "2004 Tournament games predicted correctly: 39 out of 64.0. (60.9%)\n",
      "DecTree\n",
      "Best log_loss: -0.1328, with best max_depth: 4, best min_samples_split: 3, min_samples_leaf:3, max_features_split:3\n",
      "Number of teams playing in 2005 Tournnament: 65/68 (64 + The first four).\n",
      "2005 Tournament games predicted correctly: 44 out of 64.0. (68.8%)\n",
      "DecTree\n",
      "Best log_loss: -0.1358, with best max_depth: 4, best min_samples_split: 6, min_samples_leaf:1, max_features_split:3\n",
      "Number of teams playing in 2006 Tournnament: 65/68 (64 + The first four).\n",
      "2006 Tournament games predicted correctly: 44 out of 64.0. (68.8%)\n",
      "DecTree\n",
      "Best log_loss: -0.1309, with best max_depth: 4, best min_samples_split: 4, min_samples_leaf:2, max_features_split:3\n",
      "Number of teams playing in 2007 Tournnament: 65/68 (64 + The first four).\n",
      "2007 Tournament games predicted correctly: 44 out of 64.0. (68.8%)\n",
      "DecTree\n",
      "Best log_loss: -0.1341, with best max_depth: 3, best min_samples_split: 2, min_samples_leaf:1, max_features_split:4\n",
      "Number of teams playing in 2008 Tournnament: 65/68 (64 + The first four).\n",
      "2008 Tournament games predicted correctly: 48 out of 64.0. (75.0%)\n",
      "DecTree\n",
      "Best log_loss: -0.1306, with best max_depth: 4, best min_samples_split: 2, min_samples_leaf:4, max_features_split:4\n",
      "Number of teams playing in 2009 Tournnament: 65/68 (64 + The first four).\n",
      "2009 Tournament games predicted correctly: 43 out of 64.0. (67.2%)\n",
      "DecTree\n",
      "Best log_loss: -0.1256, with best max_depth: 3, best min_samples_split: 2, min_samples_leaf:1, max_features_split:4\n",
      "Number of teams playing in 2010 Tournnament: 65/68 (64 + The first four).\n",
      "2010 Tournament games predicted correctly: 39 out of 64.0. (60.9%)\n",
      "DecTree\n",
      "Best log_loss: -0.1362, with best max_depth: 4, best min_samples_split: 2, min_samples_leaf:4, max_features_split:4\n",
      "Number of teams playing in 2011 Tournnament: 68/68 (64 + The first four).\n",
      "2011 Tournament games predicted correctly: 43 out of 67.0. (64.2%)\n",
      "DecTree\n",
      "Best log_loss: -0.1306, with best max_depth: 4, best min_samples_split: 5, min_samples_leaf:3, max_features_split:3\n",
      "Number of teams playing in 2012 Tournnament: 68/68 (64 + The first four).\n",
      "2012 Tournament games predicted correctly: 43 out of 67.0. (64.2%)\n",
      "DecTree\n",
      "Best log_loss: -0.1347, with best max_depth: 4, best min_samples_split: 4, min_samples_leaf:3, max_features_split:3\n",
      "Number of teams playing in 2013 Tournnament: 68/68 (64 + The first four).\n",
      "2013 Tournament games predicted correctly: 38 out of 67.0. (56.7%)\n",
      "DecTree\n",
      "Best log_loss: -0.1309, with best max_depth: 4, best min_samples_split: 2, min_samples_leaf:1, max_features_split:4\n",
      "Number of teams playing in 2014 Tournnament: 68/68 (64 + The first four).\n",
      "2014 Tournament games predicted correctly: 41 out of 67.0. (61.2%)\n",
      "DecTree\n",
      "Best log_loss: -0.1279, with best max_depth: 4, best min_samples_split: 6, min_samples_leaf:3, max_features_split:3\n",
      "Number of teams playing in 2015 Tournnament: 68/68 (64 + The first four).\n",
      "2015 Tournament games predicted correctly: 45 out of 67.0. (67.2%)\n",
      "DecTree\n",
      "Best log_loss: -0.1317, with best max_depth: 3, best min_samples_split: 2, min_samples_leaf:3, max_features_split:4\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 44 out of 67.0. (65.7%)\n",
      "DecTree\n",
      "Best log_loss: -0.1251, with best max_depth: 4, best min_samples_split: 4, min_samples_leaf:4, max_features_split:3\n",
      "Number of teams playing in 2017 Tournnament: 68/68 (64 + The first four).\n",
      "2017 Tournament games predicted correctly: 41 out of 67.0. (61.2%)\n",
      "DecTree\n",
      "Best log_loss: -0.1315, with best max_depth: 3, best min_samples_split: 2, min_samples_leaf:1, max_features_split:4\n",
      "Number of teams playing in 2018 Tournnament: 68/68 (64 + The first four).\n",
      "2018 Tournament games predicted correctly: 44 out of 67.0. (65.7%)\n",
      "   Season                 Columns WinsPred GamesPlayed WinsPred%  \\\n",
      "0    2003  [OR%, PIE, TOVR, eFG%]       44          64      68.8   \n",
      "1    2004  [OR%, PIE, TOVR, eFG%]       39          64      60.9   \n",
      "2    2005  [OR%, PIE, TOVR, eFG%]       44          64      68.8   \n",
      "3    2006  [OR%, PIE, TOVR, eFG%]       44          64      68.8   \n",
      "4    2007  [OR%, PIE, TOVR, eFG%]       44          64      68.8   \n",
      "5    2008  [OR%, PIE, TOVR, eFG%]       48          64        75   \n",
      "6    2009  [OR%, PIE, TOVR, eFG%]       43          64      67.2   \n",
      "7    2010  [OR%, PIE, TOVR, eFG%]       39          64      60.9   \n",
      "8    2011  [OR%, PIE, TOVR, eFG%]       43          67      64.2   \n",
      "9    2012  [OR%, PIE, TOVR, eFG%]       43          67      64.2   \n",
      "10   2013  [OR%, PIE, TOVR, eFG%]       38          67      56.7   \n",
      "11   2014  [OR%, PIE, TOVR, eFG%]       41          67      61.2   \n",
      "12   2015  [OR%, PIE, TOVR, eFG%]       45          67      67.2   \n",
      "13   2016  [OR%, PIE, TOVR, eFG%]       44          67      65.7   \n",
      "14   2017  [OR%, PIE, TOVR, eFG%]       41          67      61.2   \n",
      "15   2018  [OR%, PIE, TOVR, eFG%]       44          67      65.7   \n",
      "\n",
      "                                          Coef weight     Score  C  \n",
      "0   [0.0, 0.9899757232593034, 0.010024276740696642... -0.122416  3  \n",
      "1   [0.005160784634871491, 0.9770599212032957, 0.0... -0.134958  4  \n",
      "2   [0.001243933284847888, 0.5024670452155086, 0.0... -0.132844  4  \n",
      "3   [0.0, 0.582689553681066, 0.013928909582202807,... -0.135819  4  \n",
      "4   [0.0, 0.5456650949199638, 0.006127060946341437... -0.130868  4  \n",
      "5   [0.0, 0.9835967493941917, 0.016403250605808238... -0.134091  3  \n",
      "6   [0.003486795897064249, 0.9805696489469391, 0.0... -0.130587  4  \n",
      "7   [0.0, 0.9916663761757644, 0.008333623824235628... -0.125579  3  \n",
      "8   [0.004451624692745218, 0.9845691571301457, 0.0... -0.136233  4  \n",
      "9   [0.0008075400840837235, 0.9883264527545841, 0....  -0.13062  4  \n",
      "10  [0.0025037121598417587, 0.986151544627388, 0.0... -0.134722  4  \n",
      "11  [0.0012871562305696576, 0.9882020201014855, 0.... -0.130945  4  \n",
      "12  [0.0003499967270251435, 0.9831970374285525, 0.... -0.127858  4  \n",
      "13  [0.0, 0.9882396759528786, 0.011760324047121328... -0.131723  3  \n",
      "14  [0.0070710778809435805, 0.9828027491126556, 0....   -0.1251  4  \n",
      "15  [0.0, 0.9909007247618992, 0.009099275238100809... -0.131508  3  \n",
      "SVC\n",
      "Best log_loss: -0.1117, with best C: 21.5443469003\n",
      "Number of teams playing in 2003 Tournnament: 65/68 (64 + The first four).\n",
      "2003 Tournament games predicted correctly: 47 out of 64.0. (73.4%)\n",
      "SVC\n",
      "Best log_loss: -0.1232, with best C: 5.99484250319\n",
      "Number of teams playing in 2004 Tournnament: 65/68 (64 + The first four).\n",
      "2004 Tournament games predicted correctly: 42 out of 64.0. (65.6%)\n",
      "SVC\n",
      "Best log_loss: -0.1175, with best C: 21.5443469003\n",
      "Number of teams playing in 2005 Tournnament: 65/68 (64 + The first four).\n",
      "2005 Tournament games predicted correctly: 45 out of 64.0. (70.3%)\n",
      "SVC\n",
      "Best log_loss: -0.1207, with best C: 21.5443469003\n",
      "Number of teams playing in 2006 Tournnament: 65/68 (64 + The first four).\n",
      "2006 Tournament games predicted correctly: 46 out of 64.0. (71.9%)\n",
      "SVC\n",
      "Best log_loss: -0.1192, with best C: 77.4263682681\n",
      "Number of teams playing in 2007 Tournnament: 65/68 (64 + The first four).\n",
      "2007 Tournament games predicted correctly: 45 out of 64.0. (70.3%)\n",
      "SVC\n",
      "Best log_loss: -0.1123, with best C: 5.99484250319\n",
      "Number of teams playing in 2008 Tournnament: 65/68 (64 + The first four).\n",
      "2008 Tournament games predicted correctly: 50 out of 64.0. (78.1%)\n",
      "SVC\n",
      "Best log_loss: -0.1143, with best C: 21.5443469003\n",
      "Number of teams playing in 2009 Tournnament: 65/68 (64 + The first four).\n",
      "2009 Tournament games predicted correctly: 46 out of 64.0. (71.9%)\n",
      "SVC\n",
      "Best log_loss: -0.1112, with best C: 21.5443469003\n",
      "Number of teams playing in 2010 Tournnament: 65/68 (64 + The first four).\n",
      "2010 Tournament games predicted correctly: 38 out of 64.0. (59.4%)\n",
      "SVC\n",
      "Best log_loss: -0.119, with best C: 5.99484250319\n",
      "Number of teams playing in 2011 Tournnament: 68/68 (64 + The first four).\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2011 Tournament games predicted correctly: 44 out of 67.0. (65.7%)\n",
      "SVC\n",
      "Best log_loss: -0.1175, with best C: 5.99484250319\n",
      "Number of teams playing in 2012 Tournnament: 68/68 (64 + The first four).\n",
      "2012 Tournament games predicted correctly: 41 out of 67.0. (61.2%)\n",
      "SVC\n",
      "Best log_loss: -0.1104, with best C: 5.99484250319\n",
      "Number of teams playing in 2013 Tournnament: 68/68 (64 + The first four).\n",
      "2013 Tournament games predicted correctly: 36 out of 67.0. (53.7%)\n",
      "SVC\n",
      "Best log_loss: -0.1138, with best C: 77.4263682681\n",
      "Number of teams playing in 2014 Tournnament: 68/68 (64 + The first four).\n",
      "2014 Tournament games predicted correctly: 42 out of 67.0. (62.7%)\n",
      "SVC\n",
      "Best log_loss: -0.112, with best C: 77.4263682681\n",
      "Number of teams playing in 2015 Tournnament: 68/68 (64 + The first four).\n",
      "2015 Tournament games predicted correctly: 43 out of 67.0. (64.2%)\n",
      "SVC\n",
      "Best log_loss: -0.1136, with best C: 21.5443469003\n",
      "Number of teams playing in 2016 Tournnament: 68/68 (64 + The first four).\n",
      "2016 Tournament games predicted correctly: 50 out of 67.0. (74.6%)\n",
      "SVC\n",
      "Best log_loss: -0.1102, with best C: 77.4263682681\n",
      "Number of teams playing in 2017 Tournnament: 68/68 (64 + The first four).\n",
      "2017 Tournament games predicted correctly: 43 out of 67.0. (64.2%)\n",
      "SVC\n",
      "Best log_loss: -0.1145, with best C: 5.99484250319\n",
      "Number of teams playing in 2018 Tournnament: 68/68 (64 + The first four).\n",
      "2018 Tournament games predicted correctly: 45 out of 67.0. (67.2%)\n",
      "   Season                 Columns WinsPred GamesPlayed WinsPred% Coef weight  \\\n",
      "0    2003  [OR%, PIE, TOVR, eFG%]       47          64      73.4           0   \n",
      "1    2004  [OR%, PIE, TOVR, eFG%]       42          64      65.6           0   \n",
      "2    2005  [OR%, PIE, TOVR, eFG%]       45          64      70.3           0   \n",
      "3    2006  [OR%, PIE, TOVR, eFG%]       46          64      71.9           0   \n",
      "4    2007  [OR%, PIE, TOVR, eFG%]       45          64      70.3           0   \n",
      "5    2008  [OR%, PIE, TOVR, eFG%]       50          64      78.1           0   \n",
      "6    2009  [OR%, PIE, TOVR, eFG%]       46          64      71.9           0   \n",
      "7    2010  [OR%, PIE, TOVR, eFG%]       38          64      59.4           0   \n",
      "8    2011  [OR%, PIE, TOVR, eFG%]       44          67      65.7           0   \n",
      "9    2012  [OR%, PIE, TOVR, eFG%]       41          67      61.2           0   \n",
      "10   2013  [OR%, PIE, TOVR, eFG%]       36          67      53.7           0   \n",
      "11   2014  [OR%, PIE, TOVR, eFG%]       42          67      62.7           0   \n",
      "12   2015  [OR%, PIE, TOVR, eFG%]       43          67      64.2           0   \n",
      "13   2016  [OR%, PIE, TOVR, eFG%]       50          67      74.6           0   \n",
      "14   2017  [OR%, PIE, TOVR, eFG%]       43          67      64.2           0   \n",
      "15   2018  [OR%, PIE, TOVR, eFG%]       45          67      67.2           0   \n",
      "\n",
      "       Score        C  \n",
      "0  -0.111664  21.5443  \n",
      "1  -0.123189  5.99484  \n",
      "2  -0.117523  21.5443  \n",
      "3   -0.12074  21.5443  \n",
      "4  -0.119167  77.4264  \n",
      "5  -0.112264  5.99484  \n",
      "6  -0.114251  21.5443  \n",
      "7  -0.111206  21.5443  \n",
      "8  -0.119028  5.99484  \n",
      "9  -0.117532  5.99484  \n",
      "10  -0.11039  5.99484  \n",
      "11 -0.113801  77.4264  \n",
      "12 -0.111971  77.4264  \n",
      "13 -0.113623  21.5443  \n",
      "14 -0.110205  77.4264  \n",
      "15 -0.114529  5.99484  \n"
     ]
    }
   ],
   "source": [
    "# Train and evalute models for different seasons\n",
    "for model in ['LinReg','LogReg','DecTree','SVC']: # 'LinReg','LogReg','DecTree','SVC'\n",
    "    df_stats = pd.DataFrame(columns=['Season', 'Columns', 'WinsPred', 'GamesPlayed', 'WinsPred%', 'Coef weight', 'Score', 'C'])\n",
    "    counter = 0\n",
    "    columns = ['OR%', 'PIE', 'TOVR', 'eFG%'] # PIE', 'OR%', 'TOVR', 'eFG%'\n",
    "    for year in range(2003,2019):\n",
    "        print model\n",
    "        df_final = load_and_construct_features_in_df(year)   \n",
    "        year, columns, numOfWins, gamesPlayed, numOfWinsPerc, score, C, clf= main(df_final, columns=columns, year=year, model=model)\n",
    "        df_stats.at[counter,'Season'] = year\n",
    "        df_stats.at[counter,'Columns'] = columns\n",
    "        df_stats.at[counter,'WinsPred'] = numOfWins\n",
    "        df_stats.at[counter,'GamesPlayed'] = gamesPlayed\n",
    "        df_stats.at[counter,'WinsPred%'] = numOfWinsPerc\n",
    "        df_stats.at[counter,'Score'] = score\n",
    "        df_stats.at[counter,'C'] = C\n",
    "        if model == 'SVC':\n",
    "            df_stats.at[counter,'Coef weight'] = 0\n",
    "        elif model == 'DecTree':\n",
    "            df_stats.at[counter,'Coef weight'] = clf.best_estimator_.feature_importances_ \n",
    "        elif model == 'LogReg':\n",
    "            df_stats.at[counter,'Coef weight'] = clf.best_estimator_.coef_[0]\n",
    "        elif model == 'LinReg':\n",
    "            df_stats.at[counter,'Coef weight'] = clf.best_estimator_.coef_\n",
    "        counter += 1\n",
    "    print df_stats\n",
    "    df_stats.to_csv('ModelAnalysis_of_'+str(model)+'_2003_2018_4features.csv', index=False)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Compare models from years 2003 to 2018 based on 6 features ['FTAR', 'OR%', 'PIE', 'TDR%', 'TOVR', 'eFG%']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#plt.figure(figsize=(8,8))\n",
    "Lin_reg_2003_2018 = pd.read_csv('ModelAnalysis_of_LinReg_2003_2018_6features.csv')\n",
    "Lin_reg_2003_2018['#Features'] = Lin_reg_2003_2018['Columns'].apply(lambda x: len(x.strip('[]').split(',')))\n",
    "LogReg_2003_2018 = pd.read_csv('ModelAnalysis_of_LogReg_2003_2018_6features.csv')\n",
    "LogReg_2003_2018['#Features'] = LogReg_2003_2018['Columns'].apply(lambda x: len(x.strip('[]').split(',')))\n",
    "SVCReg_2003_2018 = pd.read_csv('ModelAnalysis_of_SVCReg_2003_2018_6features.csv')\n",
    "SVCReg_2003_2018['#Features'] = SVCReg_2003_2018['Columns'].apply(lambda x: len(x.strip('[]').split(',')))\n",
    "TreeReg_2003_2018 = pd.read_csv('ModelAnalysis_of_DecTree_2003_2018_6features.csv')\n",
    "TreeReg_2003_2018['#Features'] = TreeReg_2003_2018['Columns'].apply(lambda x: len(x.strip('[]').split(',')))\n",
    "\n",
    "plt.plot(LogReg_2003_2018['Season'], LogReg_2003_2018['WinsPred%'], zorder=4, label='LogReg, AVG='+str(round(LogReg_2003_2018['WinsPred%'].mean(),2))+'%')\n",
    "plt.plot(SVCReg_2003_2018['Season'], SVCReg_2003_2018['WinsPred%'], zorder=3, label='SVC, AVG='+str(round(SVCReg_2003_2018['WinsPred%'].mean(),2))+'%')\n",
    "plt.plot(Lin_reg_2003_2018['Season'], Lin_reg_2003_2018['WinsPred%'], zorder=2, label='LinReg, AVG='+str(round(Lin_reg_2003_2018['WinsPred%'].mean(),2))+'%')\n",
    "plt.plot(TreeReg_2003_2018['Season'], TreeReg_2003_2018['WinsPred%'], zorder=1, label='Tree, AVG='+str(round(TreeReg_2003_2018['WinsPred%'].mean(),2))+'%')\n",
    "plt.xlabel('Season')\n",
    "plt.ylabel('% of correct predictions')\n",
    "plt.legend()\n",
    "plt.savefig('ModelsCompared6features.png')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Compare models from years 2003 to 2018 based on 5 features ['FTAR', 'PIE', 'TDR%', 'TOVR', 'eFG%']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#plt.figure(figsize=(10,10))\n",
    "Lin_reg_2003_2018noOR = pd.read_csv('ModelAnalysis_of_LinReg_2003_2018_5features_noOR%.csv')\n",
    "Lin_reg_2003_2018noOR['#Features'] = Lin_reg_2003_2018noOR['Columns'].apply(lambda x: len(x.strip('[]').split(',')))\n",
    "LogReg_2003_2018noOR = pd.read_csv('ModelAnalysis_of_LogReg_2003_2018_5features_noOR%.csv')\n",
    "LogReg_2003_2018noOR['#Features'] = LogReg_2003_2018noOR['Columns'].apply(lambda x: len(x.strip('[]').split(',')))\n",
    "SVCReg_2003_2018noOR = pd.read_csv('ModelAnalysis_of_SVC_2003_2018_5features_noOR%.csv')\n",
    "SVCReg_2003_2018noOR['#Features'] = SVCReg_2003_2018noOR['Columns'].apply(lambda x: len(x.strip('[]').split(',')))\n",
    "TreeReg_2003_2018noOR = pd.read_csv('ModelAnalysis_of_DecTree_2003_2018_5features_noOR%.csv')\n",
    "TreeReg_2003_2018noOR['#Features'] = TreeReg_2003_2018noOR['Columns'].apply(lambda x: len(x.strip('[]').split(',')))\n",
    "\n",
    "plt.plot(LogReg_2003_2018noOR['Season'], LogReg_2003_2018noOR['WinsPred%'], zorder=4, label='LogReg, AVG='+str(round(LogReg_2003_2018noOR['WinsPred%'].mean(),2))+'%')\n",
    "plt.plot(SVCReg_2003_2018noOR['Season'], SVCReg_2003_2018noOR['WinsPred%'], zorder=3, label='SVC, AVG='+str(round(SVCReg_2003_2018noOR['WinsPred%'].mean(),2))+'%')\n",
    "plt.plot(Lin_reg_2003_2018noOR['Season'], Lin_reg_2003_2018noOR['WinsPred%'], zorder=2, label='LinReg, AVG='+str(round(Lin_reg_2003_2018noOR['WinsPred%'].mean(),2))+'%')\n",
    "plt.plot(TreeReg_2003_2018noOR['Season'], TreeReg_2003_2018noOR['WinsPred%'], zorder=1, label='Tree, AVG='+str(round(TreeReg_2003_2018noOR['WinsPred%'].mean(),2))+'%')\n",
    "plt.xlabel('Season')\n",
    "plt.ylabel('% of correct predictions')\n",
    "plt.legend()\n",
    "plt.savefig('ModelsCompared5featuresNoOR%.png')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Compare models from years 2003 to 2018 based on 4 features ['OR%', 'PIE', 'TOVR', 'eFG%']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "Lin_reg_2003_2018_4feat = pd.read_csv('ModelAnalysis_of_LinReg_2003_2018_4features.csv')\n",
    "Lin_reg_2003_2018_4feat['#Features'] = Lin_reg_2003_2018_4feat['Columns'].apply(lambda x: len(x.strip('[]').split(',')))\n",
    "LogReg_2003_2018_4feat = pd.read_csv('ModelAnalysis_of_LogReg_2003_2018_4features.csv')\n",
    "LogReg_2003_2018_4feat['#Features'] = LogReg_2003_2018_4feat['Columns'].apply(lambda x: len(x.strip('[]').split(',')))\n",
    "SVCReg_2003_2018_4feat = pd.read_csv('ModelAnalysis_of_SVC_2003_2018_4features.csv')\n",
    "SVCReg_2003_2018_4feat['#Features'] = SVCReg_2003_2018_4feat['Columns'].apply(lambda x: len(x.strip('[]').split(',')))\n",
    "TreeReg_2003_2018_4feat = pd.read_csv('ModelAnalysis_of_DecTree_2003_2018_4features.csv')\n",
    "TreeReg_2003_2018_4feat['#Features'] = TreeReg_2003_2018_4feat['Columns'].apply(lambda x: len(x.strip('[]').split(',')))\n",
    "\n",
    "plt.plot(LogReg_2003_2018_4feat['Season'], LogReg_2003_2018_4feat['WinsPred%'], zorder=4, label='LogReg, AVG='+str(round(LogReg_2003_2018_4feat['WinsPred%'].mean(),2))+'%')\n",
    "plt.plot(SVCReg_2003_2018_4feat['Season'], SVCReg_2003_2018_4feat['WinsPred%'], zorder=3, label='SVC, AVG='+str(round(SVCReg_2003_2018_4feat['WinsPred%'].mean(),2))+'%')\n",
    "plt.plot(Lin_reg_2003_2018_4feat['Season'], Lin_reg_2003_2018_4feat['WinsPred%'], zorder=2, label='LinReg, AVG='+str(round(Lin_reg_2003_2018_4feat['WinsPred%'].mean(),2))+'%')\n",
    "plt.plot(TreeReg_2003_2018_4feat['Season'], TreeReg_2003_2018_4feat['WinsPred%'], zorder=1, label='Tree, AVG='+str(round(TreeReg_2003_2018_4feat['WinsPred%'].mean(),2))+'%')\n",
    "plt.xlabel('Season')\n",
    "plt.ylabel('% of correct predictions')\n",
    "plt.legend()\n",
    "plt.savefig('ModelsCompared4features.png')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Takeaways\n",
    "#### + Feature visualization reveals correlations\n",
    "#### + Data over the years can be noisy and fluctuate\n",
    "#### + Models can perform similarly\n",
    "#### + Features matter more than models!\n",
    "#### + Reducing the number of features can result in better averaged results.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Context of results\n",
    "- Lower bound: Coin toss 50%\n",
    "- Our results: ≈68% (averaged)\n",
    "- Upper glass sealing**: ≈74-78% for sports predictions in general\n",
    "                        (NBA ≈ 72-74%)\n",
    "                        (Experts ≈ 68%)\n",
    "\n",
    "Room for improvement:\n",
    "- Use expert ratings/ranking, elo scoring system, etc…\n",
    "\n",
    "\n",
    "\n",
    "**A. Zimmermann, S. Moorthy, Z. Shi, Predicting college basketball match outcomes using machine learning techniques: some results and lessons learned.  [CoRRabs/1310.3607](https://dblp.org/db/journals/corr/corr1310.html), (2013).\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.15"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}