nba-allstart-sl.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In this assignment I chose to perform my supervised learning to predict the NBA All Stars for the 2019-2020 season. All data was pulled from `https://www.basketball-reference.com/`. I used the season data from the 2016-2017, 2017-2018, and 2018-2019 seasons along with the chosen all star team members from those respective seasons. \n",
    "\n",
    "I chose 2 types of supervised classifier models (knn and random forest) to perform the classifications. Both performed with high accuracy, but had a hard time classifying the all stars with a high accuracy, given that the dataset was so imbalanced (far more non all stars than all stars).\n",
    "\n",
    "In the test sets, the notable differences between actual and predicted all stars included:\n",
    "\n",
    "Actual but not Predicted (model says they shouldn't have been all stars)\n",
    "- Giannis Antetokounmpo, 2016\n",
    "- Kyle Lowry, 2018\n",
    "- Klay Thompson, 2016\n",
    "- LaMarcus Aldridge, 2017\n",
    "- Dirk Nowitzki, 2018\n",
    "- Khris Middleton, 2018\n",
    "- DeAndre Jordan, 2016\n",
    "\n",
    "Predicted but not Actual (model says they should've been all stars, but were not)\n",
    "- Marc Gasol, 2017\n",
    "- CJ McCollum, 2016\n",
    "\n",
    "Also, using the dataset from the 2019-2020 season so far, the projected all stars include\n",
    "- Pascal Siakam\n",
    "- LeBron James\n",
    "- Kawhi Leonard\n",
    "- Bradley Beal\n",
    "- Buddy Hield\n",
    "- Paul George\n",
    "- Karl-Anthony Towns\n",
    "- Donovan Mitchell\n",
    "- Brandon Ingram\n",
    "- Kyrie Irving\n",
    "- Andrew Wiggins\n",
    "- Anthony Davis\n",
    "- Damian Lillard\n",
    "- CJ McCollum\n",
    "- Giannis Antetokounmpo\n",
    "- James Harden\n",
    "- DeMar DeRozan"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.preprocessing import scale\n",
    "from sklearn import metrics\n",
    "from matplotlib import pyplot as plt\n",
    "\n",
    "pd.set_option('display.max_columns', 999)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Data Preparation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "team_dict = {'Atlanta Hawks': 'ATL', 'Boston Celtics': 'BOS', 'Brooklyn Nets': 'BRK', \n",
    "             'Charlotte Hornets': 'CHO', 'Chicago Bulls': 'CHI', 'Cleveland Cavaliers': 'CLE', \n",
    "             'Dallas Mavericks': 'DAL', 'Denver Nuggets': 'DEN', \n",
    "             'Detroit Pistons': 'DET', 'Golden State Warriors': 'GSW', \n",
    "             'Houston Rockets': 'HOU', 'Indiana Pacers': 'IND', 'Los Angeles Clippers': 'LAC', \n",
    "             'Los Angeles Lakers': 'LAL', 'Memphis Grizzlies': 'MEM', 'Miami Heat': 'MIA', \n",
    "             'Milwaukee Bucks': 'MIL', 'Minnesota Timberwolves': 'MIN', 'New Orleans Pelicans': 'NOP', \n",
    "             'New York Knicks': 'NYK', 'Oklahoma City Thunder': 'OKC', 'Orlando Magic': 'ORL', \n",
    "             'Philadelphia 76ers': 'PHI', 'Phoenix Suns': 'PHO', 'Portland Trail Blazers': 'POR', \n",
    "             'Sacramento Kings': 'SAC', 'San Antonio Spurs': 'SAS', 'Toronto Raptors': 'TOR', \n",
    "             'Utah Jazz': 'UTA', 'Washington Wizards': 'WAS'}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/sam/anaconda3/lib/python3.7/site-packages/ipykernel_launcher.py:4: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  after removing the cwd from sys.path.\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>Team</th>\n",
       "      <th>Overall</th>\n",
       "      <th>win%</th>\n",
       "      <th>Tm</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Golden State Warriors</td>\n",
       "      <td>67-15</td>\n",
       "      <td>0.817073</td>\n",
       "      <td>GSW</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>San Antonio Spurs</td>\n",
       "      <td>61-21</td>\n",
       "      <td>0.743902</td>\n",
       "      <td>SAS</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Houston Rockets</td>\n",
       "      <td>55-27</td>\n",
       "      <td>0.670732</td>\n",
       "      <td>HOU</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Boston Celtics</td>\n",
       "      <td>53-29</td>\n",
       "      <td>0.646341</td>\n",
       "      <td>BOS</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Cleveland Cavaliers</td>\n",
       "      <td>51-31</td>\n",
       "      <td>0.621951</td>\n",
       "      <td>CLE</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                    Team Overall      win%   Tm\n",
       "0  Golden State Warriors   67-15  0.817073  GSW\n",
       "1      San Antonio Spurs   61-21  0.743902  SAS\n",
       "2        Houston Rockets   55-27  0.670732  HOU\n",
       "3         Boston Celtics   53-29  0.646341  BOS\n",
       "4    Cleveland Cavaliers   51-31  0.621951  CLE"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "w_2016 = pd.read_csv('data/2016_wins.csv')\n",
    "w_2016['win%'] = w_2016.Overall.apply(lambda x: float(x.split('-')[0])/(float(x.split('-')[0])+float(x.split('-')[1])))\n",
    "w2016 = w_2016[['Team','Overall','win%']]\n",
    "w2016['Tm'] = w2016.Team.apply(lambda x: team_dict[x])\n",
    "w2016.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/sam/anaconda3/lib/python3.7/site-packages/ipykernel_launcher.py:9: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  if __name__ == '__main__':\n",
      "/home/sam/anaconda3/lib/python3.7/site-packages/ipykernel_launcher.py:28: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "/home/sam/anaconda3/lib/python3.7/site-packages/ipykernel_launcher.py:46: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "/home/sam/anaconda3/lib/python3.7/site-packages/ipykernel_launcher.py:86: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "/home/sam/anaconda3/lib/python3.7/site-packages/ipykernel_launcher.py:129: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "/home/sam/anaconda3/lib/python3.7/site-packages/ipykernel_launcher.py:142: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\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>Rk</th>\n",
       "      <th>Player</th>\n",
       "      <th>Pos</th>\n",
       "      <th>Age</th>\n",
       "      <th>Tm</th>\n",
       "      <th>G</th>\n",
       "      <th>GS</th>\n",
       "      <th>MP</th>\n",
       "      <th>FG</th>\n",
       "      <th>FGA</th>\n",
       "      <th>FG%</th>\n",
       "      <th>3P</th>\n",
       "      <th>3PA</th>\n",
       "      <th>3P%</th>\n",
       "      <th>2P</th>\n",
       "      <th>2PA</th>\n",
       "      <th>2P%</th>\n",
       "      <th>eFG%</th>\n",
       "      <th>FT</th>\n",
       "      <th>FTA</th>\n",
       "      <th>FT%</th>\n",
       "      <th>ORB</th>\n",
       "      <th>DRB</th>\n",
       "      <th>TRB</th>\n",
       "      <th>AST</th>\n",
       "      <th>STL</th>\n",
       "      <th>BLK</th>\n",
       "      <th>TOV</th>\n",
       "      <th>PF</th>\n",
       "      <th>PTS</th>\n",
       "      <th>year</th>\n",
       "      <th>name</th>\n",
       "      <th>allstar</th>\n",
       "      <th>Team</th>\n",
       "      <th>Overall</th>\n",
       "      <th>win%</th>\n",
       "      <th>GS%</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>418</th>\n",
       "      <td>243</td>\n",
       "      <td>Kyrie Irving\\irvinky01</td>\n",
       "      <td>PG</td>\n",
       "      <td>28</td>\n",
       "      <td>BRK</td>\n",
       "      <td>54</td>\n",
       "      <td>54</td>\n",
       "      <td>34.9</td>\n",
       "      <td>10.2</td>\n",
       "      <td>20.1</td>\n",
       "      <td>0.506</td>\n",
       "      <td>2.8</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0.402</td>\n",
       "      <td>7.4</td>\n",
       "      <td>13.1</td>\n",
       "      <td>0.561</td>\n",
       "      <td>0.576</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.0</td>\n",
       "      <td>0.922</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3.8</td>\n",
       "      <td>4.8</td>\n",
       "      <td>6.0</td>\n",
       "      <td>1.4</td>\n",
       "      <td>0.7</td>\n",
       "      <td>2.4</td>\n",
       "      <td>2.6</td>\n",
       "      <td>26.9</td>\n",
       "      <td>2020</td>\n",
       "      <td>Kyrie Irving</td>\n",
       "      <td>1</td>\n",
       "      <td>Brooklyn Nets</td>\n",
       "      <td>48-24</td>\n",
       "      <td>0.666667</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>429</th>\n",
       "      <td>74</td>\n",
       "      <td>Jaylen Brown\\brownja02</td>\n",
       "      <td>SG</td>\n",
       "      <td>24</td>\n",
       "      <td>BOS</td>\n",
       "      <td>58</td>\n",
       "      <td>58</td>\n",
       "      <td>34.5</td>\n",
       "      <td>9.3</td>\n",
       "      <td>19.2</td>\n",
       "      <td>0.484</td>\n",
       "      <td>2.8</td>\n",
       "      <td>7.1</td>\n",
       "      <td>0.397</td>\n",
       "      <td>6.5</td>\n",
       "      <td>12.1</td>\n",
       "      <td>0.536</td>\n",
       "      <td>0.558</td>\n",
       "      <td>3.3</td>\n",
       "      <td>4.3</td>\n",
       "      <td>0.764</td>\n",
       "      <td>1.2</td>\n",
       "      <td>4.8</td>\n",
       "      <td>6.0</td>\n",
       "      <td>3.4</td>\n",
       "      <td>1.2</td>\n",
       "      <td>0.6</td>\n",
       "      <td>2.7</td>\n",
       "      <td>2.9</td>\n",
       "      <td>24.7</td>\n",
       "      <td>2020</td>\n",
       "      <td>Jaylen Brown</td>\n",
       "      <td>1</td>\n",
       "      <td>Boston Celtics</td>\n",
       "      <td>36-36</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>437</th>\n",
       "      <td>473</td>\n",
       "      <td>Jayson Tatum\\tatumja01</td>\n",
       "      <td>SF</td>\n",
       "      <td>22</td>\n",
       "      <td>BOS</td>\n",
       "      <td>64</td>\n",
       "      <td>64</td>\n",
       "      <td>35.8</td>\n",
       "      <td>9.5</td>\n",
       "      <td>20.6</td>\n",
       "      <td>0.459</td>\n",
       "      <td>2.9</td>\n",
       "      <td>7.6</td>\n",
       "      <td>0.386</td>\n",
       "      <td>6.5</td>\n",
       "      <td>13.0</td>\n",
       "      <td>0.502</td>\n",
       "      <td>0.530</td>\n",
       "      <td>4.6</td>\n",
       "      <td>5.3</td>\n",
       "      <td>0.868</td>\n",
       "      <td>0.8</td>\n",
       "      <td>6.6</td>\n",
       "      <td>7.4</td>\n",
       "      <td>4.3</td>\n",
       "      <td>1.2</td>\n",
       "      <td>0.5</td>\n",
       "      <td>2.7</td>\n",
       "      <td>1.9</td>\n",
       "      <td>26.4</td>\n",
       "      <td>2020</td>\n",
       "      <td>Jayson Tatum</td>\n",
       "      <td>1</td>\n",
       "      <td>Boston Celtics</td>\n",
       "      <td>36-36</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>444</th>\n",
       "      <td>146</td>\n",
       "      <td>Joel Embiid\\embiijo01</td>\n",
       "      <td>C</td>\n",
       "      <td>26</td>\n",
       "      <td>PHI</td>\n",
       "      <td>51</td>\n",
       "      <td>51</td>\n",
       "      <td>31.1</td>\n",
       "      <td>9.0</td>\n",
       "      <td>17.6</td>\n",
       "      <td>0.513</td>\n",
       "      <td>1.1</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.377</td>\n",
       "      <td>7.9</td>\n",
       "      <td>14.6</td>\n",
       "      <td>0.541</td>\n",
       "      <td>0.545</td>\n",
       "      <td>9.2</td>\n",
       "      <td>10.7</td>\n",
       "      <td>0.859</td>\n",
       "      <td>2.2</td>\n",
       "      <td>8.4</td>\n",
       "      <td>10.6</td>\n",
       "      <td>2.8</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.4</td>\n",
       "      <td>3.1</td>\n",
       "      <td>2.4</td>\n",
       "      <td>28.5</td>\n",
       "      <td>2020</td>\n",
       "      <td>Joel Embiid</td>\n",
       "      <td>1</td>\n",
       "      <td>Philadelphia 76ers</td>\n",
       "      <td>49-23</td>\n",
       "      <td>0.680556</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>457</th>\n",
       "      <td>457</td>\n",
       "      <td>Ben Simmons\\simmobe01</td>\n",
       "      <td>PG</td>\n",
       "      <td>24</td>\n",
       "      <td>PHI</td>\n",
       "      <td>58</td>\n",
       "      <td>58</td>\n",
       "      <td>32.4</td>\n",
       "      <td>5.6</td>\n",
       "      <td>10.1</td>\n",
       "      <td>0.557</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.2</td>\n",
       "      <td>0.300</td>\n",
       "      <td>5.6</td>\n",
       "      <td>9.9</td>\n",
       "      <td>0.562</td>\n",
       "      <td>0.560</td>\n",
       "      <td>3.0</td>\n",
       "      <td>4.9</td>\n",
       "      <td>0.613</td>\n",
       "      <td>1.6</td>\n",
       "      <td>5.6</td>\n",
       "      <td>7.2</td>\n",
       "      <td>6.9</td>\n",
       "      <td>1.6</td>\n",
       "      <td>0.6</td>\n",
       "      <td>3.0</td>\n",
       "      <td>2.9</td>\n",
       "      <td>14.3</td>\n",
       "      <td>2020</td>\n",
       "      <td>Ben Simmons</td>\n",
       "      <td>1</td>\n",
       "      <td>Philadelphia 76ers</td>\n",
       "      <td>49-23</td>\n",
       "      <td>0.680556</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      Rk                  Player Pos  Age   Tm   G  GS    MP    FG   FGA  \\\n",
       "418  243  Kyrie Irving\\irvinky01  PG   28  BRK  54  54  34.9  10.2  20.1   \n",
       "429   74  Jaylen Brown\\brownja02  SG   24  BOS  58  58  34.5   9.3  19.2   \n",
       "437  473  Jayson Tatum\\tatumja01  SF   22  BOS  64  64  35.8   9.5  20.6   \n",
       "444  146   Joel Embiid\\embiijo01   C   26  PHI  51  51  31.1   9.0  17.6   \n",
       "457  457   Ben Simmons\\simmobe01  PG   24  PHI  58  58  32.4   5.6  10.1   \n",
       "\n",
       "       FG%   3P  3PA    3P%   2P   2PA    2P%   eFG%   FT   FTA    FT%  ORB  \\\n",
       "418  0.506  2.8  7.0  0.402  7.4  13.1  0.561  0.576  3.7   4.0  0.922  1.0   \n",
       "429  0.484  2.8  7.1  0.397  6.5  12.1  0.536  0.558  3.3   4.3  0.764  1.2   \n",
       "437  0.459  2.9  7.6  0.386  6.5  13.0  0.502  0.530  4.6   5.3  0.868  0.8   \n",
       "444  0.513  1.1  3.0  0.377  7.9  14.6  0.541  0.545  9.2  10.7  0.859  2.2   \n",
       "457  0.557  0.1  0.2  0.300  5.6   9.9  0.562  0.560  3.0   4.9  0.613  1.6   \n",
       "\n",
       "     DRB   TRB  AST  STL  BLK  TOV   PF   PTS  year          name  allstar  \\\n",
       "418  3.8   4.8  6.0  1.4  0.7  2.4  2.6  26.9  2020  Kyrie Irving        1   \n",
       "429  4.8   6.0  3.4  1.2  0.6  2.7  2.9  24.7  2020  Jaylen Brown        1   \n",
       "437  6.6   7.4  4.3  1.2  0.5  2.7  1.9  26.4  2020  Jayson Tatum        1   \n",
       "444  8.4  10.6  2.8  1.0  1.4  3.1  2.4  28.5  2020   Joel Embiid        1   \n",
       "457  5.6   7.2  6.9  1.6  0.6  3.0  2.9  14.3  2020   Ben Simmons        1   \n",
       "\n",
       "                   Team Overall      win%  GS%  \n",
       "418       Brooklyn Nets   48-24  0.666667  1.0  \n",
       "429      Boston Celtics   36-36  0.500000  1.0  \n",
       "437      Boston Celtics   36-36  0.500000  1.0  \n",
       "444  Philadelphia 76ers   49-23  0.680556  1.0  \n",
       "457  Philadelphia 76ers   49-23  0.680556  1.0  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "as2016 = pd.read_csv('data/AS_2016.csv')\n",
    "as2016['year'] = 2016\n",
    "as2016['name'] = as2016['Starters'].apply(lambda x: x.split(\"\\\\\")[0])\n",
    "as_2016 = as2016['name'].tolist()\n",
    "\n",
    "w_2016 = pd.read_csv('data/2016_wins.csv')\n",
    "w_2016['win%'] = w_2016.Overall.apply(lambda x: float(x.split('-')[0])/(float(x.split('-')[0])+float(x.split('-')[1])))\n",
    "w2016 = w_2016[['Team','Overall','win%']]\n",
    "w2016['Tm'] = w2016.Team.apply(lambda x: team_dict[x])\n",
    "\n",
    "d2016 = pd.read_csv('data/2016.csv')\n",
    "d2016['year'] = 2016\n",
    "d2016['name'] = d2016['Player'].apply(lambda x: x.split(\"\\\\\")[0])\n",
    "d2016['allstar'] = d2016['name'].apply(lambda x: 1 if x in as_2016 else 0)\n",
    "mt2016 = d2016[d2016['Tm']=='TOT']['name'].tolist()\n",
    "d2016.drop(d2016[(d2016['name'].isin(mt2016)) & (d2016['Tm']!='TOT')].index , inplace=True)\n",
    "d2016 = d2016.merge(w2016, on='Tm')\n",
    "\n",
    "\n",
    "as2017 = pd.read_csv('data/AS_2017.csv')\n",
    "as2017['year'] = 2017\n",
    "as2017['name'] = as2017['Starters'].apply(lambda x: x.split(\"\\\\\")[0])\n",
    "as_2017 = as2017['name'].tolist()\n",
    "\n",
    "w_2017 = pd.read_csv('data/2017_wins.csv')\n",
    "w_2017['win%'] = w_2017.Overall.apply(lambda x: float(x.split('-')[0])/(float(x.split('-')[0])+float(x.split('-')[1])))\n",
    "w2017 = w_2017[['Team','Overall','win%']]\n",
    "w2017['Tm'] = w2017.Team.apply(lambda x: team_dict[x])\n",
    "\n",
    "d2017 = pd.read_csv('data/2017.csv')\n",
    "d2017['year'] = 2017\n",
    "d2017['name'] = d2017['Player'].apply(lambda x: x.split(\"\\\\\")[0])\n",
    "d2017['allstar'] = d2017['name'].apply(lambda x: 1 if x in as_2017 else 0)\n",
    "mt2017 = d2017[d2017['Tm']=='TOT']['name'].tolist()\n",
    "d2017.drop(d2017[(d2017['name'].isin(mt2017)) & (d2017['Tm']!='TOT')].index , inplace=True)\n",
    "d2017 = d2017.merge(w2017, on='Tm')\n",
    "\n",
    "as2018 = pd.read_csv('data/AS_2018.csv')\n",
    "as2018['year'] = 2018\n",
    "as2018['name'] = as2018['Starters'].apply(lambda x: x.split(\"\\\\\")[0])\n",
    "as_2018 = as2018['name'].tolist()\n",
    "\n",
    "w_2018 = pd.read_csv('data/2018_wins.csv')\n",
    "w_2018['win%'] = w_2018.Overall.apply(lambda x: float(x.split('-')[0])/(float(x.split('-')[0])+float(x.split('-')[1])))\n",
    "w2018 = w_2018[['Team','Overall','win%']]\n",
    "w2018['Tm'] = w2018.Team.apply(lambda x: team_dict[x])\n",
    "\n",
    "d2018 = pd.read_csv('data/2018.csv')\n",
    "d2018['year'] = 2018\n",
    "d2018['name'] = d2018['Player'].apply(lambda x: x.split(\"\\\\\")[0])\n",
    "d2018['allstar'] = d2018['name'].apply(lambda x: 1 if x in as_2018 else 0)\n",
    "mt2018 = d2018[d2018['Tm']=='TOT']['name'].tolist()\n",
    "d2018.drop(d2018[(d2018['name'].isin(mt2018)) & (d2018['Tm']!='TOT')].index , inplace=True)\n",
    "d2018 = d2018.merge(w2018, on='Tm')\n",
    "\n",
    "as_2019 = [\n",
    "   'Bam Adebayo', \n",
    "   'Giannis Antetokounmpo', \n",
    "   'Jimmy Butler', \n",
    "   'Joel Embiid', \n",
    "   'Rudy Gobert', \n",
    "   'Brandon Ingram', \n",
    "   'Kyle Lowry', \n",
    "   'Khris Middleton', \n",
    "   'Donovan Mitchell', \n",
    "   'Pascal Siakam', \n",
    "   'Kemba Walker', \n",
    "   'Trae Young', \n",
    "   'Devin Booker', \n",
    "   'Anthony Davis', \n",
    "   'Luka Dončić', \n",
    "   'James Harden', \n",
    "   'LeBron James', \n",
    "   'Nikola Jokić', \n",
    "   'Kawhi Leonard', \n",
    "   'Damian Lillard', \n",
    "   'Chris Paul', \n",
    "   'Domantas Sabonis', \n",
    "   'Ben Simmons', \n",
    "   'Jayson Tatum'\n",
    "]\n",
    "\n",
    "w_2019 = pd.read_csv('data/2019_wins.csv')\n",
    "w_2019['win%'] = w_2019.Overall.apply(lambda x: float(x.split('-')[0])/(float(x.split('-')[0])+float(x.split('-')[1])))\n",
    "w2019 = w_2019[['Team','Overall','win%']]\n",
    "w2019['Tm'] = w2019.Team.apply(lambda x: team_dict[x])\n",
    "\n",
    "d2019 = pd.read_csv('data/2019.csv')\n",
    "d2019['year'] = 2019\n",
    "d2019['name'] = d2019['Player'].apply(lambda x: x.split(\"\\\\\")[0])\n",
    "d2019['allstar'] = d2019['name'].apply(lambda x: 1 if x in as_2019 else 0)\n",
    "mt2019 = d2019[d2019['Tm']=='TOT']['name'].tolist()\n",
    "d2019.drop(d2019[(d2019['name'].isin(mt2019)) & (d2019['Tm']!='TOT')].index , inplace=True)\n",
    "d2019 = d2019.merge(w2019, on='Tm')\n",
    "\n",
    "as_2020 = [\n",
    "  'Bradley Beal', \n",
    "  'Devin Booker', \n",
    "  'Mike Conley', \n",
    "  'Anthony Davis', \n",
    "  'Kevin Durant', \n",
    "  'Joel Embiid', \n",
    "  'James Harden', \n",
    "  'Kyrie Irving', \n",
    "  'Zach LaVine', \n",
    "  'Kawhi Leonard', \n",
    "  'Donovan Mitchell', \n",
    "  'Julius Randle', \n",
    "  'Jayson Tatum', \n",
    "  'Nikola Vučević', \n",
    "  'Zion Williamson', \n",
    "  'Giannis Antetokounmpo', \n",
    "  'Jaylen Brown', \n",
    "  'Stephen Curry', \n",
    "  'Luka Dončić', \n",
    "  'Paul George', \n",
    "  'Rudy Gobert', \n",
    "  'LeBron James', \n",
    "  'Nikola Jokić', \n",
    "  'Damian Lillard', \n",
    "  'Chris Paul', \n",
    "  'Domantas Sabonis', \n",
    "  'Ben Simmons'\n",
    "]\n",
    "\n",
    "w_2020 = pd.read_csv('data/2020_wins.csv')\n",
    "w_2020['win%'] = w_2020.Overall.apply(lambda x: float(x.split('-')[0])/(float(x.split('-')[0])+float(x.split('-')[1])))\n",
    "w2020 = w_2020[['Team','Overall','win%']]\n",
    "w2020['Tm'] = w2020.Team.apply(lambda x: team_dict[x])\n",
    "\n",
    "d2020 = pd.read_csv('data/2020.csv')\n",
    "d2020['year'] = 2020\n",
    "d2020['name'] = d2020['Player'].apply(lambda x: x.split(\"\\\\\")[0])\n",
    "d2020['allstar'] = d2020['name'].apply(lambda x: 1 if x in as_2020 else 0)\n",
    "mt2020 = d2020[d2020['Tm']=='TOT']['name'].tolist()\n",
    "d2020.drop(d2020[(d2020['name'].isin(mt2020)) & (d2020['Tm']!='TOT')].index , inplace=True)\n",
    "d2020 = d2020.merge(w2020, on='Tm')\n",
    "\n",
    "w_2021 = pd.read_csv('data/2021_wins.csv')\n",
    "w_2021['win%'] = w_2021.Overall.apply(lambda x: float(x.split('-')[0])/(float(x.split('-')[0])+float(x.split('-')[1])))\n",
    "w2021 = w_2021[['Team','Overall','win%']]\n",
    "w2021['Tm'] = w2021.Team.apply(lambda x: team_dict[x])\n",
    "\n",
    "d2021 = pd.read_csv('data/2021.csv')\n",
    "d2021['year'] = 2021\n",
    "d2021['name'] = d2021['Player'].apply(lambda x: x.split(\"\\\\\")[0])\n",
    "d2021.fillna(0, inplace=True)\n",
    "d2021['GS%'] = d2021['GS']/d2021['G']\n",
    "d2021 = d2021.merge(w2021, on='Tm')\n",
    "\n",
    "\n",
    "df = pd.concat([d2016, d2017, d2018, d2019, d2020])\n",
    "df['GS%'] = df['GS']/df['G']\n",
    "# df['GP_percentile']\n",
    "df.fillna(0, inplace=True)\n",
    "\n",
    "# df.groupby('year')['name'].count().head()\n",
    "# df[df['year'] == 2020] \n",
    "df[df['allstar']==1].tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = df.drop(columns=['Rk','Player','Pos','Tm','G','GS','name','year','allstar','Team','Overall'])\n",
    "data_cols = data.columns\n",
    "labels = df['allstar']\n",
    "\n",
    "data_2021 = d2021.drop(columns=['Rk','Player','Pos','Tm','G','GS','name','year','Team','Overall'])\n",
    "\n",
    "train_data, test_data, train_labels, test_labels = train_test_split(data, labels, test_size=0.2, random_state=42, stratify=labels)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### KNN Classifier - Non-Normalized"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAtQAAAG6CAYAAAAoO9FHAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjMsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+AADFEAAAgAElEQVR4nOzdd3yc1ZX/8c9RtSVLLrJsbFkuErbBOC7gSu8hZSEJAUxoBgLO7rLJJhs2vW92UzcbFvIDAqGEAKEkQBKWEkKP3ADbuIIkW5arZMmWLcnq5/fHPDKDkKyRpdHMyN/36zUvZu7zPPeeGdvozNW59zF3R0REREREjkxSrAMQEREREUlkSqhFRERERHpBCbWIiIiISC8ooRYRERER6QUl1CIiIiIivaCEWkRERESkF5RQi0jcMLNaMyuI8Fw3s2O7OLbYzF7r2+j6hpltMbNzo9T3aWa2Kez1VDN7y8wOmNnnzex2M/tWNMYWETmaKaEWkS51TP7MbJGZ7TWzMzo598wgyb2tQ/trZrY4kvHcfYi7l/Y68Bgys2wz+x8z2xp8QSgOXo+M9tju/qq7Tw1r+nfgJXfPcvdb3P1z7v6DaMchInK0UUItIhExs2uA24CPufvLXZxWB1xtZhP7K65oM7OUHpybBrwAnABcAGQDJwNVwLyoBHh4E4B1ve2kJ5+BiMjRSAm1iHTLzG4Efg582N3/fphT9wH3At85TF/XmdmGYKb7WTObEHbsUBmHmeWY2Z/MbL+ZrTCz/+ikjONcM3s36Os2M7P3D2X/a2Y1ZrbRzM4JOzDWzJ4ys+pgBvmGsGPfNbPHzOwBM9sPLDazeWa2Mohlt5n9dxdv72pgPPBJd1/v7m3uXuHuP3D3pzv5LOaZWZGZ7TOznWZ2a5CUYyG/MLOK4D2sMbPpwbGPmtn6oJRju5l9OWg/08y2Bc//BpwF3BrMlE8xs3vN7D/Cxv+4ma0Kxv+7mc0IO7bFzL5iZmuAOiXVIiJdU0ItIt35R+AHwDnuvjKC838IXGxmUzseMLNPAF8HPgXkAq8CD3XRz22EZryPAa4JHh19HJgLzAQuBT4cdmw+UAqMJJTg/8HMRgTHHgK2AWOBTwP/GZ5wAxcBjwHDgN8BvwR+6e7ZQCHwSBcxnws84+61XRzvqBX4YhDjQuAc4J+CY+cDpwNTgjguIzTTDXA3sMTds4DpwN86duzuZxP6fG8KSmneCT9uZicCvwGWADnAHcBTZpYedtrlwMeAYe7eEuF7EhE56iihFpHunAcsBd6O5GR33wXcDny/k8NLgP9y9w1BgvafwKzwWWoAM0sGLga+4+717r4euK+T/n7k7vvcfSvwIjAr7FgF8D/u3uzuvwc2AR8zs3zgVOAr7t7g7quAu4Crwq4tcvcnghnmg0AzcKyZjXT3Wndf2sXbzwF2HubjeR93f8Pdl7p7i7tvIZTUttenNwNZwHGABZ/ZzrBj08ws2933uvubkY4Z5gbgDndf5u6t7n4f0AgsCDvnFncvDz4DERHpghJqEenO5wjNkt7VoaTicH4MfNjMZnZonwD8Migx2AdUAwbkdTgvF0gBysPayvmgXWHP64EhYa+3u7uHvS4jNCM9Fqh29wMdjoXH0HGs6wl9BhuD8pOPdxILhGaQx3Rx7AOCMow/m9muoLzkPwnNVuPufwNuJTRTv9vM7jSz7ODSi4GPAmVm9rKZLYx0zDATgH9r/7MI/jzyCX0+7Tr7zEVEpAMl1CLSnQpCpQinAb+K5AJ3rwL+h1CpSLhyQqUKw8Iegzupy64EWoBxYW35PYw7r8MXgPHAjuAxwsyyOhzbHv4WOryfd939cmAUoS8Lj5lZZidj/pXQF4nOjnXm/wEbgclBOcnXCX3BaB/3Fnc/idAixynAzUH7Cne/KIjnCbouQTmccuCHHf4sMtw9vATHu7pYRETeo4RaRLrl7juAs4ELzOwXEV7234R2uDg+rO124GtmdgKAmQ01s0s6Ga8V+APwXTPLMLPjCC3464lRwOfNLDUY43jgaXcvB/4O/JeZDQoW4l1PqFa6U2Z2pZnlunsboYWXEKp/7ui3hBLVx83sODNLChZXft3MPtrJ+VnAfqA2eI//GDbmXDObb2aphGrJG4BWM0szsyvMbKi7NwfXdxZLd34NfC4Yw8ws08w+1uGLhoiIREAJtYhEJEhEzwY+bWb/FcH5+4GfACPC2v5IaIb34aDEYS3wkS66uAkYSqis47eEFhI29iDkZcBkYA+hhZKfDmbOIbTYbiKh2eo/EqrVfv4wfV0ArDOzWkILFBe5e0PHk9y9kdDCxI3A84SS3eWEyjiWddLvl4HPAAcIJbi/DzuWHbTtJVSSUgX8LDh2FbAl+Aw/B1x5mNg7FSwwvYFQWcleoBhY3NN+REQktNAl1jGIiHTLzH4MHOPune32ISIiEjOaoRaRuBSUTMwIyhHmESrL+GOs4xIREelIG/WLSLzKIlTmMZbQwsifA0/GNCIREZFOqORDRERERKQXVPIhIiIiItILA6bkY+TIkT5x4sRYhyEiIiLSrTfeeGOPu+fGOg7pGwMmoZ44cSIrV66MdRgiIiIi3TKzsljHIH1HJR8iIiIiIr2ghFpEREREpBeUUIuIiIiI9IISahERERGRXlBCLSIiIiLSC0qoRURERER6QQm1iIiIiEgvKKEWEREREekFJdQiIiIS17Zs2cL06dOP6NpvfOMb5OfnM2TIkPe1b926lbPOOovZs2czY8YMnn766W77uuWWWzj++OO54oor2LhxIwsXLiQ9PZ2f/exnPYqpuroaYLKZvWtmz5vZ8M7OM7PxZvacmW0ws/VmNjFoP8fM3jSzVWb2mpkdG7Snm9nvzazYzJaFnZ9qZveZ2dtBX18LG2NL0L7KzFaGtV9iZuvMrM3M5nSI62vBGJvM7MNh7cPM7DEz2xiMszBo/4GZrQnGeM7MxgbtQ83sT2a2Ohjr2qB9lpkVBW1rzOyysDHODt772uA9pQTtNwf9rwqOtZrZiODYF4O+1prZQ2Y2KGh/NeyaHWb2RNB+RTDuGjP7u5nN7PYP1d0HxOOkk05yERERGXg2b97sJ5xwwhFdW1RU5Dt27PDMzMz3td9www3+q1/9yt3d161b5xMmTOi2r6lTp3ppaam7u+/evduXL1/uX//61/2nP/1pj2K6+eabHdjm7gBfBX7sneQ2wEvAecHzIUBG8Pwd4Pjg+T8B94Y9vz14vgj4ffD8M8DDwfMMYAswMXi9BRjZydjHA1ODGOaEtU8DVgPpwCSgBEgOjt0HfDZ4ngYMC55nh13/+bAYv97+3oFcoDq4bgowOWgfC+wEhhGaCC4HpgTHvg9c30ns/wD8LXieB2wGBgevHwEWd3LN48DVwfOTgeHB848Ayzr78wl/aIZaREREEkZpaSmzZ89mxYoVEZ2/YMECxowZ84F2M2P//v0A1NTUMHbs2MP287nPfY7S0lIuvPBCfvGLXzBq1Cjmzp1Lampqj9/Dk08+CVAVvLwP+EQn8U0DUtz9eQB3r3X3+uCwA9nB86HAjuD5RUF/AI8B55iZBednBrO5g4EmYP/hYnT3De6+qZNDFxFKzhvdfTNQDMwzs2zgdODu4Pomd98XPA8fKzOIp/19ZAUxDiGUULe4+zvu/m5w7Q6gglDCnQM0uvs7wfXPAxd3EuPlwENhr1OAwcH7z+C9zwsAM8sCzgaeCMb8u7vvDQ4vBcZ18TG9bwARERGRuLdp0yYWLVrEPffcw6xZs9i0aROXXXZZp+e+9NJLDBs2rMu+vvvd73L++efzv//7v9TV1fHXv/71sGPffvvtPPPMM7z44ouMHDnysOeedtppHDhw4APtP/vZzzj33HPZvXs3QDOAu+80s1GddDMF2GdmfyA0E/xX4Kvu3gp8FnjazA4SSowXBNfkEZrBxd1bzKyGUBL6GKFEeCehhPKL7l4dXOPAc2bmwB3ufudh31xojKVhr7cFbQeBSuCeoETiDeAL7l4HYGY/BK4GaoCzgmtvBZ4ilOBmAZe5e1v4YGY2j9CsdUkQa6qZzXH3lcCngfwO52cAFwA3BZ/DdjP7GbA1iPE5d3+uw3v6JPBCh8S/3fXA/3XzmSihFhERkfhXWVnJRRddxOOPP84JJ5wAwNSpU1m1atUR9ffQQw+xePFi/u3f/o2ioiKuuuoq1q5dS1JS7395/+qrr/a6D0I52mnAbELJ4O+BxYRmgL8IfNTdl5nZzcB/E0qyrZN+HJgHtBIqnxgOvGpmf3X3UuAUd98RJPXPm9lGd3/lMHF1NUYKcCLwL0FcvyRUzvItAHf/BvCNoH77JuA7wIeBVYRmhwuD8V9tT2zNbAzwW+Ca9kTbzBYBvzCzdOA5oKVDLP8AvN7+hSGoT7+I0JeSfcCjZnaluz8Qds3lwF0feKNmZxFKqE89zOcBKKEWERGRBDB06FDy8/N5/fXXDyXUvZmhvvvuu3nmmWcAWLhwIQ0NDezZs4dRozqbLO6Z7maoR48eTU1NTSocShorOulmG/BWkPQSLJhbYGZPATPdfVlw3u+BZ8KuyQe2BeUNQwmVUXwGeMbdm4EKM3sdmAOUBiUVuHuFmf2RUPJ9uIS6fYx24wjNMG8jVBfeHtdjhBLqjh4E/kIoob4W+JGHipWLzWwzcBywPCgh+QvwTXc/NCPu7kWEvmhgZucTmskPt4j3l3ucC2x298rgmj8QqpF+IHidE7znT4Z3YmYzCCXZH3H3KroR1RpqM7sgWAFabGadfaiY2aXBytV1ZvZgWPtPgrYNZnZLUF8jIiIiR5HWNqetzUlLS+OJJ57g/vvv58EHQ+lC+wx1Z4/DJdMA48eP54UXXgBgw4YNNDQ0kJuby/bt2znnnHN6FfOrr77aaUznnnsuABdeeCGESjEArgGe7KSbFcBwM8sNXp8NrAf2AkPNrD2RPA/YEDx/KugPQuUQfwuS1a3A2RaSSahEZKOZZQb1wwTt5wNru3l7TwGLLLSjyCRgMrDc3XcB5WY2NTjvnCBezGxy2PUXAhuD51uD8zCz0YQWQZaaWRrwR+B+d380fPD28phghvorwO1hx4YCZ/D+z3MroS8iGUEueU7Y5wVwCfBnd28I62c88AfgqrB67cPrbtXikT6AZEL1LgWEal9WA9M6nDMZeIv3VlKOCltd+XrQRzJQBJx5uPG0y4eIiMjA0tra5h/95St+w61/ObTLx969e33OnDn+xBNPRNTHzTff7Hl5eW5mnpeX59/5znfcPbSzx8knn+wzZszwmTNn+rPPPuvu7itWrPDzzz+/074mTJjglZWV7u6+c+dOz8vL86ysLB86dKjn5eV5TU1NRDHt2bPHCdU+vwu8AIzwUP4zB7jL38uTzgPWAG8D9wJpQfsng7bVhHbhKAjaBwGPEloouDysfUjQvo5Qkntz0F4Q9LE6OPaNsLE/SWjWuRHYDTwbduwbQY63idAMbnv7LGBlEPMTYfnd44QS9TXAn4A8f28Hj+eC97IWuDJov5JQjfmqsMes4NhPCSXEm4B/9ffnlYsJdjPp0P49Qkn8WkIlJOlhx14CLuhw/l2Evri0j72yY58dHxZc2OeCvQe/6+4fDl5/DcDd/yvsnJ8A77j7XZ1ceyuhmhUj9KuHq9w9/BvF+8yZM8dXrlzZ1WERERFJMM+v380N968kOcl46ctnkj8iI+pj3nrrrYwfP759FjlqzOwNd5/T/ZmSCKJZ8nFopWmgfRVouCnAFDN73cyWmtkFcKg+5kVCq1F3EvpW1GUyLSIiIgPPHS+XMDo7nSSDu1/b3C9j3nTTTVFPpmXgiWZC3dUq0HAphMo+ziRYYWmhu+wcS2hD8XGEkvCzzez0DwxgdqOZrTSzlZWVlX0avIiIiMTOyi3VrCzby+fOKOSiWXk8vGIr1XVNsQ5LpFPRTKi7WgXa8Zwn3b3ZQ5uDbyKUYH8SWOqhTcxrCe3/t6DDtbj7ne4+x93n5ObmdjwsIiIiCeqOV0oZlpHKZXPzufH0Ahqa2/htUVmswxLpVDQT6hWE7lM/KVituYjQytBwTxBs7m1mIwmVgJQSWpF5hpmlmFkqoRWbKvkQERE5ChRX1PL8+t1cvWACGWkpTBmdxTnHjeK+oi0cbGqNdXgiHxC1hNrdWwht3P0soWT4EXdfZ2bfN7P24qRngSozW0+oZvpmD+319xih1aPtK1hXu/ufohWriIiIxI9fv1JKekoSV5888VDbkjMKqa5r4rE3yru+UCRGonpjF3d/Gni6Q9u3w5478KXgEX5OK7AkmrGJiIhI/KnY38Af39rOpXPHMXJI+qH2uROHM3v8MH796mY+M38CyUm6PYXEj6je2EVERESkJ+75+xZa2tr47KkF72s3M5acXsjW6nqeWbsrRtGJdE4JtYiIiMSFAw3NPLC0jI9MH8PEkZkfOH7etNEUjMzk9pdLiNZ9NESOhBJqERERiQsPLy/nQEMLN55e0Onx5CTjs6cV8Pb2GopKq/o5OpGuKaEWERGRmGtqaePu1zazoGAEM/OHdXnep07MY+SQNO54ubQfoxM5PCXUIiIiEnNPrd7Brv0NLDmj8LDnDUpN5tpTJvHyO5Vs2Lm/n6ITOTwl1CIiIhJTbW3Ona+UMHV0FmdO6f5GbVfOn0BGWjJ3vqJZaokPSqhFREQkpl56p4J3dtey5IwCzLrfDm9oRiqXzxvPU6t3sH3fwX6IUOTwlFCLiIhITN3xciljhw7iH2aOjfia606dhAG/eW1z9AITiZASahEREYmZt7buZdnmaq47dRKpyZGnJXnDBnPhzLE8tHwrNfXNUYxQpHtKqEVERCRm7nyllOxBKSyaN77H195wegH1Ta08sKwsCpGJRE4JtYiIiMTE5j11PLNuF1ctnMCQ9JQeX3/8mGzOmJLLPa9vpqG5NQoRikRGCbWIiIjExK9fLSU1KYlrTp54xH0sOaOAPbVN/OHN7X0XmEgPKaEWERGRfld5oJHH3tjGxSflMSpr0BH3s7AghxnjhvLrV0tpbdPtyCU2lFCLiIhIv7u/aAvNrW189rTObzMeKTNjyemFbN5Tx/Prd/VNcCI9pIRaRERE+lVdYwv3F5Vx/rTRFOYO6XV/F0w/hvEjMrj95VLcNUst/U8JtYiIiPSr368op+Zgc7e3GY9UcpJxw2mTWFW+jxVb9vZJnyI9oYRaRERE+k1zaxt3v7aZeRNHcOL44X3W76dPymdEZhp3vFzSZ32KREoJtYiIiPSbv6zZyfZ9B1lyRu9qpzsanJbMNQsn8sLGCt7ZfaBP+xbpjhJqERER6Rfuzu0vlzB51BDOmjqqz/u/euEEBqcmc+crpX3et8jhKKEWERGRfvHKu3vYuOsAN5xeQFKS9Xn/wzPTuGxuPk+u2s7OmoN93r9IV5RQi4iISL+485USRmenc9GssVEb4/pTJ9HmcM/rW6I2hkhHSqhFREQk6t7eVsPrxVVcd8ok0lOSozZO/ogMPvahMTy4bCv7G5qjNo5IOCXUIiIiEnV3vFJCVnoKl88fH/Wxbjy9gNrGFh5ctjXqY4kApMQ6ABERCdmx7yBfemQVe2qbYh2KJJCMtGS+8w8ncNKEvtuCrq9trarn6bd3csPpBWQPSo36eNPzhnLqsSP5zWubufaUiVGdERcBJdQiInGh5mAzi+9Zzs59DZw+JTfW4UgCWVW+j2vvWc7DNy5k2tjsWIfTqbteKyU5ybjulEn9NuaSMwq46u7lPPnWDi6dm99v48rRSQm1iEiMNba0suS3K9m8p477rp3HyceOjHVIkkC27a3n0tuLuPo3y3hkyUIK+uBW3n2pqraRR1aW88nZeYzOHtRv45567EimjcnmjldK+PRJ46Kyq4hIO9VQi4jEUFub8+VH17C0tJqfXTJTybT02LjhGfz2s/NxhyvvWsb2ffG1Xdz9RWU0NLdx4+l9eyOX7pgZS84ooKSyjhc2VvTr2HL0UUItIhJDP352I39avYOvXHAcF83Ki3U4kqAKc4dw//XzONDYwpV3LaPyQGOsQwLgYFMr9xdt4dzjR3HsqKx+H/9jHxpD3rDBuh25RJ0SahGRGLnv71u44+VSrlowgc/18W2Y5ehzwtih3HvtXHbVNHDV3cuoqY/9lnGPvlHO3vpmlpxRGJPxU5KTuOG0Saws28sbZdUxiUGODkqoRURi4Jm1u/jun9Zx3rTRfPfCEzBTfaf03kkTRnDn1SdRWlnH4nuXU9fYErNYWlrb+PWrpZw4fhhzYrgDyaVz8xmWkcodL+t25BI9SqhFRPrZG2V7+cLDbzFz3DBuWTSbZC2Wkj502uRcbrl8Nmu21XDjb1fS0Nwakzj+b+0uyqsPsuSMwph+YcxIS+HqBRN4fsNuiitqYxaHDGxKqEVE+lFpZS2fvW8FY4YO4u5r5jA4TfvjSt+7YPox/OTiGbxeXMW/PPQWza1t/Tq+u3PHKyUUjMzkvONH9+vYnbn65ImkJSdx16uapZboUEItItJPKg80cs09y0ky477r5pEzJD3WIckAdvFJ4/jehSfw/Prd/Ptja2hr834b++8lVazdvp8bTy+Ii+3qRg5J55I54/jDm9up2N8Q63BkAFJCLSLSD+qbWrj+vhVUHmjk7sVzmZCTGeuQ5ChwzckT+fL5U/jjW9v5zlPrcO+fpPqOV0rJzUrnE7PjZ+eaz55aQEtbG/f8fUusQ5EBSAm1iEiUtbS2cdODb7F2ew23Xn4is/KHxTokOYr881nHsuT0An67tIyfPrsp6uOt37GfV96pZPHJExmUGj8lTRNHZvKR6WN4YGkZtTFcrCkDU1QTajO7wMw2mVmxmX21i3MuNbP1ZrbOzB4Max9vZs+Z2Ybg+MRoxioiEg3uzreeXMvfNlbwg09M59xpsa8nlaOLmfHVjxzH5fPG86uXSvh/L0V3T+Y7XykhMy2ZK+dPiOo4R+LG0ws40NDCw8u3xjoUGWCidutxM0sGbgPOA7YBK8zsKXdfH3bOZOBrwCnuvtfMRoV1cT/wQ3d/3syGAP27okJEpA/c9mIxDy0v55/PKuSKOEww5OhgZvzHJ6ZT29jCj5/ZSNagFK5c0Pd/H7ftredPa3Zy7ckTGZqR2uf999bM/GEsKBjB3a9t5uqFE0lL0S/qpW9E82/SPKDY3UvdvQl4GLiowzk3ALe5+14Ad68AMLNpQIq7Px+017p7fRRjFRHpc4+9sY2fPfcOn5qdx5fPnxrrcOQol5xk/PelMznnuFF868m1PLlqe5+PcfdrmzHgulMn9XnffWXJGYXsrGngT6t3xDoUGUCimVDnAeVhr7cFbeGmAFPM7HUzW2pmF4S17zOzP5jZW2b202DGW0QkIbz6biVffXwNpxybw48unqEbt0hcSE1O4rYrTmTBpBy+9Mhqnl+/u8/63lvXxMPLy7lw1ljGDhvcZ/32tTOn5DJ1dBZ3vFLSb4s0ZeCLZkLd2U+Pjn9zU4DJwJnA5cBdZjYsaD8N+DIwFygAFn9gALMbzWylma2srKzsu8hFRHph3Y4a/vGBNzl21BD+35Un6dfKElcGpSbz62vmMH1sNv/84Ju8XrynT/p9YGkZB5tbufH0gj7pL1rMjCVnFPDO7lpe2qTcQfpGNP8vvw3ID3s9Duj4+5VtwJPu3uzum4FNhBLsbcBbQblIC/AEcGLHAdz9Tnef4+5zcnNzo/ImRER6Yvu+g1x7zwqyBqVw77XzyB4Uf3WkIkPSQ38/J+VkcsP9K3lz695e9dfQ3Mp9RVs4a2ouxx2T3TdBRtE/zBzL2KGDuP3l6C7QlKNHNBPqFcBkM5tkZmnAIuCpDuc8AZwFYGYjCZV6lAbXDjez9iz5bGA9IiJxrKa+mcW/Wc7B5lbuvXYexwwdFOuQRLo0PDON314/j9ysdBb/Zjkbdu4/4r4ef3Mbe2qbWHJGYR9GGD2pyUlcd+oklm2uZlX5vliHIwNA1BLqYGb5JuBZYAPwiLuvM7Pvm9mFwWnPAlVmth54EbjZ3avcvZVQuccLZvY2ofKRX0crVhGR3mpsaeWG366krKqeO6+aw9RjsmIdkki3RmUP4oHr55ORlsJVdy9n8566HvfR2ub8+pVSZo4byvxJI6IQZXQsmjee7EEp3PmKZqml96Ja2OfuT7v7FHcvdPcfBm3fdvengufu7l9y92nu/iF3fzjs2ufdfUbQvjjYKUREJO60tTn/9shqlm+u5qeXzGBhYU6sQxKJWP6IDB747Dza3LnyrmXs2HewR9c/t24XW6rqWXJGYUItvh2SHto68P/W7mLLEXyREAmnlTIiIr3042c28uc1O/naR47jolnxc6tlkUgdOyqL+6+bx/6DzVx51zL21DZGdJ27c/vLJUzIyeDDJxwT5Sj73uJTJpKalMSvXy2NdSiS4JRQi4j0wr2vb+aOV0q5ZuGEuN/dQORwpucN5TfXzmVHzUGuvns5NQebu71m2eZqVm+r4YbTCkhOSpzZ6XajsgZx8Ul5PPrGNioPRPYlQqQzSqhFRI7QM2t38b0/r+f8aaP59j+ckFC/7hbpzNyJI7j9ypN4t+IA1927gvqmlsOef8fLJeRkpvHpk8b1U4R977OnFdDc2sb9RVtiHYokMCXUIiJH4I2yar7w8FvMzh/GLZfPTsjZOZHOnDl1FL9cNJu3tu5lyW/foLGltdPzNu06wIubKll88kQGpSbuvdcKc4dw/rTR3F9URl3j4b9AiHRFCbWISA+VVNZy/X0rGTtsMHddMzehkwmRznz0Q2P40cUzePXdPXz+obdoaW37wDl3vlLK4NRkrlo4IQYR9q0lZxRSc7CZR1aWd3+ySCeUUIuI9EDlgUYW37OclCTjvmvnMSIzLdYhiUTFpXPy+fbHp/Hsut185fG3aWt772bHO2sO8uSq7Syal8+wjMT/N3Di+OHMmziCu17dTHMnXx5EuqOEWkQkQnWNLVx37wr2HGji7mvmMj4nI9YhiUTVdadO4ovnTuHxN7fx/T+vxz2UVP/mtc04cP2pk2IbYB+68fQCtu87yNNv74x1KJKAUmIdgIhIImhpbeOmB99k3Y4a7rpmDjPzh8U6JJF+8flzjmV/QzN3v7aZrEEpfPa0Ah5ctrXSh4gAACAASURBVJWPzxjDuOED50vl2ceN4thRQ7j95VIunDlWi4ylRzRDLSLSDXfnW0+u5cVNlfzHJz7E2ceNjnVIIv3GzPjmx47nsjn5/O/firn6N8upa2odcNtEJiUZN55ewIad+3n13T2xDkcSjBJqEZFu3Pq3Yh5aXs6/nH0sn5k/PtbhiPQ7M+M/P/UhPjZjDKvL93Ha5JGcMHZorMPqcxfNGsvo7HTu0O3IpYdU8iEichiPrizn58+/w6dOzONL502JdTgiMZOcZPzi0lkcNzqLj84YE+twoiI9JZmbzp7MrpqDtLa5tsOUiCmhFpGE8fz63Xzjj2/T5t79yX2kuq6J0yaP5EefmqGaSjnqpaUk8S/nTI51GFF11YLE3wZQ+p8SahFJGH/bWEFtYwufnJ3Xb2MOHZzKP55ZSFqKKuRERKRzSqhFJGGUVNQybUw2P/zkh2IdioiIyCGachGRhFFSWcuxo4bEOgwREZH3UUItIglhb10TVXVNFOYqoRYRkfiihFpEEkJxZS2AZqhFRCTuKKEWkYRQUhFKqDVDLSIi8UYJtYgkhOKKWtJTksgbPjjWoYiIiLyPEmoRSQgllbUU5A7RjRZERCTuKKEWkYRQXFlLYW5mrMMQERH5ACXUIhL3Gppb2bb3oBYkiohIXFJCLSJxr7SyDnctSBQRkfikhFpE4p62zBMRkXimhFpE4l5JRS1mMGmkaqhFRCT+KKEWkbhXXFlL/vAMBqUmxzoUERGRD1BCLSJxr6RCO3yIiEj8UkItInGttc0p3VOn+mkREYlbSqhFJK5t33uQppY2JdQiIhK3lFCLSFwrrjwAaMs8ERGJX0qoRSSulVTUAUqoRUQkfimhFpG4VlxRS05mGsMz02IdioiISKeUUItIXCuurKVQ9dMiIhLHlFCLSNxyd4oralXuISIicU0JtYjEraq6JmoONmuHDxERiWtRTajN7AIz22RmxWb21S7OudTM1pvZOjN7sMOxbDPbbma3RjNOEYlPJRW1AEqoRUQkrqVEq2MzSwZuA84DtgErzOwpd18fds5k4GvAKe6+18xGdejmB8DL0YpRROJbcWUoodZdEkVEJJ5Fc4Z6HlDs7qXu3gQ8DFzU4ZwbgNvcfS+Au1e0HzCzk4DRwHNRjFFE4lhJRR2DU5MZO3RwrEMRERHpUjQT6jygPOz1tqAt3BRgipm9bmZLzewCADNLAn4O3BzF+EQkzhVX1lKQm0lSksU6FBERkS5FreQD6OwnoHcy/mTgTGAc8KqZTQeuBJ5293Kzrn+QmtmNwI0A48eP74OQRSSelFTUMmfi8FiHISIicljRTKi3Aflhr8cBOzo5Z6m7NwObzWwToQR7IXCamf0TMARIM7Nad3/fwkZ3vxO4E2DOnDkdk3URSWD1TS1s33eQy3Lzuz9ZREQkhqJZ8rECmGxmk8wsDVgEPNXhnCeAswDMbCShEpBSd7/C3ce7+0Tgy8D9HZNpERnYSitDtxzXDh8iIhLvopZQu3sLcBPwLLABeMTd15nZ983swuC0Z4EqM1sPvAjc7O5V0YpJRBJHyaEdPpRQi4hIfItmyQfu/jTwdIe2b4c9d+BLwaOrPu4F7o1OhCISr4orakkymDgyI9ahiIiIHJbulCgicamkspYJOZmkpyTHOhQREZHDUkItInGpuKJWN3QREZGEoIRaROJOS2sbW/bUU6gFiSIikgCUUItI3Cnfe5Cm1jYtSBQRkYSghFpE4k5xRWiHD22ZJyIiiUAJtYjEHW2ZJyIiiUQJtYjEneKKWnKz0hk6ODXWoYiIiHRLCbWIxJ2SylqO1ey0iIgkCCXUIhJX3D20Zd4obZknIiKJQQm1iMSVytpGDjS0aIZaREQShhJqEYkr7Tt8aA9qERFJFEqoRSSulGjLPBERSTBKqEUkrpRU1pGZlswx2YNiHYqIiEhElFCLSFwJLUgcgpnFOhQREZGIKKEWkbhSUlmrG7qIiEhCUUItInGjtrGFnTUNqp8WEZGEooRaROJGqW45LiIiCUgJtYjEjeJDO3zopi4iIpI4lFCLSNwoqawlJcmYkKOEWkREEocSahGJG8UVtYzPySA1Wf9rEhGRxKGfWiISN4oranXLcRERSThKqEUkLjS3tlFWVa9bjouISMJRQi0icaGsqp6WNtcMtYiIJBwl1CISF0oq23f4UEItIiKJRQm1iMSF9i3zCnK1w4eIiCQWJdQiEhdKKms5JnsQWYNSYx2KiIhIjyihFpG4UFJRS6Fu6CIiIglICbWIxJy7U1JZpwWJIiKSkJRQi0jM7d7fSG1ji7bMExGRhKSEWkRirn1BomaoRUQkESmhFpGYa98yTzPUIiKSiJRQi0jMFVfUkpWewqis9FiHIiIi0mNKqEUk5koqaykcNQQzi3UoIiIiPaaEWmQA+ZeH3uJXLxXHOoweK66opVD10yIikqCUUIsMEGu31/Cn1Tt4dOW2WIfSI/sbmqk40KhbjouISMJSQi0yQPxu2VYANu+pY1dNQ4yjiVxJsMNHoW45LiIiCSqqCbWZXWBmm8ys2My+2sU5l5rZejNbZ2YPBm2zzKwoaFtjZpdFM06RRHegoZknV23nQ3lDAVhaWhXjiCJ3aMs8zVCLiEiCilpCbWbJwG3AR4BpwOVmNq3DOZOBrwGnuPsJwL8Gh+qBq4O2C4D/MbNh0YpVJNE98dZ26pta+d5FJ5A9KCWhEuqSyjpSk43xIzJiHYqIiMgRieYM9Tyg2N1L3b0JeBi4qMM5NwC3ufteAHevCP77jru/GzzfAVQAuVGMVSRhuTsPLN3K9LxsZucPY96kHIoSKKEurqhlYk4mKcmqQBMRkcQUzZ9geUB52OttQVu4KcAUM3vdzJaa2QUdOzGzeUAaUNLJsRvNbKWZraysrOzD0EUSxxtle9m0+wBXzp+AmbGgYARlVfXs2Hcw1qFFpLSyVuUeIiKS0KKZUHe2oax3eJ0CTAbOBC4H7gov7TCzMcBvgWvdve0Dnbnf6e5z3H1Obq4msOXo9MDSMrLSU7hw1lgAFhbmAIlRR93U0kZZdb22zBMRkYQWzYR6G5Af9nocsKOTc55092Z33wxsIpRgY2bZwF+Ab7r70ijGKZKwquuaePrtXXzqxDwy0lIAOP6YbIYOTk2IhLqsqo7WNtcMtYiIJLRoJtQrgMlmNsnM0oBFwFMdznkCOAvAzEYSKgEpDc7/I3C/uz8axRhFEtqjK8tpam3jigUTDrUlJRnzJ41gaWl1DCOLTPGhLfOUUIuISOKKWkLt7i3ATcCzwAbgEXdfZ2bfN7MLg9OeBarMbD3wInCzu1cBlwKnA4vNbFXwmBWtWEUSUVub8+DyrcybOIIpo7Ped2xBQQ5bq+vZHud11O0JdYH2oBYRkQSWEs3O3f1p4OkObd8Oe+7Al4JH+DkPAA9EMzaRRPda8R7Kqur50nlTPnDsUB11SRUXnzSuv0OLWEllLWOHDiIzPar/KxIREYkq7VMlkqAeWFpGTmYaF0w/5gPHpo7OYlhGatxvn1dcWUuh6qdFRCTBKaEWSUA7aw7ywsYKLpmTT3pK8geOv1dHHb8JdVubU1JRp/ppERFJeEqoRRLQw8vLaXPnM/PGd3nOwoIctu09SHl1fT9GFrmd+xs42NyqHT5ERCThKaEWSTDNrW08vGIrp0/OZXxO17frXhDn+1GXBAsSlVCLiEiiU0ItkmBe2FDB7v2NXBm2VV5npozKYkRmWtzWUWvLPBERGSiUUIskmN8tK2PM0EGcNfXwdwdtr6NeVlpNaEOd+FJcWcvQwamMHJIW61BERER6RQm1SALZsqeOV9/dw+XzxpOS3P0/34WFOWzfd5Dy6vjbj7qkopbC3EzMLNahiIiI9IoSapEE8uDyrSQnGYvm5kd0/oKC+K2jLqmsVf20iIgMCEqoRRJEQ3Mrj64s5/xpoxmVPSiiayaPGkJOHNZR76tvYk9tk+qnRURkQFBCLZIg/m/tTvbWN3e7GDGcmbGgIIelpVVxVUddUqkdPkREZOBQQi2SIB5YupVJIzNZGJRxRGpBYQ47axooq4qf/ahLKuoAJdQiIjIwKKEWSQAbdu7njbK9XDF/PElJPVvEt7BgBBBfddTFlbWkpSQxbnjX+2iLiIgkCiXUIgngd8vKSEtJ4uITx/X42sLcIYwckh5XddQlFbUUjMwkuYdfDkREROKREmqROFfb2MIf39zOx2eMYXhmz/dsDtVRj4irOuriylotSBQRkQFDCbVInHty1Xbqmlp7tBixo4WFOeze38jmPXV9GNmRaWhupby6nkLVT4uIyADRbUJtIVea2beD1+PNbF70QxMRd+eBpVs5fkw2s/OHHXE/7+1HXd1XoR2xLVV1tDkU5mbGOhQREZE+EckM9a+AhcDlwesDwG1Ri0hEDnmrfB8bdu7nygXje3VHwYKRmYzKSo+LhYnFFdoyT0REBpZIEur57v7PQAOAu+8Fel7IKRIHDjQ0xzqEHnlgaRmZaclcNCuvV/2070ddFAd11CUVdZhBwUgl1CIiMjBEklA3m1ky4ABmlgu0RTUqkSjYuGs/M7/3HI+uLI91KBHZW9fEn9fs5JMn5jEkPaXX/S0szKHyQCOlMa6jLq6sJW/YYAanJcc0DhERkb4SSUJ9C/BHYJSZ/RB4DfjPqEYlEgUvbaqkzeE7T62jNLhTXzx7/M1tNLW0ccX8I1+MGK69jrqoJLZlHyUVtSr3EBGRAaXbhNrdfwf8O/BfwE7gE+7+aLQDE+lrRSVV5A0bTFpKEl94eBVNLfH7i5a2Nud3y7Zy0oThHD8mu0/6nJiTwTHZg2JaR93W5pTu0ZZ5IiIysBw2oTazJDNb6+4b3f02d7/V3Tf0V3AifaW5tY0VW6o5+7hR/PjiGby9vYafP7cp1mF1qai0is176rhywfg+6/O9/airY1ZHvX3fQRqa2zRDLSIiA8phE2p3bwNWm1nf/VQXiYE122qob2plYWEOHz7hGK6YP547XinltXf3xDq0Tj2wtIzhGal8ZPqYPu13YWEOe2obKYlRyUtxMK5mqEVEZCCJpIZ6DLDOzF4ws6faH9EOTKQvtZc5zJ80AoBvfmwax44awpceWUVVbWMsQ/uA3fsbeG79bi6Zk8+g1L5duBfrOuoSbZknIiIDUCQJ9feAjwPfB34e9hBJGEtLq5g6OoucIekADE5L5pZFs9lX38xXHl8T863kwv1+RTmtbc5n5vX9L4bGj8hg7NBBMbvBS0llLcMzUhlxBLdQFxERiVeRLEp8GdgIZAWPDUGbSEJoamlj5Za9LCzMeV/7tLHZfPUjx/HXDRU8sLQsRtG9X0trGw8t38ppk0cycWTf30mwfT/qpTHaj7pYO3yIiMgAFMmtxy8FlgOXAJcCy8zs09EOTKSvrN62j4PNrYfKHcJde8pEzpyay3/8ZQObdh2IQXTv9+KmSnbWNPTZVnmdWVCYQ1VdE+9W9H8ddUllnRJqEREZcCIp+fgGMNfdr3H3q4F5wLeiG5ZI3ykqqcIMFhSM+MAxM+Onn55J1qBUPv/QWzQ0t8Ygwvc8sLSM0dnpnHv8qKiNsTBGddTVdU1U1zVpQaKIiAw4kSTUSe5eEfa6KsLrROLC0tIqjjsmm2EZndft5mal87NLZrBp9wH+6+nY7Qq5taqeV96tZNHc8aQkR++fWP6IDPKGDe73/ajbdxYp1Ay1iIgMMJH81H7GzJ41s8Vmthj4C/B/0Q1LpG80trTyRtneQ7OyXTlz6iiuP3US9xWV8cKG3f0U3fs9uHwrSWZcHoXFiB2111G3tfVfHXVx+w4fmqEWEZEBJpJFiTcDdwAzgJnAne7+79EOTKQvvLV1H40tbR9YkNiZf79gKsePyebmx9ZQsb+hH6J7T2NLK4+sLOec40ZxzNBBUR9vQcEI9tY3805F/9WNF1fUkp6SRN6wwf02poiISH+IZFHiJOBpd/+Su3+R0Iz1xGgHJtIXikqqSDKYN+mD9dMdpack87+Xz6K+qYV/e3R1v87ePrN2F9V1TVy5IHqLEcPFYj/qkspaCnKHkJRk/TamiIhIf4ik5ONRoC3sdWvQJhL3lpZWccLYoQwdnBrR+ceOyuLbHz+BV9/dw92vbY5ydO/53dKtTMjJ4NRjR/bLePkjMhg3vH/rqLVlnoiIDFSRJNQp7t7U/iJ4rrsySNxraG7lra37Ot3d43Aun5fPh08YzU+e3cja7TVRiu49m3YdYPmWaj4zb3y/zt4uLMhh2ebqfpmJP9jUyvZ9B1U/LSIiA1IkCXWlmV3Y/sLMLgL2RC8kkb7xZtlemlojq58OZ2b86FMzyMlM5/MPvUVdY0uUIgx5cFkZaclJXDInP6rjdLSgIId99c1s7If9t0v31OIOhaP6/mY1IiIisRZJQv054OtmttXMyoGvAEsi6dzMLjCzTWZWbGZf7eKcS81svZmtM7MHw9qvMbN3g8c1kYwnEq6otIrkJGPuxJ7NUAMMz0zjF5fNYnNVHd//0/ooRBdS19jCH97czkc/dEy/3457QfBFo6gfyj5KKusAVPIhIiIDUiS7fJS4+wJgGjDN3U929+LurjOzZOA24CPBtZeb2bQO50wGvgac4u4nAP8atI8AvgPMJ3Qjme+Y2fAevTM56hWVVDE9byhZgyKrn+5oYWEO/3RmIb9fWc5f1uzs4+hC/rR6BwcaW/ptMWK4vGGDGT8io1/qqIsrakkymJijGWoRERl4Itnl4wtmlg3UAb8wszfN7PwI+p4HFLt7aVB3/TBwUYdzbgBuc/e9AGE3kPkw8Ly7VwfHngcuiOwtiUB9Uwurt/W8frqjfz13CjPzh/G1P6xh+76DfRRdiLvzwLIypo7O4qQJsfm+uLAgh+X9UEddUlFL/ogMBqUmR3UcERGRWIik5OM6d98PnA+MAq4FfhTBdXlAedjrbUFbuCnAFDN73cyWmtkFPbgWM7vRzFaa2crKysoIQpKjxRtle2lu9W5v6NKd1OQkblk0i9Y254sPr6K1DxPP1dtqWLt9P1cuGI9ZbLaSW1A4gpqDzazfuT+q45RU1uqW4yIiMmBFklC3/6T/KHCPu68Oa4vkunAds5EUYDJwJnA5cJeZDYvwWtz9Tnef4+5zcnNzIwhJjhZFJVWkHGH9dEcTcjL5wSems3xLNb96sdtqp4j9bmkZGWnJfGL2B74r9pv2/aijWfbR2uaU7qlT/bSIiAxYkSTUb5jZc4QS6mfNLIv370vdlW1A+LYF44AdnZzzpLs3u/tmYBOhBDuSa0W6VFRaxYxxQ8lMT+mT/j45O4+LZo3lf154lzfK9va6v5r6Zv60ZgcXzco74hrvvjBm6GAm5kS3jnrb3nqaWtoozFX9tIiIDEyRJNTXA18F5rp7PaE9qK+N4LoVwGQzm2RmacAi4KkO5zwBnAVgZiMJlYCUAs8C55vZ8GAx4vlBm0i36hpbWLOtpsfb5R2OmfGDT0xnzNBBfOHht9jf0Nyr/h5/cxsNzW1cMX98H0V45BYWhvaj7stylnDFFbWAdvgQEZGBK5JdPtrc/U133xe8rnL3NRFc1wLcRCgR3gA84u7rzOz7YftaPwtUmdl64EXg5qD/auAHhJLyFcD3gzaRbq3YEkoOF/Syfrqj7EGp/HLRbHbWNPDNP67F/cgSUHfnd8vKmJU/jOl5Q/s0xiOxoCCHAw0trN8RnTrqkspQQq0aahERGaj65vfhXXD3p4GnO7R9O+y5A18KHh2v/Q3wm2jGJwNTUWkVqcnGnAm9r5/u6KQJw/nXcybz8+ff4cypuXzqxHFHFF9JZR0/u2Rmn8d3JMLrqD80ru8T/OKKWkYOSWNYhm6wKiIiA1MkJR8iCWVpSRWz8ocxOC06W7T901nHMm/iCL71xFrKqup6fP3vlm1l6OBUPj5jTBSi67nR2YMoGJkZtRu8lFTWaXZaREQGtCNKqM1MPx0lLu1vaObt7TW93i7vcJKTjF8smkVykvH5h1fR3BrJGt2QigMNPLt2F58+aVxc7cm8oDCHFZuraenBe4mEu1NcUUuh6qdFRGQAO9IZ6ujdi1mkF1ZuqabN37utdrTkDRvMjy6eweryffzPX9+J+LpHVpTT0uZ8Jg4WI4ZbUJDDgcYW1vVxHfWe2iZqDjZzrGaoRURkAOuyhtrMPlDX3H4I0E9HiUtFJVWkJSdx4vjo33nwox8aw2Vz8vnVSyWccuxITi4cedjzW9uch5aXc3JhTtyVQLTfUXJpaRUz84f1Wb+HFiRqhlpERAaww81Q/ycwHMjq8BjSzXUiMVNUWsXs8cP6rZziOxdOY1JOJl/6/Wr21jUd9tyXNlWwfd9BrlwwoV9i64lRWYMozO37OmptmSciIkeDwyXGbwJPuPv3Oj6AA/0Un0jEauqbWbdjf5/uP92djLQUbrl8NlV1jXz1D2sOu5Xe75ZtJTcrnfOmje63+HpiYRTqqEsqa8lIS2ZM9qA+61NERCTeHC6hvhYo6+LYnCjEItIryzZX4U5UFyR2ZnreUP79w8fx7LrdPLS8vNNzyqvreXFTBYvm5pOaHJ+/4FlQkENdUytvb6/psz6LK2opyM0kKcn6rE8REZF4c7if7N909z1m9oWOB9x9dxRjEjkiS0urSU9JYtb4vqsBjtT1p07itMkj+f6f11Fc8cFf4Dy0fCsGLJoXX4sRw723H3Xf3UOptLJOCxJFRGTAO1xCfZKZTQCuC24BPiL80V8BikSqqLSKkyYMJz2l/7ejS0oyfn7JTDLSUviXh1bR2NJ66FhTSxuPrCzn7ONGkTdscL/HFqmRQ9KZPGpIn9VR1zW2sH3fwbhbgCkiItLXDpdQ3w48AxwHvNHhsTL6oYlEbm9dExt27u/3co9wo7IH8dNPz2DDzv385JlNh9qfXbeLPbVNXBGHixE7WliYw8ot1T3aW7srpZWhm95oQaKIiAx0XSbU7n6Lux8P/MbdC9x9UtijoB9jFOnWss2hWdX+XJDYmXOOH801Cydw92ubeWlTBQC/W1bGuOGDOX1ybkxji8SCghzqm1pZs633ddTaMk9ERI4W3a6Ocvd/7I9ARHpjaWk1g1OTmTGu/+unO/raR49n6ugsvvzoaopKqlhaWs1n5o8nOQEW5s2f9N5+1L1VXFFLcpIxISej132JiIjEs/jcbkCkh4pKqpgzcThpKbH/Kz0oNZlbLp/NgYYWFt+znNRk49I5+bEOKyI5Q9KZOjqrTxLqkspaxo/IiElNu4iISH+KffYh0ktVtY1s2n3g0C4V8WDqMVl882PH09jSxgXTxzBySHqsQ4pYqI56L00tvaujLq6o1YJEERE5KnR563GRRNG+zVus66c7unLBBAalJnP6lPivnQ63oGAE9/59C29v38dJE45sQ5+W1ja2VNVxzvHxeRMbERGRvqQZakl4RaV7yExL5kN5Q2MdyvuYGZfMyWd0gt0lcP6k0BeTopIjL/vYWl1Pc6tTmJvZV2GJiIjELSXUkvCWllYzZ+KIuL0DYaIZnpnGccdk9eoGLyXaMk9ERI4iykAkoVUcaKC4ojbuyj0S3YKCHFaWVb/vBjU9UVyhLfNEROTooYRaEtqh+uk4WpA4ECwszKGhue2I96MurqhlVFY62YNS+zgyERGR+KOEWhJaUUkVWekpnDA2O9ahDCjzJ43A7MjrqEsqtcOHiIgcPZRQS0JbWlrFvEkjSFH9dJ8alpHG8cdkH9F+1O5OSUWt6qdFROSooSxEEtaumgY276mLq/2nB5IFBTm8Uba3x3XUlQcaOdDYooRaRESOGkqoJWG1z55qQWJ0LCzMobGljVVb9/XoukMLElXyISIiRwkl1JKwikqqyB6UwvFjVD8dDfMmBnXUPSz7KKkMJdSaoRYRkaOFEmpJWEWlVcwvyCE5yWIdyoA0NCOVE8b2vI66uKKWIekpjM5OnNuti4iI9IYSaklI2/cdZGt1veqno2zBpBze3LqPhubI66iLK2spzM3ETF90RETk6KCEWhLS0mA7N+0/HV0LC3NoamnjrR7UUZdU1Kl+WkREjipKqCUhFZVWMTwjleOOyYp1KAPa3EkjSOpBHfWBhmZ27W/QHRJFROSoooRaElJRSRXzJ+WQpPrpqMoelMr0vKER11GXVtYB2uFDRESOLkqoJeGUV9ezfd9BbZfXTxYU5LAqwjrq9i3ztMOHiIgcTZRQS8JpLz/QgsT+sbAgh6bWNt4s29vtuSWVtaQkGRNyMvohMhERkfighFoSztKSKnIy05gyWrOg/WHOxOEkJ1lEddTFFbVMyMkgVbeCFxGRo4h+6klCcXeKSqtYUJCjbdn6SVYP6qiLK2tV7iEiIkcdJdSSUMqq6tlZ08AC1U/3qwUFI1hVvo+DTV3XUTe3trG1ql4LEkVE5KijhFoSSvssqfaf7l8LC3JobnXeOEwddVlVHS1trhlqERE56kQ1oTazC8xsk5kVm9lXOzm+2MwqzWxV8Phs2LGfmNk6M9tgZreYfr8vhBYk5malU5ibGetQjipzJo4gOckOW/ZRXKEt80RE5OiUEq2OzSwZuA04D9gGrDCzp9x9fYdTf+/uN3W49mTg/7d35+Fx3fW9x99frZa8SLIt77IlZyGr7diy48QBAgUaaHA2KA4pIU4p5bZpSnsphJZLaWi5bKWXAs/lIdSQUPaEgIFAwnYDSWQ8cuI1TmJLHtvyKmsk2ZJtrd/7h47MWB7Ji3TmjDSf1/Po0cxZvzo6sj86+p7fWQEsCCY9A7wW+H9h1SuZz92pqVP/dBQmFOaxYE7JkDcm1jX2DZmnh7qIiEi2CfMK9TJgp7vXu3sn8B3glnNc14FxQAFQCOQDh0KpUkaN+iPtHD7WoXaPiCyfP4VNe1s43tmdcn7d4TZmloxjQmFov6eLiIhkpDAD9Wxgb9L7hmDaQHeY2WYze9TMKgDcvQb4DXAg+HjS3bcPRRT7ggAAIABJREFUXNHM3mtmtWZW29jYOPJfgWSUmrqgf1o3JEbiuvlT6O51auOp+6h3Nrap3UNERLJSmIE61d/kfcD7HwOV7r4A+CXwMICZXQxcDsyhL4S/3sxec8bG3L/i7tXuXl1eXj6ixUvmWVffxPRJhVTqoSGRWDKvjLxB+qjdnbrDGjJPRESyU5iBugGoSHo/B9ifvIC7N7l7R/D2IWBJ8Po2YJ27t7l7G/AzYHmItUqGc3fW1Se4Tv3TkRlfmMfCitKUfdQHj56kvbNHN4uKiEhWCjNQx4BLzKzKzAqAVcDa5AXMbGbS25VAf1vHHuC1ZpZnZvn03ZB4RsuHjDz3gX9EyAw7D7dxpK1D7R4RWz5/MpsbWmnvOL2Peudh3ZAoIiLZK7RA7e7dwH3Ak/SF4e+5+zYze9DMVgaL3R8MjbcJuB+4J5j+KFAHbAE2AZvc/cdh1Sp9vvx0HW/43NO0daS+6SxKNafGn54acSXZ7br5U+npdWLxxGnT64JAfbF6qEVEJAuFeju+uz8BPDFg2keTXn8Y+HCK9XqAvwyzNjnTz7cepK6xnU/97CU+futVUZdzmpq6JmaXFlExuSjqUrLaknll5Oca6+oT3Piqaaem72xsY+K4PMonFkZYnYiISDT0pEQB4HhnN1v3tVJSlM831u1m/a7E2VdKk95e5/e7Elw7f7L6pyNWVJDLohR91HWH27mofIK+PyIikpUUqAWAjXtb6O51PnHb1VRMLuJDj23mZFdP1GUB8MrhYyTaOzX+dIZYPn8KW/e1cuxk16lpOxs1woeIiGQvBWoBILarGTO44ZKpfPL2Bew60s5//PKVqMsCNP50plk+fwo9SeNRt57oovFYhwK1iIhkLQVqASAWT3DZjEmUFOWz4uKprFpawUO/rWdzQ0vUpVFT10TF5CLmlGn86UyweG4ZBbk5p8ajPvXIcd2QKCIiWUqBWuju6eX5Pc0srSw7Ne3Db7mc8omFfPDRzXR290ZWW3//9PIqXZ3OFAP7qPuHzNMVahERyVYK1MKLB45yvLOHpZWTT00rKcrnX2+9mpcOHuPLT9dFVtv2g0dpPdGldo8Ms/yivj7qoye7qGtsoyA3h4oyjcAiIiLZSYFaTo3okRyoAd54xXTeunAWX/j1Dl45dCyK0tQ/naGWz59Mr0NsV4K6w21UTi0mL1f/nIiISHbS/4BCbbyZislFzCgZd8a8j731CiaOy+eDj26mpzf9T1FcV99E5ZRiZpbo6mcmWTy3jIK8vj7qusZ29U+LiEhWU6DOcu59T70beHW635QJhfzzW69g494WvvbsrrTW1hP0T+vqdOYZl5/LNRWl/PaVI+xJHFf/tIiIZDUF6ixXf6SdpvZOlg0SqAFWLpzFH102jc8+9TK7m9rTVtuL+49y7GQ3yzX+dEa67qIpvHzoGD29rkAtIiJZTYE6y8WC/unqIQK1mfFvt11Nfk4ODzy2Bff0tH7U1B8B0ANdMlTyLzpq+RARkWymQJ3lYvFmpowv4KLy8UMuN6NkHP/4J5dTU9/Et9fvTUttNXVNzC8fz7RJZ/Z2S/QWVZRSmNf3T8j8s5w/IiIiY5kCdZaLxRNUV5ZhZmdddtXSCq6/aAqfeGI7B1pPhFpXd08vsXizrk5nsHH5uSyZV8acsiKKC/KiLkdERCQyCtRZ7NDRk+xJHB/0hsSBzIxP3r6A7t5e/unxraG2fmzZ10pbh/qnM92/3noVX3zn4qjLEBERiZQCdRaLxVOPPz2UuVOK+cCbXsWvXzrM2k37wyqNdfV9tSlQZ7b55RNYVFEadRkiIiKRUqDOYrFdCYoLcrly1qTzWm/1iiqumVvKx9Zu40hbRyi11dQ3ccm0CZRPLAxl+yIiIiIjRYE6i62PN3PN3NLzfsJdbo7x6TsW0N7Rw8fWbhvxurp6eqmNa/xpERERGR0UqLPU0ZNdvHTw6Hm1eyS7ZPpE/ub1F/OTzQd4atvBEa1tc0MLxzt7dEOiiIiIjAoK1Flqw+5m3BnygS5n874bL+KyGRP5yA+30nqia8Rq6++fvlaBWkREREYBBeosFduVIC/HWDT3wm8oy8/N4TNvW0hTeyef+On2Eautpq6Jy2ZMZPL4ghHbpoiIiEhYFKizVG28mStnlwx7/OCr55TwF6+ez3dr9/LMjiPDrquju4fa3QmN7iEiIiKjhgJ1Furo7mFjQwvLKstGZHvvf8MlzJ86ngd+sJn2ju5hbWvT3lZOdvXqhkQREREZNRSos9DmhlY6u3sv+IbEgcbl5/Kpty2gofkEn33q5WFtq6auCTO4tmpkahMREREJmwJ1Fup/oEv1CAVq6Hs4zN3XzePrz8XZsDtxwdtZV9/E5TMmUVqs/mkREREZHRSos1BsV4KLp00Y8Zv+PnjTZcwqKeKDj27mZFfPea9/squHDXua1e4hIiIio4oCdZbp6XVqdzePWLtHsgmFefzv26+mrrGdL/x6x3mv/8KeFjq7ezX+tIiIiIwqCtRZ5uWDxzh2spulI3RD4kCvubScty2Zw5efrmfb/tbzWremvokcg6XqnxYREZFRRIE6y9QG/c1hXKHu95E/uZyy4gI++Ohmunp6z3m9dfVNXDmrhJKi/NBqExERERlpCtRZZv2uBDNLxjGnrCi0fZQWF/Cvt17Jtv1H+cpv689pnZNdPWzc06L+aRERERl1FKiziLsTiyeorpyMmYW6r5uumslbrp7B53+1g52H2866/IbdzXT2qH9aRERERh8F6izS0HyCQ0c7RuyBLmfzLyuvoig/lw89tpneXh9y2Zq6JnJzTP3TIiIiMuooUGeR9buC/uk0hdbyiYV89OYr2LC7mUdq4kMuW1PfxNWzS5hQOLxHoYuIiIikmwJ1FonFE0wal8el0yambZ+3L57Nay8t59NPvszexPGUyxzv7GbT3haWq91DRERERiEF6izS3z+dkxNu/3QyM+MTt1+NAR/+wRbcz2z9qI03093ruiFRRERERiUF6izR1NZBXWN7qMPlDWZ2aREPvOVyntl5hO/XNpwxv6a+ibwco3peenq7RUREREZSqIHazG4ys5fNbKeZPZBi/j1m1mhmG4OP9yTNm2tmT5nZdjN70cwqw6x1rIvFmwFCe6DL2dy1bC7Lqibz8Z++yKGjJ0+bV1PXxMKKUsarf1pERERGodACtZnlAl8C3gxcAdxpZlekWPS77r4o+Phq0vRHgM+4++XAMuBwWLVmg9p4goK8HK6eUxLJ/nNyjE/dsYDO7l7+1w+3nmr9aOvoZsu+VpbP1+geIiIiMjqFeYV6GbDT3evdvRP4DnDLuawYBO88d/8FgLu3uXvqO9rknMTiCRZVlFKYlxtZDVVTx/P3b7yUp148xE+3HDhVV0+vc938qZHVJSIiIjIcYQbq2cDepPcNwbSB7jCzzWb2qJlVBNMuBVrM7Adm9oKZfSa44n0aM3uvmdWaWW1jY+PIfwVjRHtHN1v3H2VZBP3TA/35DVUsmFPCP/9oG4n2TtbVNZGfayxR/7SIiIiMUmEG6lRDSQwc4uHHQKW7LwB+CTwcTM8DXg18AFgKzAfuOWNj7l9x92p3ry4vLx+pusecF/a00NPrVEfUP50sLzeHT79tAa0nuvj4T16kpr6JayrKKCqI7sq5iIiIyHCEGagbgIqk93OA/ckLuHuTu3cEbx8CliSt+0LQLtIN/BBYHGKtY1osniDHyJirwJfNmMRfve5iHn9hH5sbWlmu4fJERERkFAszUMeAS8ysyswKgFXA2uQFzGxm0tuVwPakdcvMrP+y8+uBF0OsdUyLxRNcPnMSE8flR13KKfe97mIunT4BQDckioiIyKgWWqAOrizfBzxJX1D+nrtvM7MHzWxlsNj9ZrbNzDYB9xO0dbh7D33tHr8ysy30tY88FFatY1lXTy8v7GmJZPzpoRTk5fD5VdewamkF1fMyqzYRERGR82Gpnlw3GlVXV3ttbW3UZWScjXtbuPVLz/Kldy7mTxbMPPsKIiIiEjoz2+Du1VHXISNDT0oc42K7EkB0D3QRERERGesUqMe49fEE86YUM23SuKhLERERERmTFKjHMHenNp7IuP5pERERkbFEgXoMq2tso/l4V0Y80EVERERkrFKgHsPW72oGyIgHuoiIiIiMVQrUY1htPMHUCQVUTR0fdSkiIiIiY5YC9Ri2PuifNkv1FHgRERERGQkK1OfB3enq6Y26jHNyoPUEDc0nqFb/tIiIiEioFKjP0fHObm7+wjM89Lv6qEs5J7F4X/+0bkgUERERCZcC9TkqLsijtDifR57bPSquUsd2JRhfkMvlMydGXYqIiIjImKZAfR5WX1/FwaMn+fnWg1GXclaxeILF88rIy9W3WERERCRMSlvn4fWXTWPelGLWPLsr6lKG1Hq8i5cPHdMDXURERETSQIH6POTkGPdcX8kLe1rYuLcl6nIGtWFPAncUqEVERETSQIH6PL29uoKJhXl8LYOvUq/f1Ux+rrGoojTqUkRERETGPAXq8zShMI+3V1fw080HONh6MupyUorFE1w1u4SigtyoSxEREREZ8xSoL8A911fS485/r9sddSlnONnVw+aGFg2XJyIiIpImCtQXYO6UYt5w+XS++fvdnOzqibqc02za20JXj6t/WkRERCRNFKgv0L0rqmg+3sWPNu6LupTTxOIJAJbMK4u4EhEREZHsoEB9gZbPn8xlMyay5pk47h51OafE4s1cOn0CZeMLoi5FREREJCsoUF8gM+PeG6p4+dAxauqaoi4HgJ5e5/ndzWr3EBEREUkjBephWLlwFlPGF2TMg162HzjKsY5uBWoRERGRNFKgHoZx+bncde1cfvXSYeJH2qMuh9qgf3pplQK1iIiISLooUA/Tny2fR16O8fXn4lGXQizezOzSImaXFkVdioiIiEjWUKAepmmTxnHzglk8uqGBYye7IqvD3VkfT1BdqdE9RERERNJJgXoErF5RSVtHN9+rbYisht1Nx2k81qH+aREREZE0U6AeAQvmlLJkXhkPPxenpzeaIfT6x59epv5pERERkbRSoB4h966oYk/iOL9+6XAk+4/FE5QW53Nx+YRI9i8iIiKSrRSoR8gfXzmdWSXjWPNMNEPoxeLNVM8rIyfHItm/iIiISLZSoB4hebk53H19JTX1TWw/cDSt+2481sGuI+3qnxYRERGJgAL1CFq1tIJx+Tl8Lc0PetH40yIiIiLRUaAeQaXFBdyxeA4/3LifpraOtO13fTzBuPwcrppVkrZ9ioiIiEgfBeoRtnpFJZ3dvXzr93vSts/aeDOLKkopyNO3U0RERCTdlMBG2MXTJvKaS8v5xrrddHb3hr6/to5utu1vZZn6p0VEREQioUAdgtUrKjl8rIMnthwIfV/P726m16FagVpEREQkEgrUIXjtJeXMLx/Pmmd34R7ug15q4wlyDBbP0yPHRURERKIQaqA2s5vM7GUz22lmD6SYf4+ZNZrZxuDjPQPmTzKzfWb2xTDrHGk5Ocbq6yvZ3NDK83uaQ93X+niCK2eVMKEwL9T9iIiIiEhqoQVqM8sFvgS8GbgCuNPMrkix6HfdfVHw8dUB8z4OPB1WjWG6ffEcJo3LY82z8dD20dndywt7Wqiu1NVpERERkaiEeYV6GbDT3evdvRP4DnDLua5sZkuA6cBTIdUXqvGFeaxaNpefbz3I/pYToexjy75WOrp7dUOiiIiISITCDNSzgb1J7xuCaQPdYWabzexRM6sAMLMc4N+BfxhqB2b2XjOrNbPaxsbGkap7xNx93TzcnUdqdoey/f4HuuiGRBEREZHohBmoLcW0gXfo/RiodPcFwC+Bh4PpfwU84e57GYK7f8Xdq929ury8fNgFj7Q5ZcX88ZUz+Pb6PRzv7B7x7cfiCaqmjqd8YuGIb1tEREREzk2YgboBqEh6PwfYn7yAuze5e/8jBR8ClgSvrwPuM7M48FngbjP7ZIi1hmb1iipaT3Tx+Av7RnS7vb1OLN7MUvVPi4iIiEQqzEAdAy4xsyozKwBWAWuTFzCzmUlvVwLbAdz9Lnef6+6VwAeAR9z9jFFCRoOllWVcNXsSX3s2PqJD6O1sbKP1RBdL1e4hIiIiEqnQArW7dwP3AU/SF5S/5+7bzOxBM1sZLHa/mW0zs03A/cA9YdUTFTNj9fVV7Dzcxu92HBmx7a7f1dc/vaxKgVpEREQkSqGOQ+3uT7j7pe5+kbv/WzDto+6+Nnj9YXe/0t0Xuvvr3P2lFNv4urvfF2adYbt54UymTijka8/uGrFtxuIJyicWMndy8YhtU0RERETOn56UmAaFebn82fK5/OblRuoa20Zkm7XxZpZVTsYs1b2fIiIiIpIuCtRpcte18yjIzeHh5+LD3ta+lhPsazmhGxJFREREMoACdZqUTyzkrQtn8eiGBlpPdA1rW7FdGn9aREREJFMoUKfR6hWVHO/s4XuxIYfXPqtYPMHEwjwunzlphCoTERERkQulQJ1GV80uYVnVZL7+XJzunt4L3k4snmDxvDJyc9Q/LSIiIhI1Beo0u3dFFftaTvDL7YcuaP3m9k5eOdSm/mkRERGRDKFAnWZvvGI6c8qKWPNM/ILW37C7GUAPdBERERHJEArUaZabY9xzfSXr4wm27ms97/Vj8QQFuTksrCgNoToREREROV8K1BF4e3UFxQW5rLmAB72sjye4ek4J4/JzQ6hMRERERM6XAnUESoryefuSOfxk0wEaj3Wc83onOnvY0tCqdg8RERGRDKJAHZF3X19JZ08v3/z97nNeZ+PeFrp7nWVVuiFRREREJFMoUEdkfvkEXveqcv573W46unvOaZ1YPIEZLJmrK9QiIiIimUKBOkL33lDFkbZOfrLpwDktH4sneNX0iZQU54dcmYiIiIicKwXqCN1w8VQumTaBNc/uwt2HXLa7p5fndzerf1pEREQkwyhQR8jMWL2iim37jxKLNw+57PYDx2jv7GFplQK1iIiISCZRoI7YbdfMprQ4nzXPDD2E3vp4AkBPSBQRERHJMArUESsqyOXOZXN56sWD7E0cH3S52niCOWVFzCwpSmN1IiIiInI2CtQZ4F3L52FmPFITTznf3YnFEyxT/7SIiIhIxlGgzgCzSot481Uz+E5sL+0d3WfM33WknSNtnVQrUIuIiIhkHAXqDLF6RRXHTnbz2PMNZ8yrDW5Y1ANdRERERDKPAnWGWDy3lIUVpXz92Ti9vacPobc+nqCsOJ+LyidEVJ2IiIiIDEaBOkOYGfeuqKT+SDtPv9J42rxYPEF15WTMLKLqRERERGQwCtQZ5M1XzWT6pELWPPuHIfQOHz3J7qbjuiFRREREJEMpUGeQgrwc3rV8Hr/bcYQdh44BnHrgix7oIiIiIpKZFKgzzJ3L5lKQl8PXnosDfe0eRfm5XDlrUrSFiYiIiEhKCtQZZsqEQm5bNJsfPN9Ay/FO1u9KcM3cUvJz9a0SERERyURKaRlo9Q2VnOzq5aHf1fPSwaMsVf+0iIiISMZSoM5Al82YxPUXTeHLT9fT6yhQi4iIiGQwBeoMtXpFFT29Tm6Occ3c0qjLEREREZFB5EVdgKT2+sumUTmlmLLxBYwv1LdJREREJFMpqWWo3BzjG39+bdRliIiIiMhZKFBnsIrJxVGXICIiIiJnoR5qEREREZFhUKAWERERERkGBWoRERERkWEINVCb2U1m9rKZ7TSzB1LMv8fMGs1sY/DxnmD6IjOrMbNtZrbZzN4RZp0iIiIiIhcqtJsSzSwX+BLwRqABiJnZWnd/ccCi33X3+wZMOw7c7e47zGwWsMHMnnT3lrDqFRERERG5EGFeoV4G7HT3enfvBL4D3HIuK7r7K+6+I3i9HzgMlIdWqYiIiIjIBQozUM8G9ia9bwimDXRH0NbxqJlVDJxpZsuAAqAuxbz3mlmtmdU2NjaOVN0iIiIiIucszEBtKab5gPc/BirdfQHwS+Dh0zZgNhP4BrDa3XvP2Jj7V9y92t2ry8t1AVtERERE0i/MQN0AJF9xngPsT17A3ZvcvSN4+xCwpH+emU0Cfgp8xN3XhViniIiIiMgFCzNQx4BLzKzKzAqAVcDa5AWCK9D9VgLbg+kFwOPAI+7+/RBrFBEREREZltBG+XD3bjO7D3gSyAXWuPs2M3sQqHX3tcD9ZrYS6AYSwD3B6n8KvAaYYmb90+5x941h1SsiIiIiciHMfWBb8+hUXV3ttbW1UZchIiIiclZmtsHdq6OuQ0aGnpQoIiIiIjIMCtQiIiIiIsOgQC0iIiIiMgxjpofazBqB3VHXkUGmAkeiLiJD6dikpuOSmo7L4HRsUtNxGZyOzR/Mc3c9RGOMGDOBWk5nZrW62SE1HZvUdFxS03EZnI5Najoug9OxkbFKLR8iIiIiIsOgQC0iIiIiMgwK1GPXV6IuIIPp2KSm45KajsvgdGxS03EZnI6NjEnqoRYRERERGQZdoRYRERERGQYFahERERGRYVCgHsXMrMLMfmNm281sm5n9bYplbjSzVjPbGHx8NIpa083M4ma2Jfiaa1PMNzP7TzPbaWabzWxxFHWmm5m9Kulc2GhmR83s/QOWyZpzxszWmNlhM9uaNG2ymf3CzHYEn8sGWffdwTI7zOzd6as6fIMcl8+Y2UvBz8vjZlY6yLpD/uyNZoMcl4+Z2b6kn5e3DLLuTWb2cvBvzgPpqzo9Bjk23006LnEz2zjIumP2nJHsoR7qUczMZgIz3f15M5sIbABudfcXk5a5EfiAu98cUZmRMLM4UO3uKR8gEPyn9zfAW4Brgc+7+7XpqzB6ZpYL7AOudffdSdNvJEvOGTN7DdAGPOLuVwXTPg0k3P2TQfApc/cPDVhvMlALVANO38/eEndvTusXEJJBjsubgF+7e7eZfQpg4HEJloszxM/eaDbIcfkY0Obunx1ivVzgFeCNQAMQA+5M/rd6tEt1bAbM/3eg1d0fTDEvzhg9ZyR76Ar1KObuB9z9+eD1MWA7MDvaqkaNW+j7h9/dfR1QGvyCkk3+CKhLDtPZxt1/CyQGTL4FeDh4/TBwa4pV/xj4hbsnghD9C+Cm0ApNs1THxd2fcvfu4O06YE7aC4vYIOfLuVgG7HT3enfvBL5D33k2Zgx1bMzMgD8Fvp3WokTSSIF6jDCzSuAa4PcpZl9nZpvM7GdmdmVaC4uOA0+Z2QYze2+K+bOBvUnvG8i+X0ZWMfh/cNl4zvSb7u4HoO+XVmBaimWy/fy5F/jZIPPO9rM3Ft0XtMKsGaRFKNvPl1cDh9x9xyDzs/GckTFGgXoMMLMJwGPA+9396IDZzwPz3H0h8AXgh+muLyIr3H0x8Gbgr4M/RyazFOtkTf+TmRUAK4Hvp5idrefM+cja88fM/gnoBr45yCJn+9kba/4vcBGwCDgA/HuKZbL2fAncydBXp7PtnJExSIF6lDOzfPrC9Dfd/QcD57v7UXdvC14/AeSb2dQ0l5l27r4/+HwYeJy+P7kmawAqkt7PAfanp7qM8GbgeXc/NHBGtp4zSQ71t/8Enw+nWCYrz5/g5subgbt8kBtwzuFnb0xx90Pu3uPuvcBDpP56s/J8ATCzPOB24LuDLZNt54yMTQrUo1jQl/ZfwHZ3/9wgy8wIlsPMltH3PW9KX5XpZ2bjg5s0MbPxwJuArQMWWwvc3TfYhy2n72aZA2kuNUqDXjHKxnNmgLVA/6gd7wZ+lGKZJ4E3mVlZ8Cf+NwXTxiwzuwn4ELDS3Y8Pssy5/OyNKQPuvbiN1F9vDLjEzKqCvw6tou88ywZvAF5y94ZUM7PxnJGxKS/qAmRYVgDvArYkDUf0j8BcAHf/MvA24H+YWTdwAlg12JWlMWQ68HiQCfOAb7n7z83sfXDquDxB3wgfO4HjwOqIak07Myumb7SBv0yalnxssuacMbNvAzcCU82sAfhn4JPA98zsz4E9wNuDZauB97n7e9w9YWYfpy8oATzo7hdys1pGGuS4fBgoBH4R/Gytc/f3mdks4Kvu/hYG+dmL4EsIxSDH5UYzW0RfC0ec4Ocq+bgEI6PcR98vXbnAGnffFsGXEJpUx8bd/4sU92pk0zkj2UPD5omIiIiIDINaPkREREREhkGBWkRERERkGBSoRURERESGQYFaRERERGQYFKhFRERERIZBgVpEREREZBgUqEUkI5hZpZmd9YEOZjbTzH6SpppWmtkDZ1nmxsHqMbP4SD9l0szKzUzj9IqIZBAFahEZbf6evkc8h87d17r7J9Oxr4GCRzafwd0bgQNmtiLNJYmIyCAUqEUk45jZfDN7wcyWpph9B/DzYLl7zOwHZvZzM9thZp8+y3bbzOzfzGyTma0zs+nB9HIze8zMYsHHiqTtfzF4fVGwTszMHjSztqRNTzCzR83sJTP7Zv+j2wP/YGbrg4+Lg23NM7Nfmdnm4PPcYPrXzexzZvYb4FNm9loz2xh8vND/iGbgh8Bd531gRUQkFArUIpJRzOxVwGPAanePDZhXBTS7e0fS5EXAO4CrgXeYWcUQmx9P3yOzFwK/Bf4imP554D/cfSl9gf2rKdb9PPD5YJn9A+ZdA7wfuAKYDyRfPT7q7suALwL/J5j2ReARd18AfBP4z6TlLwXe4O7/E/gA8Nfuvgh4NX2PggeoDd6LiEgGUKAWkUxSDvwI+DN335hi/kygccC0X7l7q7ufBF4E5g2x/U6gv995A1AZvH4D8EUz2wisBSYlXQ3udx3w/eD1twbMW+/uDe7eC2xM2i7At5M+X5e0rf5tfAO4IWn577t7T/D6WeBzZnY/UOru3cH0w8CsIb5OERFJIwVqEckkrcBeTr/Cm+wEMG7AtOSr1T1Ayt7jQJe7e4plc4Dr3H1R8DHb3Y+dR91D1eCDvGaQ6e2nJvb1b78HKALWmdllwaxx/OFqtYiIREyBWkQySSdwK3C3mb0zxfxXOP3q70h5Criv/42ZLUofgR8fAAABKklEQVSxzDr62kEAVp3Htt+R9LkmeP1c0jbuAp5JtaKZXeTuW9z9U/S1efQH6kuBs46IIiIi6aFALSIZxd3bgZuBvzOzW1LMq+u/uW8E3Q9UBzcJvgi8L8Uy7wf+3szW09d60nqO2y40s98Dfwv8XdL+VpvZZuBdwbxU3m9mW81sE31XpH8WTH8d8NNz3L+IiITM/vDXTxGRzGdmtwFL3P0jad5vMXDC3d3MVgF3uvstZ1svpFp+C9zi7s1R7F9ERE43VK+hiEjGcffHzWxKBLteQt+Niwa0APdGUANmVg58TmFaRCRz6Aq1iIw5QYtF4YDJ73L3LVHUIyIiY5sCtYiIiIjIMOimRBERERGRYVCgFhEREREZBgVqEREREZFhUKAWERERERmG/w9rurGq0N4a5AAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 720x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "max_f1=0\n",
    "k_opt=0\n",
    "f1_scores = []\n",
    "for k in range(1,20):\n",
    "    knn = KNeighborsClassifier(n_neighbors = k)\n",
    "    knn.fit(train_data,train_labels)\n",
    "    predicted_labels = knn.predict(test_data)\n",
    "    f1 = metrics.f1_score(test_labels, predicted_labels)\n",
    "    f1_scores.append(f1)\n",
    "    if f1 > max_f1:\n",
    "        max_f1 = f1\n",
    "        k_opt = k\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(10,7))\n",
    "plt.plot(range(1,20), f1_scores)\n",
    "plt.title('K Neighbors Classifier')\n",
    "plt.xlabel('k (n_neighbors)')\n",
    "plt.ylabel('f1 score')\n",
    "ax.annotate('k={}, f1={}'.format(k_opt,max_f1), (k_opt, f1_scores[k_opt-1]))\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "K Neighbors Classifier (k=18)\n",
      "accuracy_score=0.9672489082969432\n",
      "precision=0.6956521739130435\n",
      "recall=0.6666666666666666\n",
      "f1_score=0.6808510638297872\n",
      "confusion matrix\n",
      "[[427   7]\n",
      " [  8  16]]\n"
     ]
    }
   ],
   "source": [
    "knn = KNeighborsClassifier(n_neighbors = k_opt)\n",
    "knn.fit(train_data,train_labels)\n",
    "predicted_labels = knn.predict(test_data)\n",
    "predicted_2021 = knn.predict(data_2021)\n",
    "\n",
    "print(\"K Neighbors Classifier (k={})\\naccuracy_score={}\\nprecision={}\\nrecall={}\\nf1_score={}\\nconfusion matrix\\n{}\".format(\n",
    "    k_opt,\n",
    "    metrics.accuracy_score(test_labels, predicted_labels),\n",
    "    metrics.precision_score(test_labels, predicted_labels),\n",
    "    metrics.recall_score(test_labels, predicted_labels),\n",
    "    metrics.f1_score(test_labels, predicted_labels),\n",
    "    metrics.confusion_matrix(test_labels, predicted_labels)\n",
    "))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "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>name</th>\n",
       "      <th>year</th>\n",
       "      <th>actual</th>\n",
       "      <th>predicted</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Klay Thompson</td>\n",
       "      <td>2018</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Donovan Mitchell</td>\n",
       "      <td>2019</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Goran Dragić</td>\n",
       "      <td>2017</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>James Harden</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Trae Young</td>\n",
       "      <td>2019</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Giannis Antetokounmpo</td>\n",
       "      <td>2018</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Jaylen Brown</td>\n",
       "      <td>2020</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Paul George</td>\n",
       "      <td>2018</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Mike Conley</td>\n",
       "      <td>2020</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Andre Drummond</td>\n",
       "      <td>2017</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>Blake Griffin</td>\n",
       "      <td>2018</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>Pascal Siakam</td>\n",
       "      <td>2019</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>Paul Millsap</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>Kevin Durant</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>Kevin Durant</td>\n",
       "      <td>2017</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>Giannis Antetokounmpo</td>\n",
       "      <td>2017</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>Gordon Hayward</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>Klay Thompson</td>\n",
       "      <td>2017</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>LeBron James</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>Luka Dončić</td>\n",
       "      <td>2019</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>Draymond Green</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>Russell Westbrook</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>Bradley Beal</td>\n",
       "      <td>2020</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>Giannis Antetokounmpo</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>Donovan Mitchell</td>\n",
       "      <td>2018</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>Zach LaVine</td>\n",
       "      <td>2019</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>Zach LaVine</td>\n",
       "      <td>2018</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>Devin Booker</td>\n",
       "      <td>2017</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>Brandon Ingram</td>\n",
       "      <td>2020</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>Jerami Grant</td>\n",
       "      <td>2020</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>Russell Westbrook</td>\n",
       "      <td>2020</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     name  year  actual  predicted\n",
       "0           Klay Thompson  2018       1          0\n",
       "1        Donovan Mitchell  2019       1          1\n",
       "2            Goran Dragić  2017       1          0\n",
       "3            James Harden  2016       1          1\n",
       "4              Trae Young  2019       1          1\n",
       "5   Giannis Antetokounmpo  2018       1          1\n",
       "6            Jaylen Brown  2020       1          1\n",
       "7             Paul George  2018       1          1\n",
       "8             Mike Conley  2020       1          0\n",
       "9          Andre Drummond  2017       1          0\n",
       "10          Blake Griffin  2018       1          1\n",
       "11          Pascal Siakam  2019       1          1\n",
       "12           Paul Millsap  2016       1          0\n",
       "13           Kevin Durant  2016       1          1\n",
       "14           Kevin Durant  2017       1          1\n",
       "15  Giannis Antetokounmpo  2017       1          1\n",
       "16         Gordon Hayward  2016       1          0\n",
       "17          Klay Thompson  2017       1          0\n",
       "18           LeBron James  2016       1          1\n",
       "19            Luka Dončić  2019       1          1\n",
       "20         Draymond Green  2016       1          0\n",
       "21      Russell Westbrook  2016       1          1\n",
       "22           Bradley Beal  2020       1          1\n",
       "23  Giannis Antetokounmpo  2016       1          1\n",
       "24       Donovan Mitchell  2018       0          1\n",
       "25            Zach LaVine  2019       0          1\n",
       "26            Zach LaVine  2018       0          1\n",
       "27           Devin Booker  2017       0          1\n",
       "28         Brandon Ingram  2020       0          1\n",
       "29           Jerami Grant  2020       0          1\n",
       "30      Russell Westbrook  2020       0          1"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "knn_results = test_data.merge(df)\n",
    "knn_results.rename({'allstar':'actual'}, axis=1, inplace=True)\n",
    "knn_results['predicted'] = predicted_labels\n",
    "knn_results.sort_values(by='actual', ascending=False, inplace=True)\n",
    "knn_results[(knn_results['actual']==1) |  (knn_results['predicted']==1)][['name','year','actual','predicted']].reset_index(drop=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "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>name</th>\n",
       "      <th>predicted</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Kevin Durant</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Devin Booker</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Jayson Tatum</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Trae Young</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Donovan Mitchell</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Giannis Antetokounmpo</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Paul George</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Ja Morant</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Anthony Davis</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Joel Embiid</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>Luka Dončić</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>Bradley Beal</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>LeBron James</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>Nikola Jokić</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>Jimmy Butler</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>DeMar DeRozan</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>Zach LaVine</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>Stephen Curry</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     name  predicted\n",
       "0            Kevin Durant          1\n",
       "1            Devin Booker          1\n",
       "2            Jayson Tatum          1\n",
       "3              Trae Young          1\n",
       "4        Donovan Mitchell          1\n",
       "5   Giannis Antetokounmpo          1\n",
       "6             Paul George          1\n",
       "7               Ja Morant          1\n",
       "8           Anthony Davis          1\n",
       "9             Joel Embiid          1\n",
       "10            Luka Dončić          1\n",
       "11           Bradley Beal          1\n",
       "12           LeBron James          1\n",
       "13           Nikola Jokić          1\n",
       "14           Jimmy Butler          1\n",
       "15          DeMar DeRozan          1\n",
       "16            Zach LaVine          1\n",
       "17          Stephen Curry          1"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_2021 = data_2021.merge(d2021)\n",
    "pred_2021['predicted'] = predicted_2021\n",
    "pred_2021.sort_values(by='predicted', ascending=False, inplace=True)\n",
    "pred_2021[pred_2021['predicted']==1][['name','predicted']].reset_index(drop=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Random Forest Classifier - Non-Normalized"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "max_f1=0\n",
    "d_opt=0\n",
    "f1_scores = []\n",
    "for d in range(1,10):\n",
    "    model = RandomForestClassifier(class_weight=\"balanced_subsample\", max_depth=d, n_estimators=100, random_state=42)\n",
    "    model.fit(train_data, train_labels)\n",
    "    predicted_labels = model.predict(test_data)\n",
    "    f1 = metrics.f1_score(test_labels, predicted_labels)\n",
    "    f1_scores.append(f1)\n",
    "    if f1 > max_f1:\n",
    "        max_f1 = f1\n",
    "        d_opt = d\n",
    "\n",
    "fig1, ax1 = plt.subplots(figsize=(10,7))\n",
    "plt.plot(range(1,10), f1_scores)\n",
    "plt.title('Random Forest Classifier')\n",
    "plt.xlabel('d (max_depth)')\n",
    "plt.ylabel('f1 score')\n",
    "ax1.annotate('d={}, f1={}'.format(d_opt,max_f1), (d_opt, f1_scores[d_opt-1]))\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random Forest Classifier (d=9)\n",
      "accuracy_score=0.9672489082969432\n",
      "precision=0.6956521739130435\n",
      "recall=0.6666666666666666\n",
      "f1_score=0.6808510638297872\n",
      "confusion matrix\n",
      "[[427   7]\n",
      " [  8  16]]\n"
     ]
    }
   ],
   "source": [
    "model = RandomForestClassifier(class_weight=\"balanced_subsample\", max_depth=d_opt, n_estimators=100, random_state=42)\n",
    "model.fit(train_data, train_labels)\n",
    "predicted_labels = model.predict(test_data)\n",
    "predicted_2021 = model.predict(data_2021)\n",
    "\n",
    "print(\"Random Forest Classifier (d={})\\naccuracy_score={}\\nprecision={}\\nrecall={}\\nf1_score={}\\nconfusion matrix\\n{}\".format(\n",
    "    d_opt,\n",
    "    metrics.accuracy_score(test_labels, predicted_labels),\n",
    "    metrics.precision_score(test_labels, predicted_labels),\n",
    "    metrics.recall_score(test_labels, predicted_labels),\n",
    "    metrics.f1_score(test_labels, predicted_labels),\n",
    "    metrics.confusion_matrix(test_labels, predicted_labels)\n",
    "))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "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>name</th>\n",
       "      <th>year</th>\n",
       "      <th>actual</th>\n",
       "      <th>predicted</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Klay Thompson</td>\n",
       "      <td>2018</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Donovan Mitchell</td>\n",
       "      <td>2019</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Goran Dragić</td>\n",
       "      <td>2017</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>James Harden</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Trae Young</td>\n",
       "      <td>2019</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Giannis Antetokounmpo</td>\n",
       "      <td>2018</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Jaylen Brown</td>\n",
       "      <td>2020</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Paul George</td>\n",
       "      <td>2018</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Mike Conley</td>\n",
       "      <td>2020</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Andre Drummond</td>\n",
       "      <td>2017</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>Blake Griffin</td>\n",
       "      <td>2018</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>Pascal Siakam</td>\n",
       "      <td>2019</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>Paul Millsap</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>Kevin Durant</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>Kevin Durant</td>\n",
       "      <td>2017</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>Giannis Antetokounmpo</td>\n",
       "      <td>2017</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>Gordon Hayward</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>Klay Thompson</td>\n",
       "      <td>2017</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>LeBron James</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>Luka Dončić</td>\n",
       "      <td>2019</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>Draymond Green</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>Russell Westbrook</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>Bradley Beal</td>\n",
       "      <td>2020</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>Giannis Antetokounmpo</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>DeMar DeRozan</td>\n",
       "      <td>2018</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>Jimmy Butler</td>\n",
       "      <td>2020</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>Blake Griffin</td>\n",
       "      <td>2016</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>Donovan Mitchell</td>\n",
       "      <td>2018</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>Jaylen Brown</td>\n",
       "      <td>2019</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>Brandon Ingram</td>\n",
       "      <td>2020</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>Russell Westbrook</td>\n",
       "      <td>2020</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     name  year  actual  predicted\n",
       "0           Klay Thompson  2018       1          0\n",
       "1        Donovan Mitchell  2019       1          1\n",
       "2            Goran Dragić  2017       1          0\n",
       "3            James Harden  2016       1          1\n",
       "4              Trae Young  2019       1          0\n",
       "5   Giannis Antetokounmpo  2018       1          1\n",
       "6            Jaylen Brown  2020       1          1\n",
       "7             Paul George  2018       1          1\n",
       "8             Mike Conley  2020       1          0\n",
       "9          Andre Drummond  2017       1          0\n",
       "10          Blake Griffin  2018       1          1\n",
       "11          Pascal Siakam  2019       1          1\n",
       "12           Paul Millsap  2016       1          0\n",
       "13           Kevin Durant  2016       1          1\n",
       "14           Kevin Durant  2017       1          1\n",
       "15  Giannis Antetokounmpo  2017       1          1\n",
       "16         Gordon Hayward  2016       1          1\n",
       "17          Klay Thompson  2017       1          0\n",
       "18           LeBron James  2016       1          1\n",
       "19            Luka Dončić  2019       1          1\n",
       "20         Draymond Green  2016       1          0\n",
       "21      Russell Westbrook  2016       1          1\n",
       "22           Bradley Beal  2020       1          1\n",
       "23  Giannis Antetokounmpo  2016       1          1\n",
       "24          DeMar DeRozan  2018       0          1\n",
       "25           Jimmy Butler  2020       0          1\n",
       "26          Blake Griffin  2016       0          1\n",
       "27       Donovan Mitchell  2018       0          1\n",
       "28           Jaylen Brown  2019       0          1\n",
       "29         Brandon Ingram  2020       0          1\n",
       "30      Russell Westbrook  2020       0          1"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rf_results = test_data.merge(df)\n",
    "rf_results.rename({'allstar':'actual'}, axis=1, inplace=True)\n",
    "rf_results['predicted'] = predicted_labels\n",
    "rf_results.sort_values(by='actual', ascending=False, inplace=True)\n",
    "rf_results[(rf_results['actual']==1) |  (rf_results['predicted']==1)][['name','year','actual','predicted']].reset_index(drop=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "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>name</th>\n",
       "      <th>predicted</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Luka Dončić</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Ja Morant</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Paul George</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Jaylen Brown</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Kevin Durant</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Bradley Beal</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Devin Booker</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Stephen Curry</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Anthony Davis</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Giannis Antetokounmpo</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>Nikola Jokić</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>Zach LaVine</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>LeBron James</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>DeMar DeRozan</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     name  predicted\n",
       "0             Luka Dončić          1\n",
       "1               Ja Morant          1\n",
       "2             Paul George          1\n",
       "3            Jaylen Brown          1\n",
       "4            Kevin Durant          1\n",
       "5            Bradley Beal          1\n",
       "6            Devin Booker          1\n",
       "7           Stephen Curry          1\n",
       "8           Anthony Davis          1\n",
       "9   Giannis Antetokounmpo          1\n",
       "10           Nikola Jokić          1\n",
       "11            Zach LaVine          1\n",
       "12           LeBron James          1\n",
       "13          DeMar DeRozan          1"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_2021_rf = data_2021.merge(d2021)\n",
    "pred_2021_rf['predicted'] = predicted_2021\n",
    "pred_2021_rf.sort_values(by='predicted', ascending=False, inplace=True)\n",
    "pred_2021_rf[pred_2021_rf['predicted']==1][['name','predicted']].reset_index(drop=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Logistic Regression Classifier - Non-Normalized"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Logistic Regression Classifier\n",
      "accuracy_score=0.9672489082969432\n",
      "precision=0.7647058823529411\n",
      "recall=0.5416666666666666\n",
      "f1_score=0.6341463414634146\n",
      "confusion matrix\n",
      "[[430   4]\n",
      " [ 11  13]]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/sam/anaconda3/lib/python3.7/site-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\n",
      "STOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n",
      "\n",
      "Increase the number of iterations (max_iter) or scale the data as shown in:\n",
      "    https://scikit-learn.org/stable/modules/preprocessing.html\n",
      "Please also refer to the documentation for alternative solver options:\n",
      "    https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n",
      "  extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n"
     ]
    }
   ],
   "source": [
    "model = LogisticRegression()\n",
    "model.fit(train_data, train_labels)\n",
    "predicted_labels = model.predict(test_data)\n",
    "probabilities = model.predict_proba(test_data)\n",
    "predicted_2021 = model.predict(data_2021)\n",
    "probabilities_2021 = model.predict_proba(data_2021)\n",
    "\n",
    "print(\"Logistic Regression Classifier\\naccuracy_score={}\\nprecision={}\\nrecall={}\\nf1_score={}\\nconfusion matrix\\n{}\".format(\n",
    "    metrics.accuracy_score(test_labels, predicted_labels),\n",
    "    metrics.precision_score(test_labels, predicted_labels),\n",
    "    metrics.recall_score(test_labels, predicted_labels),\n",
    "    metrics.f1_score(test_labels, predicted_labels),\n",
    "    metrics.confusion_matrix(test_labels, predicted_labels)\n",
    "))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "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>name</th>\n",
       "      <th>year</th>\n",
       "      <th>actual</th>\n",
       "      <th>prob_as</th>\n",
       "      <th>prob_non</th>\n",
       "      <th>predicted</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Russell Westbrook</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>0.998654</td>\n",
       "      <td>0.001346</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Giannis Antetokounmpo</td>\n",
       "      <td>2018</td>\n",
       "      <td>1</td>\n",
       "      <td>0.996200</td>\n",
       "      <td>0.003800</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>James Harden</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>0.993901</td>\n",
       "      <td>0.006099</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Kevin Durant</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>0.986933</td>\n",
       "      <td>0.013067</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Luka Dončić</td>\n",
       "      <td>2019</td>\n",
       "      <td>1</td>\n",
       "      <td>0.986886</td>\n",
       "      <td>0.013114</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>LeBron James</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>0.986138</td>\n",
       "      <td>0.013862</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Kevin Durant</td>\n",
       "      <td>2017</td>\n",
       "      <td>1</td>\n",
       "      <td>0.978423</td>\n",
       "      <td>0.021577</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Russell Westbrook</td>\n",
       "      <td>2020</td>\n",
       "      <td>0</td>\n",
       "      <td>0.965011</td>\n",
       "      <td>0.034989</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Giannis Antetokounmpo</td>\n",
       "      <td>2017</td>\n",
       "      <td>1</td>\n",
       "      <td>0.934428</td>\n",
       "      <td>0.065572</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Trae Young</td>\n",
       "      <td>2019</td>\n",
       "      <td>1</td>\n",
       "      <td>0.930983</td>\n",
       "      <td>0.069017</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>Paul George</td>\n",
       "      <td>2018</td>\n",
       "      <td>1</td>\n",
       "      <td>0.878088</td>\n",
       "      <td>0.121912</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>Jimmy Butler</td>\n",
       "      <td>2020</td>\n",
       "      <td>0</td>\n",
       "      <td>0.820719</td>\n",
       "      <td>0.179281</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>Bradley Beal</td>\n",
       "      <td>2020</td>\n",
       "      <td>1</td>\n",
       "      <td>0.809048</td>\n",
       "      <td>0.190952</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>Giannis Antetokounmpo</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>0.759341</td>\n",
       "      <td>0.240659</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>Blake Griffin</td>\n",
       "      <td>2018</td>\n",
       "      <td>1</td>\n",
       "      <td>0.601493</td>\n",
       "      <td>0.398507</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>DeMar DeRozan</td>\n",
       "      <td>2018</td>\n",
       "      <td>0</td>\n",
       "      <td>0.586416</td>\n",
       "      <td>0.413584</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>Blake Griffin</td>\n",
       "      <td>2016</td>\n",
       "      <td>0</td>\n",
       "      <td>0.520698</td>\n",
       "      <td>0.479302</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>Pascal Siakam</td>\n",
       "      <td>2019</td>\n",
       "      <td>1</td>\n",
       "      <td>0.451974</td>\n",
       "      <td>0.548026</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>Gordon Hayward</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>0.434211</td>\n",
       "      <td>0.565789</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>Jaylen Brown</td>\n",
       "      <td>2020</td>\n",
       "      <td>1</td>\n",
       "      <td>0.339341</td>\n",
       "      <td>0.660659</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>Donovan Mitchell</td>\n",
       "      <td>2019</td>\n",
       "      <td>1</td>\n",
       "      <td>0.226738</td>\n",
       "      <td>0.773262</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>Andre Drummond</td>\n",
       "      <td>2017</td>\n",
       "      <td>1</td>\n",
       "      <td>0.204056</td>\n",
       "      <td>0.795944</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>Mike Conley</td>\n",
       "      <td>2020</td>\n",
       "      <td>1</td>\n",
       "      <td>0.164757</td>\n",
       "      <td>0.835243</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>Klay Thompson</td>\n",
       "      <td>2018</td>\n",
       "      <td>1</td>\n",
       "      <td>0.158299</td>\n",
       "      <td>0.841701</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>Klay Thompson</td>\n",
       "      <td>2017</td>\n",
       "      <td>1</td>\n",
       "      <td>0.144518</td>\n",
       "      <td>0.855482</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>Draymond Green</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>0.121297</td>\n",
       "      <td>0.878703</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>Paul Millsap</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>0.111189</td>\n",
       "      <td>0.888811</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>Goran Dragić</td>\n",
       "      <td>2017</td>\n",
       "      <td>1</td>\n",
       "      <td>0.050304</td>\n",
       "      <td>0.949696</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     name  year  actual   prob_as  prob_non  predicted\n",
       "0       Russell Westbrook  2016       1  0.998654  0.001346          1\n",
       "1   Giannis Antetokounmpo  2018       1  0.996200  0.003800          1\n",
       "2            James Harden  2016       1  0.993901  0.006099          1\n",
       "3            Kevin Durant  2016       1  0.986933  0.013067          1\n",
       "4             Luka Dončić  2019       1  0.986886  0.013114          1\n",
       "5            LeBron James  2016       1  0.986138  0.013862          1\n",
       "6            Kevin Durant  2017       1  0.978423  0.021577          1\n",
       "7       Russell Westbrook  2020       0  0.965011  0.034989          1\n",
       "8   Giannis Antetokounmpo  2017       1  0.934428  0.065572          1\n",
       "9              Trae Young  2019       1  0.930983  0.069017          1\n",
       "10            Paul George  2018       1  0.878088  0.121912          1\n",
       "11           Jimmy Butler  2020       0  0.820719  0.179281          1\n",
       "12           Bradley Beal  2020       1  0.809048  0.190952          1\n",
       "13  Giannis Antetokounmpo  2016       1  0.759341  0.240659          1\n",
       "14          Blake Griffin  2018       1  0.601493  0.398507          1\n",
       "15          DeMar DeRozan  2018       0  0.586416  0.413584          1\n",
       "16          Blake Griffin  2016       0  0.520698  0.479302          1\n",
       "17          Pascal Siakam  2019       1  0.451974  0.548026          0\n",
       "18         Gordon Hayward  2016       1  0.434211  0.565789          0\n",
       "19           Jaylen Brown  2020       1  0.339341  0.660659          0\n",
       "20       Donovan Mitchell  2019       1  0.226738  0.773262          0\n",
       "21         Andre Drummond  2017       1  0.204056  0.795944          0\n",
       "22            Mike Conley  2020       1  0.164757  0.835243          0\n",
       "23          Klay Thompson  2018       1  0.158299  0.841701          0\n",
       "24          Klay Thompson  2017       1  0.144518  0.855482          0\n",
       "25         Draymond Green  2016       1  0.121297  0.878703          0\n",
       "26           Paul Millsap  2016       1  0.111189  0.888811          0\n",
       "27           Goran Dragić  2017       1  0.050304  0.949696          0"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results = test_data.merge(df)\n",
    "results.rename({'allstar':'actual'}, axis=1, inplace=True)\n",
    "results['predicted'] = predicted_labels\n",
    "results['prob_as'] = probabilities[:,1]\n",
    "results['prob_non'] = probabilities[:,0]\n",
    "results.sort_values(by='prob_as', ascending=False, inplace=True)\n",
    "results[(results['actual']==1) |  (results['predicted']==1)][['name','year','actual','prob_as','prob_non','predicted']].reset_index(drop=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "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>name</th>\n",
       "      <th>predicted</th>\n",
       "      <th>prob_as</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Nikola Jokić</td>\n",
       "      <td>1</td>\n",
       "      <td>0.996528</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Giannis Antetokounmpo</td>\n",
       "      <td>1</td>\n",
       "      <td>0.996321</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Kevin Durant</td>\n",
       "      <td>1</td>\n",
       "      <td>0.995401</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Stephen Curry</td>\n",
       "      <td>1</td>\n",
       "      <td>0.992830</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Luka Dončić</td>\n",
       "      <td>1</td>\n",
       "      <td>0.948731</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Joel Embiid</td>\n",
       "      <td>1</td>\n",
       "      <td>0.933463</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Trae Young</td>\n",
       "      <td>1</td>\n",
       "      <td>0.917845</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Ja Morant</td>\n",
       "      <td>1</td>\n",
       "      <td>0.881827</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>James Harden</td>\n",
       "      <td>1</td>\n",
       "      <td>0.881747</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>LeBron James</td>\n",
       "      <td>1</td>\n",
       "      <td>0.864427</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>Jimmy Butler</td>\n",
       "      <td>1</td>\n",
       "      <td>0.855466</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>Dejounte Murray</td>\n",
       "      <td>1</td>\n",
       "      <td>0.838802</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>Paul George</td>\n",
       "      <td>1</td>\n",
       "      <td>0.838015</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>Anthony Davis</td>\n",
       "      <td>1</td>\n",
       "      <td>0.820760</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>DeMar DeRozan</td>\n",
       "      <td>1</td>\n",
       "      <td>0.808577</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>Zach LaVine</td>\n",
       "      <td>1</td>\n",
       "      <td>0.757607</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>Rudy Gobert</td>\n",
       "      <td>1</td>\n",
       "      <td>0.711683</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>Devin Booker</td>\n",
       "      <td>1</td>\n",
       "      <td>0.645835</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>Russell Westbrook</td>\n",
       "      <td>1</td>\n",
       "      <td>0.624074</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>Chris Paul</td>\n",
       "      <td>1</td>\n",
       "      <td>0.621129</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>Tobias Harris</td>\n",
       "      <td>1</td>\n",
       "      <td>0.603189</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>LaMelo Ball</td>\n",
       "      <td>1</td>\n",
       "      <td>0.552026</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     name  predicted   prob_as\n",
       "0            Nikola Jokić          1  0.996528\n",
       "1   Giannis Antetokounmpo          1  0.996321\n",
       "2            Kevin Durant          1  0.995401\n",
       "3           Stephen Curry          1  0.992830\n",
       "4             Luka Dončić          1  0.948731\n",
       "5             Joel Embiid          1  0.933463\n",
       "6              Trae Young          1  0.917845\n",
       "7               Ja Morant          1  0.881827\n",
       "8            James Harden          1  0.881747\n",
       "9            LeBron James          1  0.864427\n",
       "10           Jimmy Butler          1  0.855466\n",
       "11        Dejounte Murray          1  0.838802\n",
       "12            Paul George          1  0.838015\n",
       "13          Anthony Davis          1  0.820760\n",
       "14          DeMar DeRozan          1  0.808577\n",
       "15            Zach LaVine          1  0.757607\n",
       "16            Rudy Gobert          1  0.711683\n",
       "17           Devin Booker          1  0.645835\n",
       "18      Russell Westbrook          1  0.624074\n",
       "19             Chris Paul          1  0.621129\n",
       "20          Tobias Harris          1  0.603189\n",
       "21            LaMelo Ball          1  0.552026"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_2021 = data_2021.merge(d2021)\n",
    "pred_2021['predicted'] = predicted_2021\n",
    "pred_2021['prob_as'] = probabilities_2021[:,1]\n",
    "pred_2021['prob_non'] = probabilities_2021[:,0]\n",
    "pred_2021.sort_values(by='prob_as', ascending=False, inplace=True)\n",
    "pred_2021[pred_2021['predicted']==1][['name','predicted', 'prob_as']].reset_index(drop=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "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>feature</th>\n",
       "      <th>weight</th>\n",
       "      <th>weight_abs</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>win%</td>\n",
       "      <td>2.423428</td>\n",
       "      <td>2.423428</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>FT%</td>\n",
       "      <td>-1.895922</td>\n",
       "      <td>1.895922</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2P%</td>\n",
       "      <td>-1.519178</td>\n",
       "      <td>1.519178</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>eFG%</td>\n",
       "      <td>-1.480835</td>\n",
       "      <td>1.480835</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>3P%</td>\n",
       "      <td>-1.400711</td>\n",
       "      <td>1.400711</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>FG%</td>\n",
       "      <td>-1.284770</td>\n",
       "      <td>1.284770</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>GS%</td>\n",
       "      <td>1.064943</td>\n",
       "      <td>1.064943</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>PF</td>\n",
       "      <td>-0.668026</td>\n",
       "      <td>0.668026</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>PTS</td>\n",
       "      <td>0.650201</td>\n",
       "      <td>0.650201</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>ORB</td>\n",
       "      <td>-0.561942</td>\n",
       "      <td>0.561942</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>3P</td>\n",
       "      <td>0.477348</td>\n",
       "      <td>0.477348</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>BLK</td>\n",
       "      <td>0.449611</td>\n",
       "      <td>0.449611</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>FG</td>\n",
       "      <td>0.435584</td>\n",
       "      <td>0.435584</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>3PA</td>\n",
       "      <td>-0.430734</td>\n",
       "      <td>0.430734</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>AST</td>\n",
       "      <td>0.420170</td>\n",
       "      <td>0.420170</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>2P</td>\n",
       "      <td>-0.415640</td>\n",
       "      <td>0.415640</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>STL</td>\n",
       "      <td>0.396478</td>\n",
       "      <td>0.396478</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>DRB</td>\n",
       "      <td>0.371954</td>\n",
       "      <td>0.371954</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>FGA</td>\n",
       "      <td>-0.295990</td>\n",
       "      <td>0.295990</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>TOV</td>\n",
       "      <td>-0.203893</td>\n",
       "      <td>0.203893</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>TRB</td>\n",
       "      <td>0.168577</td>\n",
       "      <td>0.168577</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>MP</td>\n",
       "      <td>-0.147716</td>\n",
       "      <td>0.147716</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>FTA</td>\n",
       "      <td>-0.101183</td>\n",
       "      <td>0.101183</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>FT</td>\n",
       "      <td>-0.030541</td>\n",
       "      <td>0.030541</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>2PA</td>\n",
       "      <td>-0.015882</td>\n",
       "      <td>0.015882</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>Age</td>\n",
       "      <td>0.007619</td>\n",
       "      <td>0.007619</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   feature    weight  weight_abs\n",
       "0     win%  2.423428    2.423428\n",
       "1      FT% -1.895922    1.895922\n",
       "2      2P% -1.519178    1.519178\n",
       "3     eFG% -1.480835    1.480835\n",
       "4      3P% -1.400711    1.400711\n",
       "5      FG% -1.284770    1.284770\n",
       "6      GS%  1.064943    1.064943\n",
       "7       PF -0.668026    0.668026\n",
       "8      PTS  0.650201    0.650201\n",
       "9      ORB -0.561942    0.561942\n",
       "10      3P  0.477348    0.477348\n",
       "11     BLK  0.449611    0.449611\n",
       "12      FG  0.435584    0.435584\n",
       "13     3PA -0.430734    0.430734\n",
       "14     AST  0.420170    0.420170\n",
       "15      2P -0.415640    0.415640\n",
       "16     STL  0.396478    0.396478\n",
       "17     DRB  0.371954    0.371954\n",
       "18     FGA -0.295990    0.295990\n",
       "19     TOV -0.203893    0.203893\n",
       "20     TRB  0.168577    0.168577\n",
       "21      MP -0.147716    0.147716\n",
       "22     FTA -0.101183    0.101183\n",
       "23      FT -0.030541    0.030541\n",
       "24     2PA -0.015882    0.015882\n",
       "25     Age  0.007619    0.007619"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "coefs_df = pd.DataFrame({'feature':data_cols,'weight':model.coef_[0]})\n",
    "coefs_df['weight_abs'] = coefs_df.weight.abs()\n",
    "coefs_df.sort_values(by='weight_abs', ascending=False, inplace=True)\n",
    "coefs_df.reset_index(inplace=True, drop=True)\n",
    "coefs_df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Normalize Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "names = list(df['name'])\n",
    "allstars = list(df['allstar'])\n",
    "years = list(df['year'])\n",
    "names_2021 = list(d2021['name'])\n",
    "scaled_data = scale(data, axis=0)\n",
    "scaled_names = pd.DataFrame(scaled_data)\n",
    "scaled_names.columns = data_cols\n",
    "scaled_names['name'] = names\n",
    "scaled_names['actual'] = allstars\n",
    "scaled_names['year'] = years\n",
    "scaled_data_2021 = scale(data_2021, axis=0)\n",
    "scaled_names_2021 = pd.DataFrame(scaled_data_2021)\n",
    "scaled_names_2021.columns = data_cols\n",
    "scaled_names_2021['name'] = names_2021\n",
    "train_data, test_data, train_labels, test_labels = train_test_split(scaled_data, labels, test_size=0.2, random_state=42, stratify=labels)\n",
    "test_data_df = pd.DataFrame(test_data)\n",
    "test_data_df.columns = data_cols\n",
    "\n",
    "# test_data_df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### KNN Classifier - Normalized"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "max_f1=0\n",
    "k_opt=0\n",
    "f1_scores = []\n",
    "for k in range(2,20):\n",
    "    knn = KNeighborsClassifier(n_neighbors = k)\n",
    "    knn.fit(train_data,train_labels)\n",
    "    predicted_labels = knn.predict(test_data)\n",
    "    f1 = metrics.f1_score(test_labels, predicted_labels)\n",
    "    f1_scores.append(f1)\n",
    "    if f1 > max_f1:\n",
    "        max_f1 = f1\n",
    "        k_opt = k\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(10,7))\n",
    "plt.plot(range(2,20), f1_scores)\n",
    "plt.title('K Neighbors Classifier')\n",
    "plt.xlabel('k (n_neighbors)')\n",
    "plt.ylabel('f1 score')\n",
    "ax.annotate('k={}, f1={}'.format(k_opt,max_f1), (k_opt, f1_scores[k_opt-1]))\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "K Neighbors Classifier (k=15)\n",
      "accuracy_score=0.9716157205240175\n",
      "precision=0.7619047619047619\n",
      "recall=0.6666666666666666\n",
      "f1_score=0.7111111111111111\n",
      "confusion matrix\n",
      "[[429   5]\n",
      " [  8  16]]\n"
     ]
    }
   ],
   "source": [
    "knn = KNeighborsClassifier(n_neighbors = k_opt)\n",
    "knn.fit(train_data,train_labels)\n",
    "predicted_labels = knn.predict(test_data)\n",
    "predicted_2021 = knn.predict(scaled_data_2021)\n",
    "\n",
    "print(\"K Neighbors Classifier (k={})\\naccuracy_score={}\\nprecision={}\\nrecall={}\\nf1_score={}\\nconfusion matrix\\n{}\".format(\n",
    "    k_opt,\n",
    "    metrics.accuracy_score(test_labels, predicted_labels),\n",
    "    metrics.precision_score(test_labels, predicted_labels),\n",
    "    metrics.recall_score(test_labels, predicted_labels),\n",
    "    metrics.f1_score(test_labels, predicted_labels),\n",
    "    metrics.confusion_matrix(test_labels, predicted_labels)\n",
    "))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "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>name</th>\n",
       "      <th>year</th>\n",
       "      <th>actual</th>\n",
       "      <th>predicted</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Klay Thompson</td>\n",
       "      <td>2018</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Donovan Mitchell</td>\n",
       "      <td>2019</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Goran Dragić</td>\n",
       "      <td>2017</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>James Harden</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Trae Young</td>\n",
       "      <td>2019</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Giannis Antetokounmpo</td>\n",
       "      <td>2018</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Jaylen Brown</td>\n",
       "      <td>2020</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Paul George</td>\n",
       "      <td>2018</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Mike Conley</td>\n",
       "      <td>2020</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Andre Drummond</td>\n",
       "      <td>2017</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>Blake Griffin</td>\n",
       "      <td>2018</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>Pascal Siakam</td>\n",
       "      <td>2019</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>Paul Millsap</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>Kevin Durant</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>Kevin Durant</td>\n",
       "      <td>2017</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>Giannis Antetokounmpo</td>\n",
       "      <td>2017</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>Gordon Hayward</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>Klay Thompson</td>\n",
       "      <td>2017</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>LeBron James</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>Luka Dončić</td>\n",
       "      <td>2019</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>Draymond Green</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>Russell Westbrook</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>Bradley Beal</td>\n",
       "      <td>2020</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>Giannis Antetokounmpo</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>Jimmy Butler</td>\n",
       "      <td>2020</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>Donovan Mitchell</td>\n",
       "      <td>2018</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>Zach LaVine</td>\n",
       "      <td>2019</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>Brandon Ingram</td>\n",
       "      <td>2020</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>Russell Westbrook</td>\n",
       "      <td>2020</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     name  year  actual  predicted\n",
       "0           Klay Thompson  2018       1          0\n",
       "1        Donovan Mitchell  2019       1          1\n",
       "2            Goran Dragić  2017       1          0\n",
       "3            James Harden  2016       1          1\n",
       "4              Trae Young  2019       1          1\n",
       "5   Giannis Antetokounmpo  2018       1          1\n",
       "6            Jaylen Brown  2020       1          1\n",
       "7             Paul George  2018       1          1\n",
       "8             Mike Conley  2020       1          0\n",
       "9          Andre Drummond  2017       1          0\n",
       "10          Blake Griffin  2018       1          1\n",
       "11          Pascal Siakam  2019       1          1\n",
       "12           Paul Millsap  2016       1          0\n",
       "13           Kevin Durant  2016       1          1\n",
       "14           Kevin Durant  2017       1          1\n",
       "15  Giannis Antetokounmpo  2017       1          1\n",
       "16         Gordon Hayward  2016       1          0\n",
       "17          Klay Thompson  2017       1          0\n",
       "18           LeBron James  2016       1          1\n",
       "19            Luka Dončić  2019       1          1\n",
       "20         Draymond Green  2016       1          0\n",
       "21      Russell Westbrook  2016       1          1\n",
       "22           Bradley Beal  2020       1          1\n",
       "23  Giannis Antetokounmpo  2016       1          1\n",
       "24           Jimmy Butler  2020       0          1\n",
       "25       Donovan Mitchell  2018       0          1\n",
       "26            Zach LaVine  2019       0          1\n",
       "27         Brandon Ingram  2020       0          1\n",
       "28      Russell Westbrook  2020       0          1"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "knn_results = test_data_df.merge(scaled_names)\n",
    "knn_results.rename({'allstar':'actual'}, axis=1, inplace=True)\n",
    "knn_results['predicted'] = predicted_labels\n",
    "knn_results.sort_values(by='actual', ascending=False, inplace=True)\n",
    "knn_results[(knn_results['actual']==1) |  (knn_results['predicted']==1)][['name','year','actual','predicted']].reset_index(drop=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "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>name</th>\n",
       "      <th>predicted</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>LeBron James</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Ja Morant</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Jimmy Butler</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Damian Lillard</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Trae Young</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Kevin Durant</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>James Harden</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Giannis Antetokounmpo</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Stephen Curry</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Anthony Davis</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>Karl-Anthony Towns</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>Russell Westbrook</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>Anthony Edwards</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>Bradley Beal</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>Paul George</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>Devin Booker</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>Donovan Mitchell</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>Joel Embiid</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>Nikola Jokić</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>Julius Randle</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>DeMar DeRozan</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>Bam Adebayo</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>Zach LaVine</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>Luka Dončić</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>Jayson Tatum</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     name  predicted\n",
       "0            LeBron James          1\n",
       "1               Ja Morant          1\n",
       "2            Jimmy Butler          1\n",
       "3          Damian Lillard          1\n",
       "4              Trae Young          1\n",
       "5            Kevin Durant          1\n",
       "6            James Harden          1\n",
       "7   Giannis Antetokounmpo          1\n",
       "8           Stephen Curry          1\n",
       "9           Anthony Davis          1\n",
       "10     Karl-Anthony Towns          1\n",
       "11      Russell Westbrook          1\n",
       "12        Anthony Edwards          1\n",
       "13           Bradley Beal          1\n",
       "14            Paul George          1\n",
       "15           Devin Booker          1\n",
       "16       Donovan Mitchell          1\n",
       "17            Joel Embiid          1\n",
       "18           Nikola Jokić          1\n",
       "19          Julius Randle          1\n",
       "20          DeMar DeRozan          1\n",
       "21            Bam Adebayo          1\n",
       "22            Zach LaVine          1\n",
       "23            Luka Dončić          1\n",
       "24           Jayson Tatum          1"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_2021 = scaled_names_2021.copy()\n",
    "pred_2021['predicted'] = predicted_2021\n",
    "pred_2021.sort_values(by='predicted', ascending=False, inplace=True)\n",
    "pred_2021[pred_2021['predicted']==1][['name','predicted']].reset_index(drop=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Random Forest Classifier - Normalized"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "max_f1=0\n",
    "d_opt=0\n",
    "f1_scores = []\n",
    "for d in range(1,10):\n",
    "    model = RandomForestClassifier(class_weight=\"balanced_subsample\", max_depth=d, n_estimators=100, random_state=42)\n",
    "    model.fit(train_data, train_labels)\n",
    "    predicted_labels = model.predict(test_data)\n",
    "    f1 = metrics.f1_score(test_labels, predicted_labels)\n",
    "    f1_scores.append(f1)\n",
    "    if f1 > max_f1:\n",
    "        max_f1 = f1\n",
    "        d_opt = d\n",
    "\n",
    "fig1, ax1 = plt.subplots(figsize=(10,7))\n",
    "plt.plot(range(1,10), f1_scores)\n",
    "plt.title('Random Forest Classifier')\n",
    "plt.xlabel('d (max_depth)')\n",
    "plt.ylabel('f1 score')\n",
    "ax1.annotate('d={}, f1={}'.format(d_opt,max_f1), (d_opt, f1_scores[d_opt-1]))\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random Forest Classifier (d=9)\n",
      "accuracy_score=0.9672489082969432\n",
      "precision=0.6956521739130435\n",
      "recall=0.6666666666666666\n",
      "f1_score=0.6808510638297872\n",
      "confusion matrix\n",
      "[[427   7]\n",
      " [  8  16]]\n"
     ]
    }
   ],
   "source": [
    "model = RandomForestClassifier(class_weight=\"balanced_subsample\", max_depth=d_opt, n_estimators=100, random_state=42)\n",
    "model.fit(train_data, train_labels)\n",
    "predicted_labels = model.predict(test_data)\n",
    "predicted_2021 = model.predict(scaled_data_2021)\n",
    "\n",
    "print(\"Random Forest Classifier (d={})\\naccuracy_score={}\\nprecision={}\\nrecall={}\\nf1_score={}\\nconfusion matrix\\n{}\".format(\n",
    "    d_opt,\n",
    "    metrics.accuracy_score(test_labels, predicted_labels),\n",
    "    metrics.precision_score(test_labels, predicted_labels),\n",
    "    metrics.recall_score(test_labels, predicted_labels),\n",
    "    metrics.f1_score(test_labels, predicted_labels),\n",
    "    metrics.confusion_matrix(test_labels, predicted_labels)\n",
    "))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>name</th>\n",
       "      <th>year</th>\n",
       "      <th>actual</th>\n",
       "      <th>predicted</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Klay Thompson</td>\n",
       "      <td>2018</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Donovan Mitchell</td>\n",
       "      <td>2019</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Goran Dragić</td>\n",
       "      <td>2017</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>James Harden</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Trae Young</td>\n",
       "      <td>2019</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Giannis Antetokounmpo</td>\n",
       "      <td>2018</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Jaylen Brown</td>\n",
       "      <td>2020</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Paul George</td>\n",
       "      <td>2018</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Mike Conley</td>\n",
       "      <td>2020</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Andre Drummond</td>\n",
       "      <td>2017</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>Blake Griffin</td>\n",
       "      <td>2018</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>Pascal Siakam</td>\n",
       "      <td>2019</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>Paul Millsap</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>Kevin Durant</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>Kevin Durant</td>\n",
       "      <td>2017</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>Giannis Antetokounmpo</td>\n",
       "      <td>2017</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>Gordon Hayward</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>Klay Thompson</td>\n",
       "      <td>2017</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>LeBron James</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>Luka Dončić</td>\n",
       "      <td>2019</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>Draymond Green</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>Russell Westbrook</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>Bradley Beal</td>\n",
       "      <td>2020</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>Giannis Antetokounmpo</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>DeMar DeRozan</td>\n",
       "      <td>2018</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>Jimmy Butler</td>\n",
       "      <td>2020</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>Blake Griffin</td>\n",
       "      <td>2016</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>Donovan Mitchell</td>\n",
       "      <td>2018</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>Jaylen Brown</td>\n",
       "      <td>2019</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>Brandon Ingram</td>\n",
       "      <td>2020</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>Russell Westbrook</td>\n",
       "      <td>2020</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     name  year  actual  predicted\n",
       "0           Klay Thompson  2018       1          0\n",
       "1        Donovan Mitchell  2019       1          1\n",
       "2            Goran Dragić  2017       1          0\n",
       "3            James Harden  2016       1          1\n",
       "4              Trae Young  2019       1          0\n",
       "5   Giannis Antetokounmpo  2018       1          1\n",
       "6            Jaylen Brown  2020       1          1\n",
       "7             Paul George  2018       1          1\n",
       "8             Mike Conley  2020       1          0\n",
       "9          Andre Drummond  2017       1          0\n",
       "10          Blake Griffin  2018       1          1\n",
       "11          Pascal Siakam  2019       1          1\n",
       "12           Paul Millsap  2016       1          0\n",
       "13           Kevin Durant  2016       1          1\n",
       "14           Kevin Durant  2017       1          1\n",
       "15  Giannis Antetokounmpo  2017       1          1\n",
       "16         Gordon Hayward  2016       1          1\n",
       "17          Klay Thompson  2017       1          0\n",
       "18           LeBron James  2016       1          1\n",
       "19            Luka Dončić  2019       1          1\n",
       "20         Draymond Green  2016       1          0\n",
       "21      Russell Westbrook  2016       1          1\n",
       "22           Bradley Beal  2020       1          1\n",
       "23  Giannis Antetokounmpo  2016       1          1\n",
       "24          DeMar DeRozan  2018       0          1\n",
       "25           Jimmy Butler  2020       0          1\n",
       "26          Blake Griffin  2016       0          1\n",
       "27       Donovan Mitchell  2018       0          1\n",
       "28           Jaylen Brown  2019       0          1\n",
       "29         Brandon Ingram  2020       0          1\n",
       "30      Russell Westbrook  2020       0          1"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rf_results = test_data_df.merge(scaled_names)\n",
    "rf_results.rename({'allstar':'actual'}, axis=1, inplace=True)\n",
    "rf_results['predicted'] = predicted_labels\n",
    "rf_results.sort_values(by='actual', ascending=False, inplace=True)\n",
    "rf_results[(rf_results['actual']==1) |  (rf_results['predicted']==1)][['name','year','actual','predicted']].reset_index(drop=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "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>name</th>\n",
       "      <th>predicted</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Luka Dončić</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Ja Morant</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Paul George</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Jaylen Brown</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Kevin Durant</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Bradley Beal</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Devin Booker</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Stephen Curry</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Anthony Davis</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Giannis Antetokounmpo</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>Nikola Jokić</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>Zach LaVine</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>LeBron James</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>DeMar DeRozan</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     name  predicted\n",
       "0             Luka Dončić          1\n",
       "1               Ja Morant          1\n",
       "2             Paul George          1\n",
       "3            Jaylen Brown          1\n",
       "4            Kevin Durant          1\n",
       "5            Bradley Beal          1\n",
       "6            Devin Booker          1\n",
       "7           Stephen Curry          1\n",
       "8           Anthony Davis          1\n",
       "9   Giannis Antetokounmpo          1\n",
       "10           Nikola Jokić          1\n",
       "11            Zach LaVine          1\n",
       "12           LeBron James          1\n",
       "13          DeMar DeRozan          1"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_2021_rf = scaled_names_2021.copy()\n",
    "pred_2021_rf['predicted'] = predicted_2021\n",
    "pred_2021_rf.sort_values(by='predicted', ascending=False, inplace=True)\n",
    "pred_2021_rf[pred_2021_rf['predicted']==1][['name','predicted']].reset_index(drop=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Logistic Regression Classifier - Normalized"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Logistic Regression Classifier\n",
      "accuracy_score=0.9716157205240175\n",
      "precision=0.7894736842105263\n",
      "recall=0.625\n",
      "f1_score=0.6976744186046512\n",
      "confusion matrix\n",
      "[[430   4]\n",
      " [  9  15]]\n"
     ]
    }
   ],
   "source": [
    "model = LogisticRegression()\n",
    "model.fit(train_data, train_labels)\n",
    "predicted_labels = model.predict(test_data)\n",
    "probabilities = model.predict_proba(test_data)\n",
    "predicted_2021 = model.predict(scaled_data_2021)\n",
    "probabilities_2021 = model.predict_proba(scaled_data_2021)\n",
    "\n",
    "print(\"Logistic Regression Classifier\\naccuracy_score={}\\nprecision={}\\nrecall={}\\nf1_score={}\\nconfusion matrix\\n{}\".format(\n",
    "    metrics.accuracy_score(test_labels, predicted_labels),\n",
    "    metrics.precision_score(test_labels, predicted_labels),\n",
    "    metrics.recall_score(test_labels, predicted_labels),\n",
    "    metrics.f1_score(test_labels, predicted_labels),\n",
    "    metrics.confusion_matrix(test_labels, predicted_labels)\n",
    "))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "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>name</th>\n",
       "      <th>year</th>\n",
       "      <th>actual</th>\n",
       "      <th>prob_as</th>\n",
       "      <th>prob_non</th>\n",
       "      <th>predicted</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Russell Westbrook</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>0.999659</td>\n",
       "      <td>0.000341</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Giannis Antetokounmpo</td>\n",
       "      <td>2018</td>\n",
       "      <td>1</td>\n",
       "      <td>0.998966</td>\n",
       "      <td>0.001034</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>James Harden</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>0.998927</td>\n",
       "      <td>0.001073</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>LeBron James</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>0.998849</td>\n",
       "      <td>0.001151</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Kevin Durant</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>0.997083</td>\n",
       "      <td>0.002917</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Kevin Durant</td>\n",
       "      <td>2017</td>\n",
       "      <td>1</td>\n",
       "      <td>0.992692</td>\n",
       "      <td>0.007308</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Luka Dončić</td>\n",
       "      <td>2019</td>\n",
       "      <td>1</td>\n",
       "      <td>0.991790</td>\n",
       "      <td>0.008210</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Russell Westbrook</td>\n",
       "      <td>2020</td>\n",
       "      <td>0</td>\n",
       "      <td>0.991608</td>\n",
       "      <td>0.008392</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Paul George</td>\n",
       "      <td>2018</td>\n",
       "      <td>1</td>\n",
       "      <td>0.959714</td>\n",
       "      <td>0.040286</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Giannis Antetokounmpo</td>\n",
       "      <td>2017</td>\n",
       "      <td>1</td>\n",
       "      <td>0.947542</td>\n",
       "      <td>0.052458</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>Pascal Siakam</td>\n",
       "      <td>2019</td>\n",
       "      <td>1</td>\n",
       "      <td>0.864835</td>\n",
       "      <td>0.135165</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>Bradley Beal</td>\n",
       "      <td>2020</td>\n",
       "      <td>1</td>\n",
       "      <td>0.862850</td>\n",
       "      <td>0.137150</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>Blake Griffin</td>\n",
       "      <td>2018</td>\n",
       "      <td>1</td>\n",
       "      <td>0.798908</td>\n",
       "      <td>0.201092</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>Jimmy Butler</td>\n",
       "      <td>2020</td>\n",
       "      <td>0</td>\n",
       "      <td>0.775874</td>\n",
       "      <td>0.224126</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>Trae Young</td>\n",
       "      <td>2019</td>\n",
       "      <td>1</td>\n",
       "      <td>0.729620</td>\n",
       "      <td>0.270380</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>DeMar DeRozan</td>\n",
       "      <td>2018</td>\n",
       "      <td>0</td>\n",
       "      <td>0.724669</td>\n",
       "      <td>0.275331</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>Blake Griffin</td>\n",
       "      <td>2016</td>\n",
       "      <td>0</td>\n",
       "      <td>0.704160</td>\n",
       "      <td>0.295840</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>Giannis Antetokounmpo</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>0.672419</td>\n",
       "      <td>0.327581</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>Gordon Hayward</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>0.564794</td>\n",
       "      <td>0.435206</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>Klay Thompson</td>\n",
       "      <td>2018</td>\n",
       "      <td>1</td>\n",
       "      <td>0.434331</td>\n",
       "      <td>0.565669</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>Klay Thompson</td>\n",
       "      <td>2017</td>\n",
       "      <td>1</td>\n",
       "      <td>0.401390</td>\n",
       "      <td>0.598610</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>Donovan Mitchell</td>\n",
       "      <td>2019</td>\n",
       "      <td>1</td>\n",
       "      <td>0.385678</td>\n",
       "      <td>0.614322</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>Jaylen Brown</td>\n",
       "      <td>2020</td>\n",
       "      <td>1</td>\n",
       "      <td>0.348327</td>\n",
       "      <td>0.651673</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>Draymond Green</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>0.261088</td>\n",
       "      <td>0.738912</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>Mike Conley</td>\n",
       "      <td>2020</td>\n",
       "      <td>1</td>\n",
       "      <td>0.250130</td>\n",
       "      <td>0.749870</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>Paul Millsap</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>0.196576</td>\n",
       "      <td>0.803424</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>Andre Drummond</td>\n",
       "      <td>2017</td>\n",
       "      <td>1</td>\n",
       "      <td>0.132374</td>\n",
       "      <td>0.867626</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>Goran Dragić</td>\n",
       "      <td>2017</td>\n",
       "      <td>1</td>\n",
       "      <td>0.061873</td>\n",
       "      <td>0.938127</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     name  year  actual   prob_as  prob_non  predicted\n",
       "0       Russell Westbrook  2016       1  0.999659  0.000341          1\n",
       "1   Giannis Antetokounmpo  2018       1  0.998966  0.001034          1\n",
       "2            James Harden  2016       1  0.998927  0.001073          1\n",
       "3            LeBron James  2016       1  0.998849  0.001151          1\n",
       "4            Kevin Durant  2016       1  0.997083  0.002917          1\n",
       "5            Kevin Durant  2017       1  0.992692  0.007308          1\n",
       "6             Luka Dončić  2019       1  0.991790  0.008210          1\n",
       "7       Russell Westbrook  2020       0  0.991608  0.008392          1\n",
       "8             Paul George  2018       1  0.959714  0.040286          1\n",
       "9   Giannis Antetokounmpo  2017       1  0.947542  0.052458          1\n",
       "10          Pascal Siakam  2019       1  0.864835  0.135165          1\n",
       "11           Bradley Beal  2020       1  0.862850  0.137150          1\n",
       "12          Blake Griffin  2018       1  0.798908  0.201092          1\n",
       "13           Jimmy Butler  2020       0  0.775874  0.224126          1\n",
       "14             Trae Young  2019       1  0.729620  0.270380          1\n",
       "15          DeMar DeRozan  2018       0  0.724669  0.275331          1\n",
       "16          Blake Griffin  2016       0  0.704160  0.295840          1\n",
       "17  Giannis Antetokounmpo  2016       1  0.672419  0.327581          1\n",
       "18         Gordon Hayward  2016       1  0.564794  0.435206          1\n",
       "19          Klay Thompson  2018       1  0.434331  0.565669          0\n",
       "20          Klay Thompson  2017       1  0.401390  0.598610          0\n",
       "21       Donovan Mitchell  2019       1  0.385678  0.614322          0\n",
       "22           Jaylen Brown  2020       1  0.348327  0.651673          0\n",
       "23         Draymond Green  2016       1  0.261088  0.738912          0\n",
       "24            Mike Conley  2020       1  0.250130  0.749870          0\n",
       "25           Paul Millsap  2016       1  0.196576  0.803424          0\n",
       "26         Andre Drummond  2017       1  0.132374  0.867626          0\n",
       "27           Goran Dragić  2017       1  0.061873  0.938127          0"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results = test_data_df.merge(scaled_names)\n",
    "results.rename({'allstar':'actual'}, axis=1, inplace=True)\n",
    "results['predicted'] = predicted_labels\n",
    "results['prob_as'] = probabilities[:,1]\n",
    "results['prob_non'] = probabilities[:,0]\n",
    "results.sort_values(by='prob_as', ascending=False, inplace=True)\n",
    "results[(results['actual']==1) |  (results['predicted']==1)][['name','year','actual','prob_as','prob_non','predicted']].reset_index(drop=True)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "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>name</th>\n",
       "      <th>predicted</th>\n",
       "      <th>prob_as</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Kevin Durant</td>\n",
       "      <td>1</td>\n",
       "      <td>0.998693</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Giannis Antetokounmpo</td>\n",
       "      <td>1</td>\n",
       "      <td>0.997864</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Nikola Jokić</td>\n",
       "      <td>1</td>\n",
       "      <td>0.995115</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Stephen Curry</td>\n",
       "      <td>1</td>\n",
       "      <td>0.994930</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>LeBron James</td>\n",
       "      <td>1</td>\n",
       "      <td>0.988696</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Luka Dončić</td>\n",
       "      <td>1</td>\n",
       "      <td>0.988139</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Paul George</td>\n",
       "      <td>1</td>\n",
       "      <td>0.986489</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>James Harden</td>\n",
       "      <td>1</td>\n",
       "      <td>0.972350</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Russell Westbrook</td>\n",
       "      <td>1</td>\n",
       "      <td>0.960830</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Joel Embiid</td>\n",
       "      <td>1</td>\n",
       "      <td>0.950063</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>DeMar DeRozan</td>\n",
       "      <td>1</td>\n",
       "      <td>0.935156</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>Anthony Davis</td>\n",
       "      <td>1</td>\n",
       "      <td>0.927436</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>Trae Young</td>\n",
       "      <td>1</td>\n",
       "      <td>0.910825</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>Jimmy Butler</td>\n",
       "      <td>1</td>\n",
       "      <td>0.908397</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>Zach LaVine</td>\n",
       "      <td>1</td>\n",
       "      <td>0.895787</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>Ja Morant</td>\n",
       "      <td>1</td>\n",
       "      <td>0.873518</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>Jayson Tatum</td>\n",
       "      <td>1</td>\n",
       "      <td>0.871822</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>Bradley Beal</td>\n",
       "      <td>1</td>\n",
       "      <td>0.857157</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>Dejounte Murray</td>\n",
       "      <td>1</td>\n",
       "      <td>0.850524</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>Rudy Gobert</td>\n",
       "      <td>1</td>\n",
       "      <td>0.833554</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>Tobias Harris</td>\n",
       "      <td>1</td>\n",
       "      <td>0.827535</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>Damian Lillard</td>\n",
       "      <td>1</td>\n",
       "      <td>0.809427</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>Julius Randle</td>\n",
       "      <td>1</td>\n",
       "      <td>0.782368</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>Devin Booker</td>\n",
       "      <td>1</td>\n",
       "      <td>0.767043</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>Chris Paul</td>\n",
       "      <td>1</td>\n",
       "      <td>0.683011</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>Karl-Anthony Towns</td>\n",
       "      <td>1</td>\n",
       "      <td>0.584410</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>Malcolm Brogdon</td>\n",
       "      <td>1</td>\n",
       "      <td>0.583432</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     name  predicted   prob_as\n",
       "0            Kevin Durant          1  0.998693\n",
       "1   Giannis Antetokounmpo          1  0.997864\n",
       "2            Nikola Jokić          1  0.995115\n",
       "3           Stephen Curry          1  0.994930\n",
       "4            LeBron James          1  0.988696\n",
       "5             Luka Dončić          1  0.988139\n",
       "6             Paul George          1  0.986489\n",
       "7            James Harden          1  0.972350\n",
       "8       Russell Westbrook          1  0.960830\n",
       "9             Joel Embiid          1  0.950063\n",
       "10          DeMar DeRozan          1  0.935156\n",
       "11          Anthony Davis          1  0.927436\n",
       "12             Trae Young          1  0.910825\n",
       "13           Jimmy Butler          1  0.908397\n",
       "14            Zach LaVine          1  0.895787\n",
       "15              Ja Morant          1  0.873518\n",
       "16           Jayson Tatum          1  0.871822\n",
       "17           Bradley Beal          1  0.857157\n",
       "18        Dejounte Murray          1  0.850524\n",
       "19            Rudy Gobert          1  0.833554\n",
       "20          Tobias Harris          1  0.827535\n",
       "21         Damian Lillard          1  0.809427\n",
       "22          Julius Randle          1  0.782368\n",
       "23           Devin Booker          1  0.767043\n",
       "24             Chris Paul          1  0.683011\n",
       "25     Karl-Anthony Towns          1  0.584410\n",
       "26        Malcolm Brogdon          1  0.583432"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_2021 = scaled_names_2021.copy()\n",
    "pred_2021['predicted'] = predicted_2021\n",
    "pred_2021['prob_as'] = probabilities_2021[:,1]\n",
    "pred_2021['prob_non'] = probabilities_2021[:,0]\n",
    "pred_2021.sort_values(by='prob_as', ascending=False, inplace=True)\n",
    "pred_2021[pred_2021['predicted']==1][['name','predicted', 'prob_as']].reset_index(drop=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "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>feature</th>\n",
       "      <th>weight</th>\n",
       "      <th>weight_abs</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>win%</td>\n",
       "      <td>1.344794</td>\n",
       "      <td>1.344794</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>DRB</td>\n",
       "      <td>0.679081</td>\n",
       "      <td>0.679081</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>FG</td>\n",
       "      <td>0.644780</td>\n",
       "      <td>0.644780</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>PF</td>\n",
       "      <td>-0.584156</td>\n",
       "      <td>0.584156</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Age</td>\n",
       "      <td>0.515681</td>\n",
       "      <td>0.515681</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>TRB</td>\n",
       "      <td>0.490720</td>\n",
       "      <td>0.490720</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>MP</td>\n",
       "      <td>0.482114</td>\n",
       "      <td>0.482114</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>PTS</td>\n",
       "      <td>0.476540</td>\n",
       "      <td>0.476540</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>AST</td>\n",
       "      <td>0.429448</td>\n",
       "      <td>0.429448</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>ORB</td>\n",
       "      <td>-0.411253</td>\n",
       "      <td>0.411253</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>2P</td>\n",
       "      <td>0.410214</td>\n",
       "      <td>0.410214</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>FG%</td>\n",
       "      <td>0.401318</td>\n",
       "      <td>0.401318</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>eFG%</td>\n",
       "      <td>0.316518</td>\n",
       "      <td>0.316518</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>FTA</td>\n",
       "      <td>0.311535</td>\n",
       "      <td>0.311535</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>3P</td>\n",
       "      <td>0.290717</td>\n",
       "      <td>0.290717</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>GS%</td>\n",
       "      <td>0.289177</td>\n",
       "      <td>0.289177</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>TOV</td>\n",
       "      <td>0.277071</td>\n",
       "      <td>0.277071</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>FT%</td>\n",
       "      <td>-0.225040</td>\n",
       "      <td>0.225040</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>2PA</td>\n",
       "      <td>0.123319</td>\n",
       "      <td>0.123319</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>FGA</td>\n",
       "      <td>0.118327</td>\n",
       "      <td>0.118327</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>FT</td>\n",
       "      <td>0.104504</td>\n",
       "      <td>0.104504</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>3PA</td>\n",
       "      <td>0.102147</td>\n",
       "      <td>0.102147</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>BLK</td>\n",
       "      <td>0.097510</td>\n",
       "      <td>0.097510</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>STL</td>\n",
       "      <td>0.078914</td>\n",
       "      <td>0.078914</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>2P%</td>\n",
       "      <td>-0.061550</td>\n",
       "      <td>0.061550</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>3P%</td>\n",
       "      <td>0.040622</td>\n",
       "      <td>0.040622</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   feature    weight  weight_abs\n",
       "0     win%  1.344794    1.344794\n",
       "1      DRB  0.679081    0.679081\n",
       "2       FG  0.644780    0.644780\n",
       "3       PF -0.584156    0.584156\n",
       "4      Age  0.515681    0.515681\n",
       "5      TRB  0.490720    0.490720\n",
       "6       MP  0.482114    0.482114\n",
       "7      PTS  0.476540    0.476540\n",
       "8      AST  0.429448    0.429448\n",
       "9      ORB -0.411253    0.411253\n",
       "10      2P  0.410214    0.410214\n",
       "11     FG%  0.401318    0.401318\n",
       "12    eFG%  0.316518    0.316518\n",
       "13     FTA  0.311535    0.311535\n",
       "14      3P  0.290717    0.290717\n",
       "15     GS%  0.289177    0.289177\n",
       "16     TOV  0.277071    0.277071\n",
       "17     FT% -0.225040    0.225040\n",
       "18     2PA  0.123319    0.123319\n",
       "19     FGA  0.118327    0.118327\n",
       "20      FT  0.104504    0.104504\n",
       "21     3PA  0.102147    0.102147\n",
       "22     BLK  0.097510    0.097510\n",
       "23     STL  0.078914    0.078914\n",
       "24     2P% -0.061550    0.061550\n",
       "25     3P%  0.040622    0.040622"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "coefs_df = pd.DataFrame({'feature':data_cols,'weight':model.coef_[0]})\n",
    "coefs_df['weight_abs'] = coefs_df.weight.abs()\n",
    "coefs_df.sort_values(by='weight_abs', ascending=False, inplace=True)\n",
    "coefs_df.reset_index(inplace=True, drop=True)\n",
    "coefs_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}