Chess_winner_prediction.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Prediction of White player winning probability"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Prediction of White player winning probability using 10 classification model and comparing the performance of all these models and select the best one according the F1 score and AUC score "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "#import libraries\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "plt.rcParams.update({'figure.figsize':(12,8), 'figure.dpi':100})\n",
    "import seaborn as sns \n",
    "\n",
    "\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn.naive_bayes import GaussianNB\n",
    "from sklearn.svm import SVC\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.ensemble import AdaBoostClassifier\n",
    "from catboost import CatBoostClassifier\n",
    "from xgboost import XGBClassifier\n",
    "from lightgbm import LGBMClassifier\n",
    "\n",
    "\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "from sklearn.metrics import classification_report\n",
    "from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score, confusion_matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Load dataset\n",
    "df = pd.read_csv('games.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id</th>\n",
       "      <th>rated</th>\n",
       "      <th>created_at</th>\n",
       "      <th>last_move_at</th>\n",
       "      <th>turns</th>\n",
       "      <th>victory_status</th>\n",
       "      <th>winner</th>\n",
       "      <th>increment_code</th>\n",
       "      <th>white_id</th>\n",
       "      <th>white_rating</th>\n",
       "      <th>black_id</th>\n",
       "      <th>black_rating</th>\n",
       "      <th>moves</th>\n",
       "      <th>opening_eco</th>\n",
       "      <th>opening_name</th>\n",
       "      <th>opening_ply</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>TZJHLljE</td>\n",
       "      <td>False</td>\n",
       "      <td>1.504210e+12</td>\n",
       "      <td>1.504210e+12</td>\n",
       "      <td>13</td>\n",
       "      <td>outoftime</td>\n",
       "      <td>white</td>\n",
       "      <td>15+2</td>\n",
       "      <td>bourgris</td>\n",
       "      <td>1500</td>\n",
       "      <td>a-00</td>\n",
       "      <td>1191</td>\n",
       "      <td>d4 d5 c4 c6 cxd5 e6 dxe6 fxe6 Nf3 Bb4+ Nc3 Ba5...</td>\n",
       "      <td>D10</td>\n",
       "      <td>Slav Defense: Exchange Variation</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>l1NXvwaE</td>\n",
       "      <td>True</td>\n",
       "      <td>1.504130e+12</td>\n",
       "      <td>1.504130e+12</td>\n",
       "      <td>16</td>\n",
       "      <td>resign</td>\n",
       "      <td>black</td>\n",
       "      <td>5+10</td>\n",
       "      <td>a-00</td>\n",
       "      <td>1322</td>\n",
       "      <td>skinnerua</td>\n",
       "      <td>1261</td>\n",
       "      <td>d4 Nc6 e4 e5 f4 f6 dxe5 fxe5 fxe5 Nxe5 Qd4 Nc6...</td>\n",
       "      <td>B00</td>\n",
       "      <td>Nimzowitsch Defense: Kennedy Variation</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>mIICvQHh</td>\n",
       "      <td>True</td>\n",
       "      <td>1.504130e+12</td>\n",
       "      <td>1.504130e+12</td>\n",
       "      <td>61</td>\n",
       "      <td>mate</td>\n",
       "      <td>white</td>\n",
       "      <td>5+10</td>\n",
       "      <td>ischia</td>\n",
       "      <td>1496</td>\n",
       "      <td>a-00</td>\n",
       "      <td>1500</td>\n",
       "      <td>e4 e5 d3 d6 Be3 c6 Be2 b5 Nd2 a5 a4 c5 axb5 Nc...</td>\n",
       "      <td>C20</td>\n",
       "      <td>King's Pawn Game: Leonardis Variation</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>kWKvrqYL</td>\n",
       "      <td>True</td>\n",
       "      <td>1.504110e+12</td>\n",
       "      <td>1.504110e+12</td>\n",
       "      <td>61</td>\n",
       "      <td>mate</td>\n",
       "      <td>white</td>\n",
       "      <td>20+0</td>\n",
       "      <td>daniamurashov</td>\n",
       "      <td>1439</td>\n",
       "      <td>adivanov2009</td>\n",
       "      <td>1454</td>\n",
       "      <td>d4 d5 Nf3 Bf5 Nc3 Nf6 Bf4 Ng4 e3 Nc6 Be2 Qd7 O...</td>\n",
       "      <td>D02</td>\n",
       "      <td>Queen's Pawn Game: Zukertort Variation</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>9tXo1AUZ</td>\n",
       "      <td>True</td>\n",
       "      <td>1.504030e+12</td>\n",
       "      <td>1.504030e+12</td>\n",
       "      <td>95</td>\n",
       "      <td>mate</td>\n",
       "      <td>white</td>\n",
       "      <td>30+3</td>\n",
       "      <td>nik221107</td>\n",
       "      <td>1523</td>\n",
       "      <td>adivanov2009</td>\n",
       "      <td>1469</td>\n",
       "      <td>e4 e5 Nf3 d6 d4 Nc6 d5 Nb4 a3 Na6 Nc3 Be7 b4 N...</td>\n",
       "      <td>C41</td>\n",
       "      <td>Philidor Defense</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         id  rated    created_at  last_move_at  turns victory_status winner  \\\n",
       "0  TZJHLljE  False  1.504210e+12  1.504210e+12     13      outoftime  white   \n",
       "1  l1NXvwaE   True  1.504130e+12  1.504130e+12     16         resign  black   \n",
       "2  mIICvQHh   True  1.504130e+12  1.504130e+12     61           mate  white   \n",
       "3  kWKvrqYL   True  1.504110e+12  1.504110e+12     61           mate  white   \n",
       "4  9tXo1AUZ   True  1.504030e+12  1.504030e+12     95           mate  white   \n",
       "\n",
       "  increment_code       white_id  white_rating      black_id  black_rating  \\\n",
       "0           15+2       bourgris          1500          a-00          1191   \n",
       "1           5+10           a-00          1322     skinnerua          1261   \n",
       "2           5+10         ischia          1496          a-00          1500   \n",
       "3           20+0  daniamurashov          1439  adivanov2009          1454   \n",
       "4           30+3      nik221107          1523  adivanov2009          1469   \n",
       "\n",
       "                                               moves opening_eco  \\\n",
       "0  d4 d5 c4 c6 cxd5 e6 dxe6 fxe6 Nf3 Bb4+ Nc3 Ba5...         D10   \n",
       "1  d4 Nc6 e4 e5 f4 f6 dxe5 fxe5 fxe5 Nxe5 Qd4 Nc6...         B00   \n",
       "2  e4 e5 d3 d6 Be3 c6 Be2 b5 Nd2 a5 a4 c5 axb5 Nc...         C20   \n",
       "3  d4 d5 Nf3 Bf5 Nc3 Nf6 Bf4 Ng4 e3 Nc6 Be2 Qd7 O...         D02   \n",
       "4  e4 e5 Nf3 d6 d4 Nc6 d5 Nb4 a3 Na6 Nc3 Be7 b4 N...         C41   \n",
       "\n",
       "                             opening_name  opening_ply  \n",
       "0        Slav Defense: Exchange Variation            5  \n",
       "1  Nimzowitsch Defense: Kennedy Variation            4  \n",
       "2   King's Pawn Game: Leonardis Variation            3  \n",
       "3  Queen's Pawn Game: Zukertort Variation            3  \n",
       "4                        Philidor Defense            5  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#print head in order to have a quick view\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(20058, 16)"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#print the shape of data\n",
    "df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['id', 'rated', 'created_at', 'last_move_at', 'turns', 'victory_status',\n",
       "       'winner', 'increment_code', 'white_id', 'white_rating', 'black_id',\n",
       "       'black_rating', 'moves', 'opening_eco', 'opening_name', 'opening_ply'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Print the columns of dataset\n",
    "df.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "id                 object\n",
       "rated                bool\n",
       "created_at        float64\n",
       "last_move_at      float64\n",
       "turns               int64\n",
       "victory_status     object\n",
       "winner             object\n",
       "increment_code     object\n",
       "white_id           object\n",
       "white_rating        int64\n",
       "black_id           object\n",
       "black_rating        int64\n",
       "moves              object\n",
       "opening_eco        object\n",
       "opening_name       object\n",
       "opening_ply         int64\n",
       "dtype: object"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#lets see the datatypes of features\n",
    "df.dtypes"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As we can see there is many object types of data "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 20058 entries, 0 to 20057\n",
      "Data columns (total 16 columns):\n",
      " #   Column          Non-Null Count  Dtype  \n",
      "---  ------          --------------  -----  \n",
      " 0   id              20058 non-null  object \n",
      " 1   rated           20058 non-null  bool   \n",
      " 2   created_at      20058 non-null  float64\n",
      " 3   last_move_at    20058 non-null  float64\n",
      " 4   turns           20058 non-null  int64  \n",
      " 5   victory_status  20058 non-null  object \n",
      " 6   winner          20058 non-null  object \n",
      " 7   increment_code  20058 non-null  object \n",
      " 8   white_id        20058 non-null  object \n",
      " 9   white_rating    20058 non-null  int64  \n",
      " 10  black_id        20058 non-null  object \n",
      " 11  black_rating    20058 non-null  int64  \n",
      " 12  moves           20058 non-null  object \n",
      " 13  opening_eco     20058 non-null  object \n",
      " 14  opening_name    20058 non-null  object \n",
      " 15  opening_ply     20058 non-null  int64  \n",
      "dtypes: bool(1), float64(2), int64(4), object(9)\n",
      "memory usage: 2.3+ MB\n"
     ]
    }
   ],
   "source": [
    "df.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "id                0\n",
       "rated             0\n",
       "created_at        0\n",
       "last_move_at      0\n",
       "turns             0\n",
       "victory_status    0\n",
       "winner            0\n",
       "increment_code    0\n",
       "white_id          0\n",
       "white_rating      0\n",
       "black_id          0\n",
       "black_rating      0\n",
       "moves             0\n",
       "opening_eco       0\n",
       "opening_name      0\n",
       "opening_ply       0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#check the missing values\n",
    "df.isnull().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "So, there is no missing values "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>count</th>\n",
       "      <th>unique</th>\n",
       "      <th>top</th>\n",
       "      <th>freq</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>min</th>\n",
       "      <th>25%</th>\n",
       "      <th>50%</th>\n",
       "      <th>75%</th>\n",
       "      <th>max</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>id</th>\n",
       "      <td>20058</td>\n",
       "      <td>19113</td>\n",
       "      <td>XRuQPSzH</td>\n",
       "      <td>5</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>rated</th>\n",
       "      <td>20058</td>\n",
       "      <td>2</td>\n",
       "      <td>True</td>\n",
       "      <td>16155</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>created_at</th>\n",
       "      <td>20058.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1483616852629.092529</td>\n",
       "      <td>28501509421.004791</td>\n",
       "      <td>1376771633173.0</td>\n",
       "      <td>1477547500000.0</td>\n",
       "      <td>1496010000000.0</td>\n",
       "      <td>1503170000000.0</td>\n",
       "      <td>1504493143790.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>last_move_at</th>\n",
       "      <td>20058.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1483617722336.14209</td>\n",
       "      <td>28501400588.890018</td>\n",
       "      <td>1376771863841.0</td>\n",
       "      <td>1477547500000.0</td>\n",
       "      <td>1496010000000.0</td>\n",
       "      <td>1503170000000.0</td>\n",
       "      <td>1504493827262.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>turns</th>\n",
       "      <td>20058.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>60.465999</td>\n",
       "      <td>33.570585</td>\n",
       "      <td>1.0</td>\n",
       "      <td>37.0</td>\n",
       "      <td>55.0</td>\n",
       "      <td>79.0</td>\n",
       "      <td>349.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>victory_status</th>\n",
       "      <td>20058</td>\n",
       "      <td>4</td>\n",
       "      <td>resign</td>\n",
       "      <td>11147</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>winner</th>\n",
       "      <td>20058</td>\n",
       "      <td>3</td>\n",
       "      <td>white</td>\n",
       "      <td>10001</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>increment_code</th>\n",
       "      <td>20058</td>\n",
       "      <td>400</td>\n",
       "      <td>10+0</td>\n",
       "      <td>7721</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>white_id</th>\n",
       "      <td>20058</td>\n",
       "      <td>9438</td>\n",
       "      <td>taranga</td>\n",
       "      <td>72</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>white_rating</th>\n",
       "      <td>20058.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1596.631868</td>\n",
       "      <td>291.253376</td>\n",
       "      <td>784.0</td>\n",
       "      <td>1398.0</td>\n",
       "      <td>1567.0</td>\n",
       "      <td>1793.0</td>\n",
       "      <td>2700.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>black_id</th>\n",
       "      <td>20058</td>\n",
       "      <td>9331</td>\n",
       "      <td>taranga</td>\n",
       "      <td>82</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>black_rating</th>\n",
       "      <td>20058.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1588.831987</td>\n",
       "      <td>291.036126</td>\n",
       "      <td>789.0</td>\n",
       "      <td>1391.0</td>\n",
       "      <td>1562.0</td>\n",
       "      <td>1784.0</td>\n",
       "      <td>2723.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>moves</th>\n",
       "      <td>20058</td>\n",
       "      <td>18920</td>\n",
       "      <td>e4 e5</td>\n",
       "      <td>27</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>opening_eco</th>\n",
       "      <td>20058</td>\n",
       "      <td>365</td>\n",
       "      <td>A00</td>\n",
       "      <td>1007</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>opening_name</th>\n",
       "      <td>20058</td>\n",
       "      <td>1477</td>\n",
       "      <td>Van't Kruijs Opening</td>\n",
       "      <td>368</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>opening_ply</th>\n",
       "      <td>20058.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4.816981</td>\n",
       "      <td>2.797152</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>28.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  count unique                   top   freq  \\\n",
       "id                20058  19113              XRuQPSzH      5   \n",
       "rated             20058      2                  True  16155   \n",
       "created_at      20058.0    NaN                   NaN    NaN   \n",
       "last_move_at    20058.0    NaN                   NaN    NaN   \n",
       "turns           20058.0    NaN                   NaN    NaN   \n",
       "victory_status    20058      4                resign  11147   \n",
       "winner            20058      3                 white  10001   \n",
       "increment_code    20058    400                  10+0   7721   \n",
       "white_id          20058   9438               taranga     72   \n",
       "white_rating    20058.0    NaN                   NaN    NaN   \n",
       "black_id          20058   9331               taranga     82   \n",
       "black_rating    20058.0    NaN                   NaN    NaN   \n",
       "moves             20058  18920                 e4 e5     27   \n",
       "opening_eco       20058    365                   A00   1007   \n",
       "opening_name      20058   1477  Van't Kruijs Opening    368   \n",
       "opening_ply     20058.0    NaN                   NaN    NaN   \n",
       "\n",
       "                                mean                 std              min  \\\n",
       "id                               NaN                 NaN              NaN   \n",
       "rated                            NaN                 NaN              NaN   \n",
       "created_at      1483616852629.092529  28501509421.004791  1376771633173.0   \n",
       "last_move_at     1483617722336.14209  28501400588.890018  1376771863841.0   \n",
       "turns                      60.465999           33.570585              1.0   \n",
       "victory_status                   NaN                 NaN              NaN   \n",
       "winner                           NaN                 NaN              NaN   \n",
       "increment_code                   NaN                 NaN              NaN   \n",
       "white_id                         NaN                 NaN              NaN   \n",
       "white_rating             1596.631868          291.253376            784.0   \n",
       "black_id                         NaN                 NaN              NaN   \n",
       "black_rating             1588.831987          291.036126            789.0   \n",
       "moves                            NaN                 NaN              NaN   \n",
       "opening_eco                      NaN                 NaN              NaN   \n",
       "opening_name                     NaN                 NaN              NaN   \n",
       "opening_ply                 4.816981            2.797152              1.0   \n",
       "\n",
       "                            25%              50%              75%  \\\n",
       "id                          NaN              NaN              NaN   \n",
       "rated                       NaN              NaN              NaN   \n",
       "created_at      1477547500000.0  1496010000000.0  1503170000000.0   \n",
       "last_move_at    1477547500000.0  1496010000000.0  1503170000000.0   \n",
       "turns                      37.0             55.0             79.0   \n",
       "victory_status              NaN              NaN              NaN   \n",
       "winner                      NaN              NaN              NaN   \n",
       "increment_code              NaN              NaN              NaN   \n",
       "white_id                    NaN              NaN              NaN   \n",
       "white_rating             1398.0           1567.0           1793.0   \n",
       "black_id                    NaN              NaN              NaN   \n",
       "black_rating             1391.0           1562.0           1784.0   \n",
       "moves                       NaN              NaN              NaN   \n",
       "opening_eco                 NaN              NaN              NaN   \n",
       "opening_name                NaN              NaN              NaN   \n",
       "opening_ply                 3.0              4.0              6.0   \n",
       "\n",
       "                            max  \n",
       "id                          NaN  \n",
       "rated                       NaN  \n",
       "created_at      1504493143790.0  \n",
       "last_move_at    1504493827262.0  \n",
       "turns                     349.0  \n",
       "victory_status              NaN  \n",
       "winner                      NaN  \n",
       "increment_code              NaN  \n",
       "white_id                    NaN  \n",
       "white_rating             2700.0  \n",
       "black_id                    NaN  \n",
       "black_rating             2723.0  \n",
       "moves                       NaN  \n",
       "opening_eco                 NaN  \n",
       "opening_name                NaN  \n",
       "opening_ply                28.0  "
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.describe(include = 'all').T"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# I will clean all the columns one by one"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### first I will drop useless columns "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0             0.0\n",
       "1             0.0\n",
       "2             0.0\n",
       "3             0.0\n",
       "4             0.0\n",
       "           ...   \n",
       "20053   -321734.0\n",
       "20054   -744219.0\n",
       "20055   -172834.0\n",
       "20056   -946699.0\n",
       "20057   -736699.0\n",
       "Name: difference, Length: 20058, dtype: float64"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['difference'] = df['created_at']-df['last_move_at']\n",
    "df['difference']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "I think these two columns does not make sense because if the player started and 13 tuens(I see in turn column for this perticular id) then they should take some time to play but the start and last move time difference is zero, hence for this is not useful for our prediction I will drop it "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = df.drop(['created_at', 'last_move_at'],  axis = 1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "I think black_id, white_id and increment_code is also not usefull so, I will drop it too"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = df.drop(['increment_code', 'white_id', 'black_id', 'difference'], axis =1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id</th>\n",
       "      <th>rated</th>\n",
       "      <th>turns</th>\n",
       "      <th>victory_status</th>\n",
       "      <th>winner</th>\n",
       "      <th>white_rating</th>\n",
       "      <th>black_rating</th>\n",
       "      <th>moves</th>\n",
       "      <th>opening_eco</th>\n",
       "      <th>opening_name</th>\n",
       "      <th>opening_ply</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>TZJHLljE</td>\n",
       "      <td>False</td>\n",
       "      <td>13</td>\n",
       "      <td>outoftime</td>\n",
       "      <td>white</td>\n",
       "      <td>1500</td>\n",
       "      <td>1191</td>\n",
       "      <td>d4 d5 c4 c6 cxd5 e6 dxe6 fxe6 Nf3 Bb4+ Nc3 Ba5...</td>\n",
       "      <td>D10</td>\n",
       "      <td>Slav Defense: Exchange Variation</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>l1NXvwaE</td>\n",
       "      <td>True</td>\n",
       "      <td>16</td>\n",
       "      <td>resign</td>\n",
       "      <td>black</td>\n",
       "      <td>1322</td>\n",
       "      <td>1261</td>\n",
       "      <td>d4 Nc6 e4 e5 f4 f6 dxe5 fxe5 fxe5 Nxe5 Qd4 Nc6...</td>\n",
       "      <td>B00</td>\n",
       "      <td>Nimzowitsch Defense: Kennedy Variation</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>mIICvQHh</td>\n",
       "      <td>True</td>\n",
       "      <td>61</td>\n",
       "      <td>mate</td>\n",
       "      <td>white</td>\n",
       "      <td>1496</td>\n",
       "      <td>1500</td>\n",
       "      <td>e4 e5 d3 d6 Be3 c6 Be2 b5 Nd2 a5 a4 c5 axb5 Nc...</td>\n",
       "      <td>C20</td>\n",
       "      <td>King's Pawn Game: Leonardis Variation</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>kWKvrqYL</td>\n",
       "      <td>True</td>\n",
       "      <td>61</td>\n",
       "      <td>mate</td>\n",
       "      <td>white</td>\n",
       "      <td>1439</td>\n",
       "      <td>1454</td>\n",
       "      <td>d4 d5 Nf3 Bf5 Nc3 Nf6 Bf4 Ng4 e3 Nc6 Be2 Qd7 O...</td>\n",
       "      <td>D02</td>\n",
       "      <td>Queen's Pawn Game: Zukertort Variation</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>9tXo1AUZ</td>\n",
       "      <td>True</td>\n",
       "      <td>95</td>\n",
       "      <td>mate</td>\n",
       "      <td>white</td>\n",
       "      <td>1523</td>\n",
       "      <td>1469</td>\n",
       "      <td>e4 e5 Nf3 d6 d4 Nc6 d5 Nb4 a3 Na6 Nc3 Be7 b4 N...</td>\n",
       "      <td>C41</td>\n",
       "      <td>Philidor Defense</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         id  rated  turns victory_status winner  white_rating  black_rating  \\\n",
       "0  TZJHLljE  False     13      outoftime  white          1500          1191   \n",
       "1  l1NXvwaE   True     16         resign  black          1322          1261   \n",
       "2  mIICvQHh   True     61           mate  white          1496          1500   \n",
       "3  kWKvrqYL   True     61           mate  white          1439          1454   \n",
       "4  9tXo1AUZ   True     95           mate  white          1523          1469   \n",
       "\n",
       "                                               moves opening_eco  \\\n",
       "0  d4 d5 c4 c6 cxd5 e6 dxe6 fxe6 Nf3 Bb4+ Nc3 Ba5...         D10   \n",
       "1  d4 Nc6 e4 e5 f4 f6 dxe5 fxe5 fxe5 Nxe5 Qd4 Nc6...         B00   \n",
       "2  e4 e5 d3 d6 Be3 c6 Be2 b5 Nd2 a5 a4 c5 axb5 Nc...         C20   \n",
       "3  d4 d5 Nf3 Bf5 Nc3 Nf6 Bf4 Ng4 e3 Nc6 Be2 Qd7 O...         D02   \n",
       "4  e4 e5 Nf3 d6 d4 Nc6 d5 Nb4 a3 Na6 Nc3 Be7 b4 N...         C41   \n",
       "\n",
       "                             opening_name  opening_ply  \n",
       "0        Slav Defense: Exchange Variation            5  \n",
       "1  Nimzowitsch Defense: Kennedy Variation            4  \n",
       "2   King's Pawn Game: Leonardis Variation            3  \n",
       "3  Queen's Pawn Game: Zukertort Variation            3  \n",
       "4                        Philidor Defense            5  "
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Moves"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "I am going to predict the probability of winning on the basis of first move of the both black and white player. "
   ]
  },
  {
   "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>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>d4</td>\n",
       "      <td>d5</td>\n",
       "      <td>c4 c6 cxd5 e6 dxe6 fxe6 Nf3 Bb4+ Nc3 Ba5 Bf4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>d4</td>\n",
       "      <td>Nc6</td>\n",
       "      <td>e4 e5 f4 f6 dxe5 fxe5 fxe5 Nxe5 Qd4 Nc6 Qe5+ N...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>e4</td>\n",
       "      <td>e5</td>\n",
       "      <td>d3 d6 Be3 c6 Be2 b5 Nd2 a5 a4 c5 axb5 Nc6 bxc6...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>d4</td>\n",
       "      <td>d5</td>\n",
       "      <td>Nf3 Bf5 Nc3 Nf6 Bf4 Ng4 e3 Nc6 Be2 Qd7 O-O O-O...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>e4</td>\n",
       "      <td>e5</td>\n",
       "      <td>Nf3 d6 d4 Nc6 d5 Nb4 a3 Na6 Nc3 Be7 b4 Nf6 Bg5...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20053</th>\n",
       "      <td>d4</td>\n",
       "      <td>f5</td>\n",
       "      <td>e3 e6 Nf3 Nf6 Nc3 b6 Be2 Bb7 O-O Be7 Ne5 d6 Bh...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20054</th>\n",
       "      <td>d4</td>\n",
       "      <td>d6</td>\n",
       "      <td>Bf4 e5 Bg3 Nf6 e3 exd4 exd4 d5 c3 Bd6 Bd3 O-O ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20055</th>\n",
       "      <td>d4</td>\n",
       "      <td>d5</td>\n",
       "      <td>Bf4 Nc6 e3 Nf6 c3 e6 Nf3 Be7 Bd3 O-O Nbd2 b6 N...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20056</th>\n",
       "      <td>e4</td>\n",
       "      <td>d6</td>\n",
       "      <td>d4 Nf6 e5 dxe5 dxe5 Qxd1+ Kxd1 Nd5 c4 Nb6 c5 N...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20057</th>\n",
       "      <td>d4</td>\n",
       "      <td>d5</td>\n",
       "      <td>Bf4 Na6 e3 e6 c3 Nf6 Nf3 Bd7 Nbd2 b5 Bd3 Qc8 e...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>20058 rows × 3 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        0    1                                                  2\n",
       "0      d4   d5       c4 c6 cxd5 e6 dxe6 fxe6 Nf3 Bb4+ Nc3 Ba5 Bf4\n",
       "1      d4  Nc6  e4 e5 f4 f6 dxe5 fxe5 fxe5 Nxe5 Qd4 Nc6 Qe5+ N...\n",
       "2      e4   e5  d3 d6 Be3 c6 Be2 b5 Nd2 a5 a4 c5 axb5 Nc6 bxc6...\n",
       "3      d4   d5  Nf3 Bf5 Nc3 Nf6 Bf4 Ng4 e3 Nc6 Be2 Qd7 O-O O-O...\n",
       "4      e4   e5  Nf3 d6 d4 Nc6 d5 Nb4 a3 Na6 Nc3 Be7 b4 Nf6 Bg5...\n",
       "...    ..  ...                                                ...\n",
       "20053  d4   f5  e3 e6 Nf3 Nf6 Nc3 b6 Be2 Bb7 O-O Be7 Ne5 d6 Bh...\n",
       "20054  d4   d6  Bf4 e5 Bg3 Nf6 e3 exd4 exd4 d5 c3 Bd6 Bd3 O-O ...\n",
       "20055  d4   d5  Bf4 Nc6 e3 Nf6 c3 e6 Nf3 Be7 Bd3 O-O Nbd2 b6 N...\n",
       "20056  e4   d6  d4 Nf6 e5 dxe5 dxe5 Qxd1+ Kxd1 Nd5 c4 Nb6 c5 N...\n",
       "20057  d4   d5  Bf4 Na6 e3 e6 c3 Nf6 Nf3 Bd7 Nbd2 b5 Bd3 Qc8 e...\n",
       "\n",
       "[20058 rows x 3 columns]"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "moves = df.moves.str.split(n=2, expand=True)\n",
    "moves"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Splited the moves into three column, first column contains the first move of white player and second column contains the first move of black player and third column contains the rest of the moves which I am not going to consider into my model because it will take more time to do data manipulation I will do it later "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['white_1move'] = moves[0]\n",
    "df['black_1move']= moves[1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "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>winner</th>\n",
       "      <th>black</th>\n",
       "      <th>draw</th>\n",
       "      <th>white</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>white_1move</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Na3</th>\n",
       "      <td>0.750000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.250000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Nc3</th>\n",
       "      <td>0.555556</td>\n",
       "      <td>0.020202</td>\n",
       "      <td>0.424242</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Nf3</th>\n",
       "      <td>0.402759</td>\n",
       "      <td>0.082759</td>\n",
       "      <td>0.514483</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Nh3</th>\n",
       "      <td>0.400000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.600000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>a3</th>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.037037</td>\n",
       "      <td>0.629630</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "winner          black      draw     white\n",
       "white_1move                              \n",
       "Na3          0.750000  0.000000  0.250000\n",
       "Nc3          0.555556  0.020202  0.424242\n",
       "Nf3          0.402759  0.082759  0.514483\n",
       "Nh3          0.400000  0.000000  0.600000\n",
       "a3           0.333333  0.037037  0.629630"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_wm = pd.crosstab(df.white_1move, columns = df.winner, normalize = 'index')\n",
    "df_wm.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "dict(df_wm['white'])\n",
    "df['white_1move'] = df['white_1move'].map(dict(df_wm['white']))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "probability of winning of white player by using their first move"
   ]
  },
  {
   "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>winner</th>\n",
       "      <th>black</th>\n",
       "      <th>draw</th>\n",
       "      <th>white</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>black_1move</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Na6</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Nc6</th>\n",
       "      <td>0.382653</td>\n",
       "      <td>0.053571</td>\n",
       "      <td>0.563776</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Nf6</th>\n",
       "      <td>0.486119</td>\n",
       "      <td>0.062611</td>\n",
       "      <td>0.451270</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Nh6</th>\n",
       "      <td>0.166667</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.833333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>a5</th>\n",
       "      <td>0.260870</td>\n",
       "      <td>0.130435</td>\n",
       "      <td>0.608696</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "winner          black      draw     white\n",
       "black_1move                              \n",
       "Na6          0.000000  0.000000  1.000000\n",
       "Nc6          0.382653  0.053571  0.563776\n",
       "Nf6          0.486119  0.062611  0.451270\n",
       "Nh6          0.166667  0.000000  0.833333\n",
       "a5           0.260870  0.130435  0.608696"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_bm = pd.crosstab(df.black_1move, columns = df.winner, normalize = 'index')\n",
    "df_bm.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "dict(df_bm['white'])\n",
    "df['black_1move'] = df['black_1move'].map(dict(df_bm['white']))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "probability of winning of white player on the basis of first move of black player"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "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>winner</th>\n",
       "      <th>black</th>\n",
       "      <th>draw</th>\n",
       "      <th>white</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>opening_eco</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>A00</th>\n",
       "      <td>0.566038</td>\n",
       "      <td>0.038729</td>\n",
       "      <td>0.395233</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>A01</th>\n",
       "      <td>0.487805</td>\n",
       "      <td>0.048780</td>\n",
       "      <td>0.463415</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>A02</th>\n",
       "      <td>0.548780</td>\n",
       "      <td>0.060976</td>\n",
       "      <td>0.390244</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>A03</th>\n",
       "      <td>0.487179</td>\n",
       "      <td>0.128205</td>\n",
       "      <td>0.384615</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>A04</th>\n",
       "      <td>0.354386</td>\n",
       "      <td>0.098246</td>\n",
       "      <td>0.547368</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>E92</th>\n",
       "      <td>0.200000</td>\n",
       "      <td>0.600000</td>\n",
       "      <td>0.200000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>E94</th>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.500000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>E95</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>E97</th>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>E98</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>365 rows × 3 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "winner          black      draw     white\n",
       "opening_eco                              \n",
       "A00          0.566038  0.038729  0.395233\n",
       "A01          0.487805  0.048780  0.463415\n",
       "A02          0.548780  0.060976  0.390244\n",
       "A03          0.487179  0.128205  0.384615\n",
       "A04          0.354386  0.098246  0.547368\n",
       "...               ...       ...       ...\n",
       "E92          0.200000  0.600000  0.200000\n",
       "E94          0.500000  0.000000  0.500000\n",
       "E95          0.000000  0.000000  1.000000\n",
       "E97          0.333333  0.333333  0.333333\n",
       "E98          0.000000  0.000000  1.000000\n",
       "\n",
       "[365 rows x 3 columns]"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_ = pd.crosstab(df.opening_eco, columns = df.winner, normalize = 'index')\n",
    "df_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'A00': 0.3952333664349553,\n",
       " 'A01': 0.4634146341463415,\n",
       " 'A02': 0.3902439024390244,\n",
       " 'A03': 0.38461538461538464,\n",
       " 'A04': 0.5473684210526316,\n",
       " 'A05': 0.5625,\n",
       " 'A06': 0.472,\n",
       " 'A07': 0.5151515151515151,\n",
       " 'A08': 0.46153846153846156,\n",
       " 'A09': 0.5294117647058824,\n",
       " 'A10': 0.5354330708661418,\n",
       " 'A11': 0.5,\n",
       " 'A12': 0.0,\n",
       " 'A13': 0.7346938775510204,\n",
       " 'A14': 0.0,\n",
       " 'A15': 0.5333333333333333,\n",
       " 'A16': 0.48,\n",
       " 'A17': 0.5,\n",
       " 'A18': 0.6666666666666666,\n",
       " 'A20': 0.54,\n",
       " 'A21': 0.5396825396825397,\n",
       " 'A22': 0.4642857142857143,\n",
       " 'A23': 1.0,\n",
       " 'A24': 0.0,\n",
       " 'A25': 0.7272727272727273,\n",
       " 'A27': 0.5454545454545454,\n",
       " 'A28': 0.5,\n",
       " 'A29': 0.75,\n",
       " 'A30': 0.5833333333333334,\n",
       " 'A31': 0.0,\n",
       " 'A33': 0.0,\n",
       " 'A34': 0.5909090909090909,\n",
       " 'A35': 0.6923076923076923,\n",
       " 'A36': 0.75,\n",
       " 'A38': 0.5714285714285714,\n",
       " 'A39': 0.6666666666666666,\n",
       " 'A40': 0.5469255663430421,\n",
       " 'A41': 0.6,\n",
       " 'A42': 0.47619047619047616,\n",
       " 'A43': 0.47435897435897434,\n",
       " 'A44': 0.7142857142857143,\n",
       " 'A45': 0.40476190476190477,\n",
       " 'A46': 0.34177215189873417,\n",
       " 'A47': 0.25,\n",
       " 'A48': 0.375,\n",
       " 'A49': 0.5,\n",
       " 'A50': 0.5,\n",
       " 'A51': 0.3,\n",
       " 'A52': 0.20833333333333334,\n",
       " 'A53': 0.6111111111111112,\n",
       " 'A54': 0.6,\n",
       " 'A56': 0.25,\n",
       " 'A57': 0.4666666666666667,\n",
       " 'A58': 0.5,\n",
       " 'A59': 0.0,\n",
       " 'A60': 0.75,\n",
       " 'A65': 0.0,\n",
       " 'A67': 1.0,\n",
       " 'A70': 0.5714285714285714,\n",
       " 'A71': 0.0,\n",
       " 'A73': 0.0,\n",
       " 'A80': 0.5272727272727272,\n",
       " 'A81': 0.25,\n",
       " 'A82': 0.3333333333333333,\n",
       " 'A83': 0.5,\n",
       " 'A84': 0.4642857142857143,\n",
       " 'A85': 0.4827586206896552,\n",
       " 'A89': 0.0,\n",
       " 'A91': 0.5,\n",
       " 'A97': 1.0,\n",
       " 'A99': 1.0,\n",
       " 'B00': 0.5973813420621932,\n",
       " 'B01': 0.5,\n",
       " 'B02': 0.525,\n",
       " 'B03': 0.6086956521739131,\n",
       " 'B04': 0.5,\n",
       " 'B05': 0.6,\n",
       " 'B06': 0.4602272727272727,\n",
       " 'B07': 0.555921052631579,\n",
       " 'B08': 0.38636363636363635,\n",
       " 'B09': 0.5263157894736842,\n",
       " 'B10': 0.48186528497409326,\n",
       " 'B11': 0.5161290322580645,\n",
       " 'B12': 0.47904191616766467,\n",
       " 'B13': 0.4588235294117647,\n",
       " 'B14': 0.4,\n",
       " 'B15': 0.7419354838709677,\n",
       " 'B17': 1.0,\n",
       " 'B18': 0.40625,\n",
       " 'B19': 0.4444444444444444,\n",
       " 'B20': 0.3932980599647266,\n",
       " 'B21': 0.5063291139240507,\n",
       " 'B22': 0.44680851063829785,\n",
       " 'B23': 0.4336283185840708,\n",
       " 'B24': 0.9,\n",
       " 'B25': 0.4528301886792453,\n",
       " 'B27': 0.5230769230769231,\n",
       " 'B28': 0.3333333333333333,\n",
       " 'B29': 0.7368421052631579,\n",
       " 'B30': 0.42731277533039647,\n",
       " 'B31': 0.7857142857142857,\n",
       " 'B32': 0.5564516129032258,\n",
       " 'B33': 0.4791666666666667,\n",
       " 'B34': 0.5151515151515151,\n",
       " 'B35': 0.16666666666666666,\n",
       " 'B36': 0.3333333333333333,\n",
       " 'B38': 0.42857142857142855,\n",
       " 'B39': 0.0,\n",
       " 'B40': 0.3953488372093023,\n",
       " 'B41': 0.3076923076923077,\n",
       " 'B42': 0.625,\n",
       " 'B43': 0.3448275862068966,\n",
       " 'B44': 0.3888888888888889,\n",
       " 'B45': 0.631578947368421,\n",
       " 'B46': 0.21428571428571427,\n",
       " 'B47': 0.6666666666666666,\n",
       " 'B48': 0.2,\n",
       " 'B50': 0.43805309734513276,\n",
       " 'B51': 0.5714285714285714,\n",
       " 'B52': 0.45454545454545453,\n",
       " 'B53': 0.6666666666666666,\n",
       " 'B54': 0.5316455696202531,\n",
       " 'B56': 0.4375,\n",
       " 'B57': 0.35135135135135137,\n",
       " 'B58': 1.0,\n",
       " 'B60': 1.0,\n",
       " 'B62': 0.0,\n",
       " 'B70': 0.5,\n",
       " 'B71': 0.5,\n",
       " 'B72': 0.6,\n",
       " 'B73': 0.5,\n",
       " 'B74': 1.0,\n",
       " 'B75': 1.0,\n",
       " 'B76': 0.5,\n",
       " 'B77': 0.6666666666666666,\n",
       " 'B78': 0.42857142857142855,\n",
       " 'B80': 0.4666666666666667,\n",
       " 'B81': 1.0,\n",
       " 'B82': 0.5,\n",
       " 'B83': 1.0,\n",
       " 'B84': 0.5,\n",
       " 'B86': 0.2857142857142857,\n",
       " 'B87': 1.0,\n",
       " 'B88': 0.0,\n",
       " 'B90': 0.5346534653465347,\n",
       " 'B91': 0.5,\n",
       " 'B92': 0.5833333333333334,\n",
       " 'B93': 0.5,\n",
       " 'B94': 0.5789473684210527,\n",
       " 'B95': 0.4,\n",
       " 'B96': 0.4,\n",
       " 'B98': 0.5,\n",
       " 'C00': 0.4940758293838863,\n",
       " 'C01': 0.5253164556962026,\n",
       " 'C02': 0.47653429602888087,\n",
       " 'C03': 0.5,\n",
       " 'C04': 0.0,\n",
       " 'C05': 0.3076923076923077,\n",
       " 'C06': 0.8333333333333334,\n",
       " 'C07': 0.375,\n",
       " 'C08': 0.6666666666666666,\n",
       " 'C09': 1.0,\n",
       " 'C10': 0.543859649122807,\n",
       " 'C11': 0.4473684210526316,\n",
       " 'C12': 0.0,\n",
       " 'C13': 0.4,\n",
       " 'C14': 0.5,\n",
       " 'C15': 0.5882352941176471,\n",
       " 'C16': 0.4,\n",
       " 'C17': 0.5,\n",
       " 'C18': 0.6666666666666666,\n",
       " 'C20': 0.44296296296296295,\n",
       " 'C21': 0.5755813953488372,\n",
       " 'C22': 0.38372093023255816,\n",
       " 'C23': 0.5419354838709678,\n",
       " 'C24': 0.6160714285714286,\n",
       " 'C25': 0.4482758620689655,\n",
       " 'C26': 0.5294117647058824,\n",
       " 'C27': 0.5555555555555556,\n",
       " 'C28': 0.5263157894736842,\n",
       " 'C29': 0.5,\n",
       " 'C30': 0.532608695652174,\n",
       " 'C31': 0.4166666666666667,\n",
       " 'C32': 0.0,\n",
       " 'C33': 0.6153846153846154,\n",
       " 'C34': 0.6533333333333333,\n",
       " 'C35': 0.5,\n",
       " 'C36': 0.6,\n",
       " 'C37': 0.6363636363636364,\n",
       " 'C38': 0.6666666666666666,\n",
       " 'C39': 0.6666666666666666,\n",
       " 'C40': 0.5964125560538116,\n",
       " 'C41': 0.573082489146165,\n",
       " 'C42': 0.5474006116207951,\n",
       " 'C43': 0.2631578947368421,\n",
       " 'C44': 0.52,\n",
       " 'C45': 0.5035211267605634,\n",
       " 'C46': 0.5580524344569289,\n",
       " 'C47': 0.46296296296296297,\n",
       " 'C48': 0.5679012345679012,\n",
       " 'C49': 0.375,\n",
       " 'C50': 0.44609665427509293,\n",
       " 'C51': 0.5,\n",
       " 'C52': 0.5714285714285714,\n",
       " 'C53': 0.4880952380952381,\n",
       " 'C54': 0.2926829268292683,\n",
       " 'C55': 0.5288461538461539,\n",
       " 'C56': 0.3333333333333333,\n",
       " 'C57': 0.5537190082644629,\n",
       " 'C58': 0.4444444444444444,\n",
       " 'C59': 0.6666666666666666,\n",
       " 'C60': 0.5625,\n",
       " 'C61': 0.4444444444444444,\n",
       " 'C62': 0.49635036496350365,\n",
       " 'C63': 0.6,\n",
       " 'C64': 0.5614035087719298,\n",
       " 'C65': 0.5142857142857142,\n",
       " 'C66': 0.6666666666666666,\n",
       " 'C67': 0.5172413793103449,\n",
       " 'C68': 0.5588235294117647,\n",
       " 'C69': 0.5652173913043478,\n",
       " 'C70': 0.6666666666666666,\n",
       " 'C71': 0.5,\n",
       " 'C72': 0.5,\n",
       " 'C74': 0.0,\n",
       " 'C75': 0.0,\n",
       " 'C77': 0.3829787234042553,\n",
       " 'C78': 0.5757575757575758,\n",
       " 'C79': 1.0,\n",
       " 'C80': 0.7142857142857143,\n",
       " 'C81': 0.0,\n",
       " 'C83': 1.0,\n",
       " 'C84': 0.6,\n",
       " 'C88': 0.23529411764705882,\n",
       " 'C89': 0.14285714285714285,\n",
       " 'C90': 0.0,\n",
       " 'C91': 0.16666666666666666,\n",
       " 'C92': 0.0,\n",
       " 'C94': 0.0,\n",
       " 'D00': 0.4614343707713126,\n",
       " 'D01': 0.3146067415730337,\n",
       " 'D02': 0.511520737327189,\n",
       " 'D03': 0.4,\n",
       " 'D04': 0.3333333333333333,\n",
       " 'D05': 0.42857142857142855,\n",
       " 'D06': 0.6369047619047619,\n",
       " 'D07': 0.5306122448979592,\n",
       " 'D08': 0.6206896551724138,\n",
       " 'D10': 0.4,\n",
       " 'D11': 0.5121951219512195,\n",
       " 'D12': 0.4444444444444444,\n",
       " 'D13': 1.0,\n",
       " 'D14': 0.25,\n",
       " 'D15': 0.5531914893617021,\n",
       " 'D16': 1.0,\n",
       " 'D17': 0.5,\n",
       " 'D18': 0.5,\n",
       " 'D19': 1.0,\n",
       " 'D20': 0.5260663507109005,\n",
       " 'D21': 0.75,\n",
       " 'D22': 0.0,\n",
       " 'D24': 0.5,\n",
       " 'D25': 0.5,\n",
       " 'D26': 0.8,\n",
       " 'D30': 0.5862068965517241,\n",
       " 'D31': 0.5641025641025641,\n",
       " 'D32': 0.5833333333333334,\n",
       " 'D34': 0.25,\n",
       " 'D35': 0.4444444444444444,\n",
       " 'D36': 0.0,\n",
       " 'D37': 0.5,\n",
       " 'D38': 1.0,\n",
       " 'D40': 0.4,\n",
       " 'D41': 0.42857142857142855,\n",
       " 'D43': 0.36363636363636365,\n",
       " 'D44': 0.6666666666666666,\n",
       " 'D45': 0.48148148148148145,\n",
       " 'D46': 0.6,\n",
       " 'D47': 1.0,\n",
       " 'D48': 0.0,\n",
       " 'D50': 0.4,\n",
       " 'D51': 0.14285714285714285,\n",
       " 'D52': 0.5714285714285714,\n",
       " 'D53': 0.6875,\n",
       " 'D55': 0.5555555555555556,\n",
       " 'D56': 0.0,\n",
       " 'D57': 0.6666666666666666,\n",
       " 'D58': 0.0,\n",
       " 'D59': 0.0,\n",
       " 'D60': 0.5,\n",
       " 'D70': 0.5,\n",
       " 'D74': 1.0,\n",
       " 'D76': 1.0,\n",
       " 'D77': 0.0,\n",
       " 'D78': 0.6666666666666666,\n",
       " 'D80': 0.5555555555555556,\n",
       " 'D81': 1.0,\n",
       " 'D82': 1.0,\n",
       " 'D85': 0.6666666666666666,\n",
       " 'D86': 1.0,\n",
       " 'D87': 1.0,\n",
       " 'D90': 0.75,\n",
       " 'D91': 0.25,\n",
       " 'D92': 0.0,\n",
       " 'D93': 0.3333333333333333,\n",
       " 'D94': 0.375,\n",
       " 'D95': 1.0,\n",
       " 'E00': 0.48936170212765956,\n",
       " 'E01': 1.0,\n",
       " 'E03': 0.0,\n",
       " 'E06': 1.0,\n",
       " 'E10': 0.55,\n",
       " 'E11': 0.4,\n",
       " 'E12': 0.7333333333333333,\n",
       " 'E14': 0.2857142857142857,\n",
       " 'E15': 0.5714285714285714,\n",
       " 'E16': 0.3333333333333333,\n",
       " 'E17': 0.16666666666666666,\n",
       " 'E18': 1.0,\n",
       " 'E20': 0.45454545454545453,\n",
       " 'E21': 0.8571428571428571,\n",
       " 'E24': 0.4444444444444444,\n",
       " 'E27': 0.0,\n",
       " 'E30': 0.5,\n",
       " 'E32': 0.2,\n",
       " 'E33': 1.0,\n",
       " 'E34': 0.5,\n",
       " 'E35': 0.0,\n",
       " 'E38': 0.4,\n",
       " 'E40': 1.0,\n",
       " 'E42': 1.0,\n",
       " 'E44': 1.0,\n",
       " 'E46': 0.0,\n",
       " 'E47': 0.0,\n",
       " 'E48': 0.0,\n",
       " 'E50': 0.75,\n",
       " 'E51': 0.8,\n",
       " 'E52': 1.0,\n",
       " 'E53': 0.5,\n",
       " 'E54': 0.6666666666666666,\n",
       " 'E56': 0.0,\n",
       " 'E59': 0.3333333333333333,\n",
       " 'E60': 0.36363636363636365,\n",
       " 'E61': 0.423728813559322,\n",
       " 'E63': 1.0,\n",
       " 'E64': 0.5,\n",
       " 'E67': 0.4,\n",
       " 'E68': 0.5,\n",
       " 'E70': 0.6363636363636364,\n",
       " 'E71': 0.7142857142857143,\n",
       " 'E72': 0.0,\n",
       " 'E73': 0.4,\n",
       " 'E76': 0.3888888888888889,\n",
       " 'E77': 0.4,\n",
       " 'E80': 0.3333333333333333,\n",
       " 'E81': 0.5714285714285714,\n",
       " 'E87': 0.3333333333333333,\n",
       " 'E88': 1.0,\n",
       " 'E90': 0.4827586206896552,\n",
       " 'E91': 0.7142857142857143,\n",
       " 'E92': 0.2,\n",
       " 'E94': 0.5,\n",
       " 'E95': 1.0,\n",
       " 'E97': 0.3333333333333333,\n",
       " 'E98': 1.0}"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dict(df_['white'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['opening_eco']= df['opening_eco'].map(dict(df_['white']))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id</th>\n",
       "      <th>rated</th>\n",
       "      <th>turns</th>\n",
       "      <th>victory_status</th>\n",
       "      <th>winner</th>\n",
       "      <th>white_rating</th>\n",
       "      <th>black_rating</th>\n",
       "      <th>moves</th>\n",
       "      <th>opening_eco</th>\n",
       "      <th>opening_name</th>\n",
       "      <th>opening_ply</th>\n",
       "      <th>white_1move</th>\n",
       "      <th>black_1move</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>TZJHLljE</td>\n",
       "      <td>False</td>\n",
       "      <td>13</td>\n",
       "      <td>outoftime</td>\n",
       "      <td>white</td>\n",
       "      <td>1500</td>\n",
       "      <td>1191</td>\n",
       "      <td>d4 d5 c4 c6 cxd5 e6 dxe6 fxe6 Nf3 Bb4+ Nc3 Ba5...</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>Slav Defense: Exchange Variation</td>\n",
       "      <td>5</td>\n",
       "      <td>0.499337</td>\n",
       "      <td>0.493591</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>l1NXvwaE</td>\n",
       "      <td>True</td>\n",
       "      <td>16</td>\n",
       "      <td>resign</td>\n",
       "      <td>black</td>\n",
       "      <td>1322</td>\n",
       "      <td>1261</td>\n",
       "      <td>d4 Nc6 e4 e5 f4 f6 dxe5 fxe5 fxe5 Nxe5 Qd4 Nc6...</td>\n",
       "      <td>0.597381</td>\n",
       "      <td>Nimzowitsch Defense: Kennedy Variation</td>\n",
       "      <td>4</td>\n",
       "      <td>0.499337</td>\n",
       "      <td>0.563776</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>mIICvQHh</td>\n",
       "      <td>True</td>\n",
       "      <td>61</td>\n",
       "      <td>mate</td>\n",
       "      <td>white</td>\n",
       "      <td>1496</td>\n",
       "      <td>1500</td>\n",
       "      <td>e4 e5 d3 d6 Be3 c6 Be2 b5 Nd2 a5 a4 c5 axb5 Nc...</td>\n",
       "      <td>0.442963</td>\n",
       "      <td>King's Pawn Game: Leonardis Variation</td>\n",
       "      <td>3</td>\n",
       "      <td>0.505715</td>\n",
       "      <td>0.515312</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>kWKvrqYL</td>\n",
       "      <td>True</td>\n",
       "      <td>61</td>\n",
       "      <td>mate</td>\n",
       "      <td>white</td>\n",
       "      <td>1439</td>\n",
       "      <td>1454</td>\n",
       "      <td>d4 d5 Nf3 Bf5 Nc3 Nf6 Bf4 Ng4 e3 Nc6 Be2 Qd7 O...</td>\n",
       "      <td>0.511521</td>\n",
       "      <td>Queen's Pawn Game: Zukertort Variation</td>\n",
       "      <td>3</td>\n",
       "      <td>0.499337</td>\n",
       "      <td>0.493591</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>9tXo1AUZ</td>\n",
       "      <td>True</td>\n",
       "      <td>95</td>\n",
       "      <td>mate</td>\n",
       "      <td>white</td>\n",
       "      <td>1523</td>\n",
       "      <td>1469</td>\n",
       "      <td>e4 e5 Nf3 d6 d4 Nc6 d5 Nb4 a3 Na6 Nc3 Be7 b4 N...</td>\n",
       "      <td>0.573082</td>\n",
       "      <td>Philidor Defense</td>\n",
       "      <td>5</td>\n",
       "      <td>0.505715</td>\n",
       "      <td>0.515312</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         id  rated  turns victory_status winner  white_rating  black_rating  \\\n",
       "0  TZJHLljE  False     13      outoftime  white          1500          1191   \n",
       "1  l1NXvwaE   True     16         resign  black          1322          1261   \n",
       "2  mIICvQHh   True     61           mate  white          1496          1500   \n",
       "3  kWKvrqYL   True     61           mate  white          1439          1454   \n",
       "4  9tXo1AUZ   True     95           mate  white          1523          1469   \n",
       "\n",
       "                                               moves  opening_eco  \\\n",
       "0  d4 d5 c4 c6 cxd5 e6 dxe6 fxe6 Nf3 Bb4+ Nc3 Ba5...     0.400000   \n",
       "1  d4 Nc6 e4 e5 f4 f6 dxe5 fxe5 fxe5 Nxe5 Qd4 Nc6...     0.597381   \n",
       "2  e4 e5 d3 d6 Be3 c6 Be2 b5 Nd2 a5 a4 c5 axb5 Nc...     0.442963   \n",
       "3  d4 d5 Nf3 Bf5 Nc3 Nf6 Bf4 Ng4 e3 Nc6 Be2 Qd7 O...     0.511521   \n",
       "4  e4 e5 Nf3 d6 d4 Nc6 d5 Nb4 a3 Na6 Nc3 Be7 b4 N...     0.573082   \n",
       "\n",
       "                             opening_name  opening_ply  white_1move  \\\n",
       "0        Slav Defense: Exchange Variation            5     0.499337   \n",
       "1  Nimzowitsch Defense: Kennedy Variation            4     0.499337   \n",
       "2   King's Pawn Game: Leonardis Variation            3     0.505715   \n",
       "3  Queen's Pawn Game: Zukertort Variation            3     0.499337   \n",
       "4                        Philidor Defense            5     0.505715   \n",
       "\n",
       "   black_1move  \n",
       "0     0.493591  \n",
       "1     0.563776  \n",
       "2     0.515312  \n",
       "3     0.493591  \n",
       "4     0.515312  "
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "20"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.white_1move.nunique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "20"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.black_1move.nunique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "id                 0\n",
       "rated              0\n",
       "turns              0\n",
       "victory_status     0\n",
       "winner             0\n",
       "white_rating       0\n",
       "black_rating       0\n",
       "moves              0\n",
       "opening_eco        0\n",
       "opening_name       0\n",
       "opening_ply        0\n",
       "white_1move        0\n",
       "black_1move       18\n",
       "dtype: int64"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.isnull().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "There is 18 missing value acceptable because it is possible that they started the game white payer is the first who start the game maybe after start of the game black player gaveup to play"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id</th>\n",
       "      <th>rated</th>\n",
       "      <th>turns</th>\n",
       "      <th>victory_status</th>\n",
       "      <th>winner</th>\n",
       "      <th>white_rating</th>\n",
       "      <th>black_rating</th>\n",
       "      <th>moves</th>\n",
       "      <th>opening_eco</th>\n",
       "      <th>opening_name</th>\n",
       "      <th>opening_ply</th>\n",
       "      <th>white_1move</th>\n",
       "      <th>black_1move</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>TZJHLljE</td>\n",
       "      <td>False</td>\n",
       "      <td>13</td>\n",
       "      <td>outoftime</td>\n",
       "      <td>white</td>\n",
       "      <td>1500</td>\n",
       "      <td>1191</td>\n",
       "      <td>d4 d5 c4 c6 cxd5 e6 dxe6 fxe6 Nf3 Bb4+ Nc3 Ba5...</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>Slav Defense: Exchange Variation</td>\n",
       "      <td>5</td>\n",
       "      <td>0.499337</td>\n",
       "      <td>0.493591</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>l1NXvwaE</td>\n",
       "      <td>True</td>\n",
       "      <td>16</td>\n",
       "      <td>resign</td>\n",
       "      <td>black</td>\n",
       "      <td>1322</td>\n",
       "      <td>1261</td>\n",
       "      <td>d4 Nc6 e4 e5 f4 f6 dxe5 fxe5 fxe5 Nxe5 Qd4 Nc6...</td>\n",
       "      <td>0.597381</td>\n",
       "      <td>Nimzowitsch Defense: Kennedy Variation</td>\n",
       "      <td>4</td>\n",
       "      <td>0.499337</td>\n",
       "      <td>0.563776</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>mIICvQHh</td>\n",
       "      <td>True</td>\n",
       "      <td>61</td>\n",
       "      <td>mate</td>\n",
       "      <td>white</td>\n",
       "      <td>1496</td>\n",
       "      <td>1500</td>\n",
       "      <td>e4 e5 d3 d6 Be3 c6 Be2 b5 Nd2 a5 a4 c5 axb5 Nc...</td>\n",
       "      <td>0.442963</td>\n",
       "      <td>King's Pawn Game: Leonardis Variation</td>\n",
       "      <td>3</td>\n",
       "      <td>0.505715</td>\n",
       "      <td>0.515312</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>kWKvrqYL</td>\n",
       "      <td>True</td>\n",
       "      <td>61</td>\n",
       "      <td>mate</td>\n",
       "      <td>white</td>\n",
       "      <td>1439</td>\n",
       "      <td>1454</td>\n",
       "      <td>d4 d5 Nf3 Bf5 Nc3 Nf6 Bf4 Ng4 e3 Nc6 Be2 Qd7 O...</td>\n",
       "      <td>0.511521</td>\n",
       "      <td>Queen's Pawn Game: Zukertort Variation</td>\n",
       "      <td>3</td>\n",
       "      <td>0.499337</td>\n",
       "      <td>0.493591</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>9tXo1AUZ</td>\n",
       "      <td>True</td>\n",
       "      <td>95</td>\n",
       "      <td>mate</td>\n",
       "      <td>white</td>\n",
       "      <td>1523</td>\n",
       "      <td>1469</td>\n",
       "      <td>e4 e5 Nf3 d6 d4 Nc6 d5 Nb4 a3 Na6 Nc3 Be7 b4 N...</td>\n",
       "      <td>0.573082</td>\n",
       "      <td>Philidor Defense</td>\n",
       "      <td>5</td>\n",
       "      <td>0.505715</td>\n",
       "      <td>0.515312</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         id  rated  turns victory_status winner  white_rating  black_rating  \\\n",
       "0  TZJHLljE  False     13      outoftime  white          1500          1191   \n",
       "1  l1NXvwaE   True     16         resign  black          1322          1261   \n",
       "2  mIICvQHh   True     61           mate  white          1496          1500   \n",
       "3  kWKvrqYL   True     61           mate  white          1439          1454   \n",
       "4  9tXo1AUZ   True     95           mate  white          1523          1469   \n",
       "\n",
       "                                               moves  opening_eco  \\\n",
       "0  d4 d5 c4 c6 cxd5 e6 dxe6 fxe6 Nf3 Bb4+ Nc3 Ba5...     0.400000   \n",
       "1  d4 Nc6 e4 e5 f4 f6 dxe5 fxe5 fxe5 Nxe5 Qd4 Nc6...     0.597381   \n",
       "2  e4 e5 d3 d6 Be3 c6 Be2 b5 Nd2 a5 a4 c5 axb5 Nc...     0.442963   \n",
       "3  d4 d5 Nf3 Bf5 Nc3 Nf6 Bf4 Ng4 e3 Nc6 Be2 Qd7 O...     0.511521   \n",
       "4  e4 e5 Nf3 d6 d4 Nc6 d5 Nb4 a3 Na6 Nc3 Be7 b4 N...     0.573082   \n",
       "\n",
       "                             opening_name  opening_ply  white_1move  \\\n",
       "0        Slav Defense: Exchange Variation            5     0.499337   \n",
       "1  Nimzowitsch Defense: Kennedy Variation            4     0.499337   \n",
       "2   King's Pawn Game: Leonardis Variation            3     0.505715   \n",
       "3  Queen's Pawn Game: Zukertort Variation            3     0.499337   \n",
       "4                        Philidor Defense            5     0.505715   \n",
       "\n",
       "   black_1move  \n",
       "0     0.493591  \n",
       "1     0.563776  \n",
       "2     0.515312  \n",
       "3     0.493591  \n",
       "4     0.515312  "
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.dropna().head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "#df.moves.unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "23"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.opening_ply.nunique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1477"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.opening_name.nunique()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Now I will treat the categorical varibles"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### first lets check what is rated columns "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([False,  True])"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['rated'].unique()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "so, it include two category true and false which seems if rated is true and if not rated is false which make sense, lets convert it into o and 1 form"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['outoftime', 'resign', 'mate', 'draw'], dtype=object)"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.victory_status.unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.drop(columns = ['opening_name', 'id','victory_status', 'turns','moves'], axis =1, inplace =True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "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>rated</th>\n",
       "      <th>winner</th>\n",
       "      <th>white_rating</th>\n",
       "      <th>black_rating</th>\n",
       "      <th>opening_eco</th>\n",
       "      <th>opening_ply</th>\n",
       "      <th>white_1move</th>\n",
       "      <th>black_1move</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>False</td>\n",
       "      <td>white</td>\n",
       "      <td>1500</td>\n",
       "      <td>1191</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>5</td>\n",
       "      <td>0.499337</td>\n",
       "      <td>0.493591</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>True</td>\n",
       "      <td>black</td>\n",
       "      <td>1322</td>\n",
       "      <td>1261</td>\n",
       "      <td>0.597381</td>\n",
       "      <td>4</td>\n",
       "      <td>0.499337</td>\n",
       "      <td>0.563776</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>True</td>\n",
       "      <td>white</td>\n",
       "      <td>1496</td>\n",
       "      <td>1500</td>\n",
       "      <td>0.442963</td>\n",
       "      <td>3</td>\n",
       "      <td>0.505715</td>\n",
       "      <td>0.515312</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>True</td>\n",
       "      <td>white</td>\n",
       "      <td>1439</td>\n",
       "      <td>1454</td>\n",
       "      <td>0.511521</td>\n",
       "      <td>3</td>\n",
       "      <td>0.499337</td>\n",
       "      <td>0.493591</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>True</td>\n",
       "      <td>white</td>\n",
       "      <td>1523</td>\n",
       "      <td>1469</td>\n",
       "      <td>0.573082</td>\n",
       "      <td>5</td>\n",
       "      <td>0.505715</td>\n",
       "      <td>0.515312</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   rated winner  white_rating  black_rating  opening_eco  opening_ply  \\\n",
       "0  False  white          1500          1191     0.400000            5   \n",
       "1   True  black          1322          1261     0.597381            4   \n",
       "2   True  white          1496          1500     0.442963            3   \n",
       "3   True  white          1439          1454     0.511521            3   \n",
       "4   True  white          1523          1469     0.573082            5   \n",
       "\n",
       "   white_1move  black_1move  \n",
       "0     0.499337     0.493591  \n",
       "1     0.499337     0.563776  \n",
       "2     0.505715     0.515312  \n",
       "3     0.499337     0.493591  \n",
       "4     0.505715     0.515312  "
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.replace({False: 0, True: 1}, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "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>rated</th>\n",
       "      <th>winner</th>\n",
       "      <th>white_rating</th>\n",
       "      <th>black_rating</th>\n",
       "      <th>opening_eco</th>\n",
       "      <th>opening_ply</th>\n",
       "      <th>white_1move</th>\n",
       "      <th>black_1move</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>white</td>\n",
       "      <td>1500</td>\n",
       "      <td>1191</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>5</td>\n",
       "      <td>0.499337</td>\n",
       "      <td>0.493591</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>black</td>\n",
       "      <td>1322</td>\n",
       "      <td>1261</td>\n",
       "      <td>0.597381</td>\n",
       "      <td>4</td>\n",
       "      <td>0.499337</td>\n",
       "      <td>0.563776</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>white</td>\n",
       "      <td>1496</td>\n",
       "      <td>1500</td>\n",
       "      <td>0.442963</td>\n",
       "      <td>3</td>\n",
       "      <td>0.505715</td>\n",
       "      <td>0.515312</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>white</td>\n",
       "      <td>1439</td>\n",
       "      <td>1454</td>\n",
       "      <td>0.511521</td>\n",
       "      <td>3</td>\n",
       "      <td>0.499337</td>\n",
       "      <td>0.493591</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>white</td>\n",
       "      <td>1523</td>\n",
       "      <td>1469</td>\n",
       "      <td>0.573082</td>\n",
       "      <td>5</td>\n",
       "      <td>0.505715</td>\n",
       "      <td>0.515312</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   rated winner  white_rating  black_rating  opening_eco  opening_ply  \\\n",
       "0      0  white          1500          1191     0.400000            5   \n",
       "1      1  black          1322          1261     0.597381            4   \n",
       "2      1  white          1496          1500     0.442963            3   \n",
       "3      1  white          1439          1454     0.511521            3   \n",
       "4      1  white          1523          1469     0.573082            5   \n",
       "\n",
       "   white_1move  black_1move  \n",
       "0     0.499337     0.493591  \n",
       "1     0.499337     0.563776  \n",
       "2     0.505715     0.515312  \n",
       "3     0.499337     0.493591  \n",
       "4     0.505715     0.515312  "
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.replace({'white' : 1, 'black': 0, 'draw': 0}, inplace = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "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>rated</th>\n",
       "      <th>winner</th>\n",
       "      <th>white_rating</th>\n",
       "      <th>black_rating</th>\n",
       "      <th>opening_eco</th>\n",
       "      <th>opening_ply</th>\n",
       "      <th>white_1move</th>\n",
       "      <th>black_1move</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1500</td>\n",
       "      <td>1191</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>5</td>\n",
       "      <td>0.499337</td>\n",
       "      <td>0.493591</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1322</td>\n",
       "      <td>1261</td>\n",
       "      <td>0.597381</td>\n",
       "      <td>4</td>\n",
       "      <td>0.499337</td>\n",
       "      <td>0.563776</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1496</td>\n",
       "      <td>1500</td>\n",
       "      <td>0.442963</td>\n",
       "      <td>3</td>\n",
       "      <td>0.505715</td>\n",
       "      <td>0.515312</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1439</td>\n",
       "      <td>1454</td>\n",
       "      <td>0.511521</td>\n",
       "      <td>3</td>\n",
       "      <td>0.499337</td>\n",
       "      <td>0.493591</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1523</td>\n",
       "      <td>1469</td>\n",
       "      <td>0.573082</td>\n",
       "      <td>5</td>\n",
       "      <td>0.505715</td>\n",
       "      <td>0.515312</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   rated  winner  white_rating  black_rating  opening_eco  opening_ply  \\\n",
       "0      0       1          1500          1191     0.400000            5   \n",
       "1      1       0          1322          1261     0.597381            4   \n",
       "2      1       1          1496          1500     0.442963            3   \n",
       "3      1       1          1439          1454     0.511521            3   \n",
       "4      1       1          1523          1469     0.573082            5   \n",
       "\n",
       "   white_1move  black_1move  \n",
       "0     0.499337     0.493591  \n",
       "1     0.499337     0.563776  \n",
       "2     0.505715     0.515312  \n",
       "3     0.499337     0.493591  \n",
       "4     0.505715     0.515312  "
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7fc9adb3a2d0>"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1200x800 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df.winner.value_counts().plot(kind = 'bar')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7fc9adc64690>"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA+IAAAKyCAYAAABR42QIAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAAPYQAAD2EBqD+naQAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+j8jraAAAgAElEQVR4nOzdf7RndX3f+9dbB0arDIR7F0O8SEPblCohxRCVMUatOGpEU+2qmotZEaNJ8SeiCZUmcgl1LRS7kFan2igmaEnt6sXqSogZ7kSNSe5IK8ZoFHNriyjFGYwExh8wKH7uH9996peTM8x8Z4b3OQOPx1p7nfPd+/Pd57MXfpc82fu7d40xAgAAAPR40GpPAAAAAB5IhDgAAAA0EuIAAADQSIgDAABAIyEOAAAAjYQ4AAAANBLiAAAA0Gjdak/gvlJVleQRSb652nMBAADgAeOIJDePMcaeBtxvQzyzCL9ptScBAADAA85xSf7nnjben0P8m0ny1a9+NRs2bFjtuQAAAHA/t2vXrjzykY9M9nJl9v05xJMkGzZsEOIAAACsGW7WBgAAAI2EOAAAADQS4gAAANBIiAMAAEAjIQ4AAACNhDgAAAA0EuIAAADQSIgDAABAIyEOAAAAjYQ4AAAANBLiAAAA0EiIAwAAQCMhDgAAAI2EOAAAADQS4gAAANBIiAMAAEAjIQ4AAACNhDgAAAA0EuIAAADQSIgDAABAIyEOAAAAjYQ4AAAANBLiAAAA0EiIAwAAQCMhDgAAAI2EOAAAADRat9oT4MD9yBuuXu0pcBB9+c1nrPYUAACA+5Az4gAAANBIiAMAAEAjIQ4AAACNFvqOeFU9OMmFSX4+ybFJbk7y20neNMYY05hK8htJfinJUUn+NMnLxxj/bW4/Ryd5e5LnJPl+kquSnDPG+NbcmB9PsiXJY5N8PcnbxxiX7M9BAgCsxH1W7l/cZwU4VCx6RvyfJ3l5klcledT0+rwkr54bc16S1yQ5O8njk3w7ydaqesjcmCuTnJRkc5JnJ3lSkt9c2lhVG5Jck+TGJKcm+dUkF1bVLy84XwAAAFhTFr1r+hOSfHiMsfSfj79cVf9nkscl/+ts+GszO0P+4WndLyTZmeS5ST5QVY9K8swkjx1jfGoa8+okv19VvzLGuDnJi5IcnuQXxxh3Jfl8VZ2S5HWZC/Z5VbU+yfq5VUcseGwAAABwn1v0jPj/m+T0qvr7SVJV/zDJE5N8ZNp+QmaXrG9besMY4/Yk1ybZNK3alOS2pQifbMvsEvXHz435xBThS7YmObGqfmgPczs/ye1zy00LHhsAAADc5xYN8Tcn+UCSL1bVd5P8WZLLxhhXTtuPnX7uXPa+nXPbjk1yy/zGMcb3kty6bMxK+5j/G8tdnOTIueW4fTgeAAAAaLXopekvyOyy8TOTfD7JKUkuq6qbxxhXHOzJLWKMsTvJ7qXXs6vkAQAAYG1ZNMTfmuTNY4wPTK8/V1V/O7PLwq9IsmNavzHJ1+betzHJZ6bfdyQ5Zn6nVbUuydFz798xvWfexrltAAAAcEha9NL0v5XZd7nn3T23nxsyC+XTlzZOd0B/fJLt06rtSY6qqlPn9vHUaR/Xzo15UlUdNjdmc5K/HGP89YJzBgAAgDVj0RD/3SS/VlVnVNWPVNXzMruT+X9OkulZ4pcl+fWq+tmqOjnJ+zJ73viHpjHXJ/mDJO+uqsdV1U8leUeSD0x3TE+S30lyV5LLq+qkqnphknOSXHogBwsAAACrbdFL01+d5F8m+beZXV5+c5J/l+SiuTGXJHlYZo8ZOyrJnyR55hjjzrkxL8osvv8wszPsV2X27PEkszutV9XTk2xJcl2Sv0py0RhjxUeXAQAAwKFioRAfY3wzs+eEv/ZexowkF0zLnsbcmtkN3+7tb302yU8vMj8AAABY6xa9NB0AAAA4AEIcAAAAGglxAAAAaCTEAQAAoJEQBwAAgEZCHAAAABoJcQAAAGgkxAEAAKCREAcAAIBGQhwAAAAaCXEAAABoJMQBAACgkRAHAACARkIcAAAAGglxAAAAaCTEAQAAoJEQBwAAgEZCHAAAABoJcQAAAGgkxAEAAKCREAcAAIBGQhwAAAAaCXEAAABoJMQBAACgkRAHAACARkIcAAAAGglxAAAAaCTEAQAAoJEQBwAAgEZCHAAAABoJcQAAAGgkxAEAAKCREAcAAIBGQhwAAAAaCXEAAABoJMQBAACgkRAHAACARkIcAAAAGglxAAAAaCTEAQAAoJEQBwAAgEZCHAAAABoJcQAAAGgkxAEAAKCREAcAAIBGQhwAAAAaCXEAAABoJMQBAACgkRAHAACARkIcAAAAGglxAAAAaCTEAQAAoJEQBwAAgEZCHAAAABoJcQAAAGgkxAEAAKCREAcAAIBGQhwAAAAaCXEAAABoJMQBAACg0UIhXlVfrqqxwrJl2v6QqtpSVd+oqm9V1VVVtXHZPo6vqqur6jtVdUtVvbWq1i0b85Sq+nRV7a6qL1XVWQd8pAAAALAGLHpG/LFJfnhu2Tyt/0/Tz7cleU6S5yd5cpJHJPng0pur6sFJrk5yeJInJHlxkrOSXDQ35oRpzMeSnJLksiTvqapnLDhXAAAAWHPW7X3ID4wxvj7/uqrekOS/J/mjqjoyyUuTnDnG+Oi0/SVJrq+q08YYn0zy9CSPTvK0McbOJJ+pqjcmeUtVXTjGuCvJ2UluGGO8fvoz11fVE5Ocm2TrnuZWVeuTrJ9bdcQixwYAAAAd9vs74lV1eJKfT/LeMcZIcmqSw5JsWxozxvhikq8k2TSt2pTkc1OEL9maZEOSk+bGbMs9bZ3bx56cn+T2ueWmBQ8JAAAA7nMHcrO25yY5KslvT6+PTXLXGOO2ZeN2TtuWxuxcYXv2YcyGqnrovczn4iRHzi3H7f0QAAAAoNdCl6Yv89IkHxlj3HywJnMgxhi7k+xeel1VqzgbAAAAWNl+nRGvqr+d5GlJ3jO3ekeSw6vqqGXDN07blsZsXGF79mHMrjHGHfszXwAAAFgr9vfS9JckuSWzu5svuS7Jd5OcvrSiqk5McnyS7dOq7UlOrqpj5t63OcmuJF+YG3N67mnz3D4AAADgkLXwpelV9aDMQvyKMcb3ltaPMW6vqsuTXFpVt2YW129Psn26Y3qSXJNZcL+/qs7L7Pvgb0qyZbq0PEneleRVVXVJkvcmeWqSFyQ5Y38OEAAAANaS/fmO+NMyO8v93hW2nZvk+0muyuxRYluTvGJp4xjj7qp6dpJ3ZnaG+9tJrkhywdyYG6rqjMyeSX5OZnc/f9kYY4+PLgMAAIBDxcIhPsa4JsmKd0IbY9yZ5JXTsqf335jkWXv5Gx9P8phF5wYAAABr3YE8vgwAAABYkBAHAACARkIcAAAAGglxAAAAaCTEAQAAoJEQBwAAgEZCHAAAABoJcQAAAGgkxAEAAKCREAcAAIBGQhwAAAAaCXEAAABoJMQBAACgkRAHAACARkIcAAAAGglxAAAAaCTEAQAAoJEQBwAAgEZCHAAAABoJcQAAAGgkxAEAAKCREAcAAIBGQhwAAAAaCXEAAABoJMQBAACgkRAHAACARkIcAAAAGglxAAAAaCTEAQAAoJEQBwAAgEZCHAAAABoJcQAAAGgkxAEAAKCREAcAAIBGQhwAAAAaCXEAAABoJMQBAACgkRAHAACARkIcAAAAGglxAAAAaCTEAQAAoJEQBwAAgEZCHAAAABoJcQAAAGgkxAEAAKCREAcAAIBGQhwAAAAaCXEAAABoJMQBAACgkRAHAACARkIcAAAAGglxAAAAaCTEAQAAoJEQBwAAgEZCHAAAABoJcQAAAGgkxAEAAKCREAcAAIBGQhwAAAAaCXEAAABotHCIV9X/UVX/vqq+UVV3VNXnquon57ZXVV1UVV+btm+rqh9dto+jq+rKqtpVVbdV1eVV9fBlY368qv64qu6sqq9W1Xn7f5gAAACwNiwU4lX1Q0n+NMl3k/xMkkcneX2Sv54bdl6S1yQ5O8njk3w7ydaqesjcmCuTnJRkc5JnJ3lSkt+c+zsbklyT5MYkpyb51SQXVtUvLzJfAAAAWGvWLTj+nyf56hjjJXPrblj6paoqyWuTvGmM8eFp3S8k2ZnkuUk+UFWPSvLMJI8dY3xqGvPqJL9fVb8yxrg5yYuSHJ7kF8cYdyX5fFWdkuR1mQv2eVW1Psn6uVVHLHhsAAAAcJ9b9NL0n03yqar6T1V1S1X9WVX90tz2E5Icm2Tb0ooxxu1Jrk2yaVq1KcltSxE+2Zbk+5mdQV8a84kpwpdsTXLidFZ+JecnuX1uuWnBYwMAAID73KIh/neSvDzJf0vyjCTvTPJvqurF0/Zjp587l71v59y2Y5PcMr9xjPG9JLcuG7PSPub/xnIXJzlybjlu74cDAAAAvRa9NP1BST41xvgX0+s/q6ofy+z74Fcc1JktaIyxO8nupdezq+QBAABgbVn0jPjXknxh2brrkxw//b5j+rlx2ZiNc9t2JDlmfmNVrUty9LIxK+1j/m8AAADAIWfREP/TJCcuW/f3M7u7eTK7cduOJKcvbZzugP74JNunVduTHFVVp87t46nTXK6dG/OkqjpsbszmJH85xpi/QzsAAAAcUhYN8bclOa2q/kVV/b2qOjPJLyfZkiRjjJHksiS/XlU/W1UnJ3lfkpuTfGgac32SP0jy7qp6XFX9VJJ3JPnAdMf0JPmdJHclubyqTqqqFyY5J8mlB3KwAAAAsNoW+o74GOO/VtXzMrsx2gWZnQF/7RjjyrlhlyR5WGaPGTsqyZ8keeYY4865MS/KLL7/MLO7pV+V2bPHl/7O7VX19MwC/7okf5XkojHGio8uAwAAgEPFojdryxjj95L83r1sH5lF+gX3MubWJGfu5e98NslPLzo/AAAAWMsWvTQdAAAAOABCHAAAABoJcQAAAGgkxAEAAKCREAcAAIBGQhwAAAAaCXEAAABoJMQBAACgkRAHAACARkIcAAAAGglxAAAAaCTEAQAAoJEQBwAAgEZCHAAAABoJcQAAAGgkxAEAAKCREAcAAIBGQhwAAAAaCXEAAABoJMQBAACgkRAHAACARkIcAAAAGglxAAAAaCTEAQAAoJEQBwAAgEZCHAAAABoJcQAAAGgkxAEAAKCREAcAAIBGQhwAAAAaCXEAAABoJMQBAACgkRAHAACARkIcAAAAGglxAAAAaCTEAQAAoJEQBwAAgEZCHAAAABoJcQAAAGgkxAEAAKCREAcAAIBGQhwAAAAaCXEAAABoJMQBAACgkRAHAACARkIcAAAAGglxAAAAaCTEAQAAoJEQBwAAgEZCHAAAABoJcQAAAGgkxAEAAKCREAcAAIBGQhwAAAAaCXEAAABoJMQBAACgkRAHAACARkIcAAAAGglxAAAAaCTEAQAAoNFCIV5VF1bVWLZ8cW77Q6pqS1V9o6q+VVVXVdXGZfs4vqqurqrvVNUtVfXWqlq3bMxTqurTVbW7qr5UVWcd0FECAADAGrE/Z8Q/n+SH55Ynzm17W5LnJHl+kicneUSSDy5trKoHJ7k6yeFJnpDkxUnOSnLR3JgTpjEfS3JKksuSvKeqnrEfcwUAAIA1Zd3eh/wN3xtj7Fi+sqqOTPLSJGeOMT46rXtJkuur6rQxxieTPD3Jo5M8bYyxM8lnquqNSd5SVReOMe5KcnaSG8YYr592fX1VPTHJuUm27mlSVbU+yfq5VUfsx7EBAADAfWp/zoj/aFXdXFX/o6qurKrjp/WnJjksybalgWOMLyb5SpJN06pNST43RfiSrUk2JDlpbsy23NPWuX3syflJbp9bblroqAAAAKDBoiF+bWaXkj8zycuTnJDkj6vqiCTHJrlrjHHbsvfsnLZl+rlzhe3ZhzEbquqh9zK3i5McObcctw/HAwAAAK0WujR9jPGRuZefraprk9yY5AVJ7jiYE1vUGGN3kt1Lr6tqFWcDAAAAKzugx5dNZ7//vyR/L8mOJIdX1VHLhm2ctmX6uXGF7dmHMbvGGKsa+wAAAHCgDijEq+rhSf5ukq8luS7Jd5OcPrf9xCTHJ9k+rdqe5OSqOmZuN5uT7Eryhbkxp+eeNs/tAwAAAA5Ziz5H/F9V1ZOr6keq6glJ/nOSu5P8hzHG7UkuT3JpVf2jqjo1yW8l2T7dMT1JrsksuN9fVf9weiTZm5JsmS4tT5J3Jfk7VXVJVf2DqnpFZpe+v+1ADxYAAABW26KPLzsuyX9I8r8l+XqSP0ly2hjj69P2c5N8P8lVmT1KbGuSVyy9eYxxd1U9O8k7MzvD/e0kVyS5YG7MDVV1RmbhfU5mdz9/2Rhjj48uAwAAgEPFojdr+7m9bL8zySunZU9jbkzyrL3s5+NJHrPI3AAAAOBQcEDfEQcAAAAWI8QBAACgkRAHAACARkIcAAAAGglxAAAAaCTEAQAAoJEQBwAAgEZCHAAAABoJcQAAAGgkxAEAAKCREAcAAIBGQhwAAAAaCXEAAABoJMQBAACgkRAHAACARkIcAAAAGglxAAAAaCTEAQAAoJEQBwAAgEZCHAAAABoJcQAAAGgkxAEAAKCREAcAAIBGQhwAAAAaCXEAAABoJMQBAACgkRAHAACARkIcAAAAGglxAAAAaCTEAQAAoJEQBwAAgEZCHAAAABoJcQAAAGgkxAEAAKCREAcAAIBGQhwAAAAaCXEAAABoJMQBAACgkRAHAACARkIcAAAAGglxAAAAaCTEAQAAoJEQBwAAgEZCHAAAABoJcQAAAGgkxAEAAKCREAcAAIBGQhwAAAAaCXEAAABoJMQBAACgkRAHAACARkIcAAAAGglxAAAAaCTEAQAAoJEQBwAAgEZCHAAAABoJcQAAAGgkxAEAAKCREAcAAIBGQhwAAAAaHVCIV9UbqmpU1WVz6x5SVVuq6htV9a2quqqqNi573/FVdXVVfaeqbqmqt1bVumVjnlJVn66q3VX1pao660DmCgAAAGvBfod4VT02yT9L8tllm96W5DlJnp/kyUkekeSDc+97cJKrkxye5AlJXpzkrCQXzY05YRrzsSSnJLksyXuq6hn7O18AAABYC/YrxKvq4UmuTPJLSf56bv2RSV6a5HVjjI+OMa5L8pIkT6iq06ZhT0/y6CQ/P8b4zBjjI0nemOSVVXX4NObsJDeMMV4/xrh+jPGOJP93knPvZU7rq2rD0pLkiP05NgAAALgv7e8Z8S1Jrh5jbFu2/tQkhyX5X+vHGF9M8pUkm6ZVm5J8boyxc+59W5NsSHLS3Jjl+946t4+VnJ/k9rnlpn09GAAAAOiycIhX1c8l+YnMwne5Y5PcNca4bdn6ndO2pTE7V9iefRizoaoeuoepXZzkyLnluHs5DAAAAFgV6/Y+5Aeq6pFJ/nWSzWOMO++bKe2fMcbuJLuXXlfVKs4GAAAAVrboGfFTkxyT5NNV9b2q+l5mN2R7zfT7ziSHV9VRy963McmO6fcd0+vl27MPY3aNMe5YcM4AAACwZiwa4n+Y5OTM7mS+tHwqsxu3Lf3+3SSnL72hqk5McnyS7dOq7UlOrqpj5va7OcmuJF+YG3N67mnz3D4AAADgkLTQpeljjG8m+Yv5dVX17STfGGP8xfT68iSXVtWtmcX125NsH2N8cnrLNZkF9/ur6rzMvg/+piRbpsvLk+RdSV5VVZckeW+SpyZ5QZIzFj9EAAAAWDsWCvF9dG6S7ye5Ksn6zO52/oqljWOMu6vq2UnemdkZ7m8nuSLJBXNjbqiqMzJ7Jvk5md0B/WVjjK33wXwBAACgzQGH+BjjKcte35nkldOyp/fcmORZe9nvx5M85kDnBwAAAGvJ/j5HHAAAANgPQhwAAAAaCXEAAABoJMQBAACgkRAHAACARkIcAAAAGglxAAAAaCTEAQAAoJEQBwAAgEZCHAAAABoJcQAAAGgkxAEAAKCREAcAAIBGQhwAAAAaCXEAAABoJMQBAACgkRAHAACARkIcAAAAGglxAAAAaCTEAQAAoJEQBwAAgEZCHAAAABoJcQAAAGgkxAEAAKCREAcAAIBGQhwAAAAaCXEAAABoJMQBAACgkRAHAACARkIcAAAAGglxAAAAaCTEAQAAoJEQBwAAgEZCHAAAABoJcQAAAGgkxAEAAKCREAcAAIBGQhwAAAAaCXEAAABoJMQBAACgkRAHAACARkIcAAAAGglxAAAAaCTEAQAAoJEQBwAAgEZCHAAAABoJcQAAAGgkxAEAAKCREAcAAIBGQhwAAAAaCXEAAABoJMQBAACgkRAHAACARkIcAAAAGglxAAAAaCTEAQAAoJEQBwAAgEZCHAAAABoJcQAAAGgkxAEAAKDRQiFeVS+vqs9W1a5p2V5VPzO3/SFVtaWqvlFV36qqq6pq47J9HF9VV1fVd6rqlqp6a1WtWzbmKVX16araXVVfqqqzDugoAQAAYI1Y9Iz4TUnekOTUJD+Z5KNJPlxVJ03b35bkOUmen+TJSR6R5INLb66qBye5OsnhSZ6Q5MVJzkpy0dyYE6YxH0tySpLLkrynqp6x4FwBAABgzVm39yE/MMb43WWrfq2qXp7ktKq6KclLk5w5xvhoklTVS5JcX1WnjTE+meTpSR6d5GljjJ1JPlNVb0zylqq6cIxxV5Kzk9wwxnj99Deur6onJjk3ydY9za2q1idZP7fqiEWODQAAADrs93fEq+rBVfVzSR6WZHtmZ8kPS7JtacwY44tJvpJk07RqU5LPTRG+ZGuSDUlOmhuzLfe0dW4fe3J+ktvnlpsWPCQAAAC4zy0c4lV1clV9K8nuJO9K8rwxxheSHJvkrjHGbcvesnPalunnzhW2Zx/GbKiqh97L1C5OcuTccty+HREAAAD0WejS9MlfZvbd7SOT/NMkV1TVkw/qrPbDGGN3Zv9xIElSVas4GwAAAFjZwiE+fY/7S9PL66rqsUnOSfIfkxxeVUctOyu+McmO6fcdSR63bJcb57Yt/dy4wphdY4w7Fp0vAAAArCUH4zniD8rsJmnXJfluktOXNlTViUmOz+w75Jl+nlxVx8y9f3OSXUm+MDfm9NzT5rl9AAAAwCFroTPiVXVxko9kdgO2I5KcmeQpSZ4xxri9qi5PcmlV3ZpZXL89yfbpjulJck1mwf3+qjovs++DvynJlunS8mT2vfNXVdUlSd6b5KlJXpDkjP0+SgAAAFgjFr00/Zgk70vyw5ndmfyzmUX4/zNtPzfJ95NcldlZ8q1JXrH05jHG3VX17CTvzOwM97eTXJHkgrkxN1TVGZk9k/yczO5+/rIxxh4fXQYAAACHikWfI/7SvWy/M8krp2VPY25M8qy97OfjSR6zyNwAAADgUHAwviMOAAAA7CMhDgAAAI2EOAAAADQS4gAAANBIiAMAAEAjIQ4AAACNhDgAAAA0EuIAAADQSIgDAABAIyEOAAAAjYQ4AAAANBLiAAAA0EiIAwAAQCMhDgAAAI2EOAAAADQS4gAAANBIiAMAAEAjIQ4AAACNhDgAAAA0EuIAAADQSIgDAABAIyEOAAAAjYQ4AAAANBLiAAAA0EiIAwAAQCMhDgAAAI2EOAAAADQS4gAAANBIiAMAAEAjIQ4AAACNhDgAAAA0EuIAAADQSIgDAABAIyEOAAAAjYQ4AAAANBLiAAAA0EiIAwAAQCMhDgAAAI2EOAAAADQS4gAAANBIiAMAAEAjIQ4AAACNhDgAAAA0EuIAAADQSIgDAABAIyEOAAAAjYQ4AAAANBLiAAAA0EiIAwAAQCMhDgAAAI2EOAAAADQS4gAAANBIiAMAAEAjIQ4AAACNhDgAAAA0EuIAAADQSIgDAABAIyEOAAAAjYQ4AAAANBLiAAAA0GihEK+q86vqv1bVN6vqlqr6UFWduGzMQ6pqS1V9o6q+VVVXVdXGZWOOr6qrq+o7037eWlXrlo15SlV9uqp2V9WXquqs/T5KAAAAWCMWPSP+5CRbkpyWZHOSw5JcU1UPmxvztiTPSfL8afwjknxwaWNVPTjJ1UkOT/KEJC9OclaSi+bGnDCN+ViSU5JcluQ9VfWMBecLAAAAa8q6vQ/5gTHGM+dfT2epb0lyapJPVNWRSV6a5MwxxkenMS9Jcn1VnTbG+GSSpyd5dJKnjTF2JvlMVb0xyVuq6sIxxl1Jzk5ywxjj9dOfur6qnpjk3CRbV5pbVa1Psn5u1RGLHBsAAAB0ONDviB85/bx1+nlqZmfJty0NGGN8MclXkmyaVm1K8rkpwpdsTbIhyUlzY7blnrbO7WMl5ye5fW65aZEDAQAAgA77HeJV9aDMLhn/0zHGX0yrj01y1xjjtmXDd07blsbsXGF79mHMhqp66B6mdHFm/2FgaTluHw8FAAAA2ix0afoyW5L8WJInHqS5HJAxxu4ku5deV9UqzgYAAABWtl9nxKvqHUmeneQfjTHmLwHfkeTwqjpq2Vs2TtuWxmxcYXv2YcyuMcYd+zNnAAAAWAsWOiNes9PMb0/yvCRPGWPcsGzIdUm+m+T0JFdN7zkxyfFJtk9jtif5tao6Zoxxy7Ruc5JdSb4wN+ZZy/a9eW4fAIeEH3nD1as9BQ6yL7/5jNWeAsADgv8Pvf/x/6E/sOil6VuSnJnkHyf5ZlUtfaf79jHGHWOM26vq8iSXVtWtmcX125Nsn+6YniTXZBbc76+q8zL7PvibkmyZLi9PkncleVVVXZLkvUmemuQFSfyTAwAA4JC26KXpL8/sRmgfT/K1ueWFc2POTfJ7mZ0R/0Rml5n/k6WNY4y7M7us/e7MznD/+yTvS3LB3JgbMovuzUn+PMnrk7xsjLHio8sAAADgULHoc8T3ege0McadSV45LXsac2P+5qXny8d8PMljFpkfAAAArHUH+hxxAAAAYAFCHAAAABoJcQAAAGgkxAEAAKCREAcAAIBGQhwAAAAaCXEAAABoJMQBAACgkRAHAACARkIcAAAAGglxAAAAaCTEAQAAoJEQBwAAgEZCHAAAABoJcQAAAGgkxAEAAKCREAcAAIBGQhwAAAAaCXEAAABoJMQBAACgkRAHAACARkIcAAAAGglxAAAAaCTEAQAAoJEQBwAAgEZCHAAAABoJcQAAAGgkxAEAAKCREAcAAIBGQhwAAAAaCXEAAABoJMQBAACgkRAHANNQvHUAABG2SURBVACARkIcAAAAGglxAAAAaCTEAQAAoJEQBwAAgEZCHAAAABoJcQAAAGgkxAEAAKCREAcAAIBGQhwAAAAaCXEAAABoJMQBAACgkRAHAACARkIcAAAAGglxAAAAaCTEAQAAoJEQBwAAgEZCHAAAABoJcQAAAGgkxAEAAKCREAcAAIBGQhwAAAAaCXEAAABoJMQBAACgkRAHAACARkIcAAAAGglxAAAAaCTEAQAAoNHCIV5VT6qq362qm6tqVNVzl22vqrqoqr5WVXdU1baq+tFlY46uqiuraldV3VZVl1fVw5eN+fGq+uOqurOqvlpV5+3fIQIAAMDasT9nxB+W5M+TvHIP289L8pokZyd5fJJvJ9laVQ+ZG3NlkpOSbE7y7CRPSvKbSxurakOSa5LcmOTUJL+a5MKq+uX9mC8AAACsGesWfcMY4yNJPpIkVXWPbTVb8dokbxpjfHha9wtJdiZ5bpIPVNWjkjwzyWPHGJ+axrw6ye9X1a+MMW5O8qIkhyf5xTHGXUk+X1WnJHld5oJ92d9en2T93KojFj02AAAAuK8d7O+In5Dk2CTbllaMMW5Pcm2STdOqTUluW4rwybYk38/sDPrSmE9MEb5ka5ITq+qH9vC3z09y+9xy04EdCgAAABx8BzvEj51+7ly2fufctmOT3DK/cYzxvSS3Lhuz0j7m/8ZyFyc5cm45bpGJAwAAQIeFL01fq8YYu5PsXnq9/LJ5AAAAWAsO9hnxHdPPjcvWb5zbtiPJMfMbq2pdkqOXjVlpH/N/AwAAAA45BzvEb8gslE9fWjHdAf3xSbZPq7YnOaqqTp1731OnuVw7N+ZJVXXY3JjNSf5yjPHXB3nOAAAA0GZ/niP+8Ko6ZbqLeZKcML0+fowxklyW5Ner6mer6uQk70tyc5IPJckY4/okf5Dk3VX1uKr6qSTvSPKB6Y7pSfI7Se5KcnlVnVRVL0xyTpJLD+BYAQAAYNXtz3fEfzLJx+ZeL8XxFUnOSnJJZs8a/80kRyX5kyTPHGPcOfeeF2UW33+Y2d3Sr8rs2eNJZndar6qnJ9mS5Lokf5XkojHGio8uAwAAgEPF/jxH/ONJ9ngntOms+AXTsqcxtyY5cy9/57NJfnrR+QEAAMBadrC/Iw4AAADcCyEOAAAAjYQ4AAAANBLiAAAA0EiIAwAAQCMhDgAAAI2EOAAAADQS4gAAANBIiAMAAEAjIQ4AAACNhDgAAAA0EuIAAADQSIgDAABAIyEOAAAAjYQ4AAAANBLiAAAA0EiIAwAAQCMhDgAAAI2EOAAAADQS4gAAANBIiAMAAEAjIQ4AAACNhDgAAAA0EuIAAADQSIgDAABAIyEOAAAAjYQ4AAAANBLiAAAA0EiIAwAAQCMhDgAAAI2EOAAAADQS4gAAANBIiAMAAEAjIQ4AAACNhDgAAAA0EuIAAADQSIgDAABAIyEOAAAAjYQ4AAAANBLiAAAA0EiIAwAAQCMhDgAAAI2EOAAAADQS4gAAANBIiAMAAEAjIQ4AAACNhDgAAAA0EuIAAADQSIgDAABAIyEOAAAAjYQ4AAAANBLiAAAA0EiIAwAAQCMhDgAAAI2EOAAAADQS4gAAANBIiAMAAEAjIQ4AAACNhDgAAAA0EuIAAADQaE2HeFW9sqq+XFV3VtW1VfW41Z4TAAAAHIg1G+JV9cIklyb5jSQ/keTPk2ytqmNWdWIAAABwANat9gTuxeuSvHuM8VtJUlVnJzkjyS8mefPywVW1Psn6uVVHJMmuXbvu+5musu/v/s5qT4GD6IHwv9kHEp/P+x+f0fsXn9H7F5/P+xefz/ufB8JndF+PscYY9/FUFldVhyf5TpJ/Osb40Nz6K5IcNcb4xyu858Ik/1fbJAEAAGBlx40x/ueeNq7VM+L/e5IHJ9m5bP3OJP9gD++5OLNL2ecdneTWgzs1VskRSW5KclySb67yXIC/yWcU1i6fT1jbfEbvf45IcvO9DVirIb6wMcbuJLuXrb7/X/vwAFFVS79+c4zhnyusMT6jsHb5fMLa5jN6v7TXf45r9WZtf5Xk7iQbl63fmGRH/3QAAADg4FiTIT7GuCvJdUlOX1pXVQ+aXm9frXkBAADAgVrLl6ZfmuSKqvpUkv+S5LVJHpbkt1Z1VqyW3Zk9ym751w+AtcFnFNYun09Y23xGH4DW5F3Tl1TVq5L8apJjk3wmyWvGGNeu7qwAAABg/63pEAcAAID7mzX5HXEAAAC4vxLiAAAA0EiIAwAAQCMhDgAAAI2EOGteVa1f7TkAwKGkqg6rqkdW1YlVdfRqzwdYmX/PfeAS4qw5VfUzVXVFVf2Pqvpuku9U1a6q+qOq+rWqesRqzxEeyKrqUVX1G1X10ar671X1tar67PS5PdO/VMDqqKojqurlVfVHSXYl+XKS65N8vapurKp3V9VjV3WS8ADn33NZ4vFlrBlV9bwkb0lyRJLfT/Jfktyc5I4kRyf5sSQ/nWRTkt9O8sYxxtdXZbLwAFRVP5HkkiT/f3t3HyxXfddx/P0JASc8WFGUUKZUxAJNxoaHQqFCkxkERJgW2zo6dkpTbZVR2w4WaUdGoJV28KHWVqc+tGqhD6LOlHYc0UjLJHRqqoUbQAxgFIKdIcHGGlpbaGLz9Y89t9ns3d2AkHPO7b5fMzvJ7u+3dz+bzHfOfu/vnN+eA3yOyTX6nc28362qb3STVpotSX4ZuBr4d+CvmVyflwL/CLypqrZ0k1aaPX7O1SgbcfVGko3A9cDfVtWeKfOOBd4EPFZV720rnzTrkjwM/Bbw8araOWXe2cBbgHur6t1t5ZNmWZI/B66vqn/Zz7zvAF4P7KqqP20lnCQ/52oBG3FJ0lOS5OCq2n2g5kuSJM0KG3FJkiRJklrkZm1aVJodYD2VTupQkqOSXJXkliQbm9stSX4lyfd2nU/SeElOSHJ71zmkWZZkRZIPJNnUbHa6rfn7B5Ks6Dqf2uOKuBaVJKuAuao6qOss0ixqdlxeB3wd+DTwWDN0NHAecChwYVXd2U1CSZN4DJW6leQi4JPAHINj6fAx9HzgdOAVVbWum4Rqk424eiXJy/cz5QeA9/ghQupGks8D9wCX18gBJEmAPwReVFVnd5FPmmVJ3ryfKccCV3oMlbqR5B7gU1V1zYTx64BXVtWLWg2mTtiIq1eS7AEKyJRp5YcIqRtJngBOraoHJoyfDGyqqmXtJpPUHEO3AbsmTDkEWO4xVOpGcww9paoenDB+EnC3x9DZ4DXi6pttDH4TuGTcDTit64DSjNsOnDll/Ez2nmonqV2PAFdU1fHjbsDFXQeUZtxWptfhxQzqWDNgadcBpBF3Mbg+5lMTxve3Wi7pwPpt4I+TnA58hoXXiL8RuLKjbNKsmz+G/uWEcY+hUreuAT6eZA3j91n5UeCnu4mmtnlqunolybnAYVX1dxPGDwNeXFUb2k0maV6SnwSuYPCBf/4U128yaAJ+p6omNQGSDqBmx+VDJ22WmORg4LlV5Yqb1JEkLwXeDJwNLG8e3g5sBN5XVRu7yqZ22YhLkv5fmg/1RzV3d1TV7i7zSJIkLRY24pIkSZIktcjN2iRJkiRJapGNuCRJkiRJLbIRlyRJkiSpRTbikiRJkiS1yEZci06SlyV5Ttc5JI1njUr9leSyJCd0nUPSeNbo7LAR12K0HngoyVu7DiJprPVYo1JffRjYnOT3ug4iaawPY43OBBtxLUbHA68Gju46iKSxrFGpp6pqCXAycH/XWSQtZI3ODr9HXJIkSZKkFi3tOoA0TpKlwEpgefPQdmBzVe3uLpWkedaotPg0dfvcqvqPrrNIWsganS024uqVJEuAdwK/CIxu9vR4kt8Hrq2qPa2Hk2SNSovbSmAOOKjrIJLGskZniI24+uYGYC3wdmAd8Fjz+NHABcCvA4cAb+sinCRrVJIk6ZnyGnH1SpLtwOuqat2E8QuBm6rKTaCkDlijUn8lmdvPlGXAiVXlapvUAWtUw1wRV98cATw6ZXwbcFhLWSQtZI1K/bUCuBl4eML4McCJ7cWRNMIa1be4Iq5eSfI3DH5B9Jqq2jEydhTwEeCbVXVJF/mkWWeNSv2V5E7gT6rqDyaMnwLc5Wqb1A1rVMNcEVffXA7cCmxL8s/se/3pDwGbAT/gS92xRqX++hxw0pTxrwJ3tJRF0kLWqL7FFXH1TrMr84XAWez71Ugbgb93N2apW9aoJEnSM2MjLkmSJElSi5Z0HUCal+S4pzn/2AOVRdJC1qjUX9an1G/WqEbZiKtPvpDkj5KcMWlCkuckeWOS+4BXtZhNkjUq9Zn1KfWbNap9uFmb+mQFcDVwW5IngbsYfE3Sk8CRzfhKYA64qqpu7SqoNKOsUam/rE+p36xR7cNrxNU7SZYBFwPnAM8HlgE7gE3Auqq6r8N40syzRqX+sj6lfrNGNc9GXJIkSZKkFnmNuCRJkiRJLbIRlyRJkiSpRTbikiRJkiS1yEZckiRJkqQW2YhLkiRJktQiG3FJknomydokO7vO0XdJKsmlXeeQJOnpshGXJKl//gI4sesQB4LNsyRJsLTrAJIkaV9V9QTwRNc5khxSVbuewryDgKqqPS3EkiRp0XNFXJKkFiS5JMnOpmklySnN6vANQ3M+lOSjo6emJ7kuyd1JXptka5LHk9yc5IihOeuTvD/Jbyb5cpLtSa4byfBdzWt8KclXktyeZNWY13lDkoeBJye8l7XNe3l5ks3AN4DjkpyR5LYkO5qMG5KcNvS8rc1fb2ne+9ahsVckmUvyZJKHklybZOnQ+AuS3NGMb05y/tP8L5AkqTdsxCVJasdngSOAU5v7q4EdwJqhOauB9ROefwJwKXBJc1sNvH1kzuuArwEvAa4CrhlpWP8K+D7gIuB0YA74TJLvHprzg8CrgFcCp0x5P4cCbwPeAKwE/rN5fzcC5wBnAVuAW4d+YXBG8+frgWPm7yc5F7gJeB+wAvh5YC1wdTO+BPgEsKt5b5cDvzElmyRJvWYjLklSC6rqceBu9jbea4D3AqcmOTzJsQya4A0TfsQSYG1V3VdVnwU+Apw3MufeqnpHVW2pqpuAO+fnJDkHOBP4iaq6s5lzJbATePXQzzgEuKyqNlXVvVPe0sHAL1TVP1TVg1X19aq6vao+WlUPVNX9wM8xaNhXN/8GX2qeu7Oqtg/dvxa4oapurKqHquo24NcYNOQAPwKc3OS6p6ruAH51SjZJknrNRlySpPZsANYkCXAug1Xe+xmsIK8GHq2qLROeu7Wqvjp0fxuD1e1ho43z8JxVwOHAfyX5n/kbcDyD1fZ5jww1yNPsGn29JEcn+WCSLUkeB77SvOZx+/lZqxis3g/n+iBwTJJDgRcCX6yqR4ees/EpZJQkqZfcrE2SpPasB36GQeO5u6oeSLKewer4kUxeDQfYPXK/WPgL9WlzDmfQmK8Z87OHvyrta1MyDHuiqmrksRuB7wHeAjzC4NrxjQxW2ac5nMGq+CfGjI29Tl2SpMXMRlySpPbMXyd+BXub7vUMrvU+EnjPAXztOWA58L9VtfUAvcYPMzhd/VaAJM8DjhqZsxs4aEy2k6rq38b90CT3A89LckxVbWsePuvZiy1JUrs8NV2SpJZU1X8zOJ37NezdlO0O4DQG3xs+bUX8mfo0g9XpTya5IMn3J3lpknclefGz9BpbgNcmeWGSlwAfY+HXsG0FzkuyPMmRzWPvBC5rdkpf2Tz/p5JcP5T9X4Ebk6xqNnd717OUWZKk1tmIS5LUrg0MVoTXA1TVl4HNwPaqevBAvWhzGvmPMWj8/4xBY3sz8HzgsWnPbb6ubPQ09HF+lsHK/hyDzeTez2A39WFvBc4HvghsarKtY7AT/AXAF4DPMzhr4JFmfA/w48Ay4J+AD9HsqC5J0mKUhZd3SZIk7ZXkHcDqqlrTdRZJkr4deI24JEnan4uAX+o6hCRJ3y5cEZckSZIkqUVeIy5JkiRJUotsxCVJkiRJapGNuCRJkiRJLbIRlyRJkiSpRTbikiRJkiS1yEZckiRJkqQW2YhLkiRJktQiG3FJkiRJklpkIy5JkiRJUov+Dy6HkCkz+4GgAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 1200x800 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df.groupby('winner').rated.value_counts().plot(kind = 'bar')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = df.dropna().copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.to_csv('cleaned_data')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "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>rated</th>\n",
       "      <th>white_rating</th>\n",
       "      <th>black_rating</th>\n",
       "      <th>opening_eco</th>\n",
       "      <th>opening_ply</th>\n",
       "      <th>white_1move</th>\n",
       "      <th>black_1move</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>1500</td>\n",
       "      <td>1191</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>5</td>\n",
       "      <td>0.499337</td>\n",
       "      <td>0.493591</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>1322</td>\n",
       "      <td>1261</td>\n",
       "      <td>0.597381</td>\n",
       "      <td>4</td>\n",
       "      <td>0.499337</td>\n",
       "      <td>0.563776</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>1496</td>\n",
       "      <td>1500</td>\n",
       "      <td>0.442963</td>\n",
       "      <td>3</td>\n",
       "      <td>0.505715</td>\n",
       "      <td>0.515312</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>1439</td>\n",
       "      <td>1454</td>\n",
       "      <td>0.511521</td>\n",
       "      <td>3</td>\n",
       "      <td>0.499337</td>\n",
       "      <td>0.493591</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>1523</td>\n",
       "      <td>1469</td>\n",
       "      <td>0.573082</td>\n",
       "      <td>5</td>\n",
       "      <td>0.505715</td>\n",
       "      <td>0.515312</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20053</th>\n",
       "      <td>1</td>\n",
       "      <td>1691</td>\n",
       "      <td>1220</td>\n",
       "      <td>0.527273</td>\n",
       "      <td>2</td>\n",
       "      <td>0.499337</td>\n",
       "      <td>0.506024</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20054</th>\n",
       "      <td>1</td>\n",
       "      <td>1233</td>\n",
       "      <td>1196</td>\n",
       "      <td>0.600000</td>\n",
       "      <td>2</td>\n",
       "      <td>0.499337</td>\n",
       "      <td>0.561694</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20055</th>\n",
       "      <td>1</td>\n",
       "      <td>1219</td>\n",
       "      <td>1286</td>\n",
       "      <td>0.461434</td>\n",
       "      <td>3</td>\n",
       "      <td>0.499337</td>\n",
       "      <td>0.493591</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20056</th>\n",
       "      <td>1</td>\n",
       "      <td>1360</td>\n",
       "      <td>1227</td>\n",
       "      <td>0.555921</td>\n",
       "      <td>4</td>\n",
       "      <td>0.505715</td>\n",
       "      <td>0.561694</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20057</th>\n",
       "      <td>1</td>\n",
       "      <td>1235</td>\n",
       "      <td>1339</td>\n",
       "      <td>0.461434</td>\n",
       "      <td>3</td>\n",
       "      <td>0.499337</td>\n",
       "      <td>0.493591</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>20040 rows × 7 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       rated  white_rating  black_rating  opening_eco  opening_ply  \\\n",
       "0          0          1500          1191     0.400000            5   \n",
       "1          1          1322          1261     0.597381            4   \n",
       "2          1          1496          1500     0.442963            3   \n",
       "3          1          1439          1454     0.511521            3   \n",
       "4          1          1523          1469     0.573082            5   \n",
       "...      ...           ...           ...          ...          ...   \n",
       "20053      1          1691          1220     0.527273            2   \n",
       "20054      1          1233          1196     0.600000            2   \n",
       "20055      1          1219          1286     0.461434            3   \n",
       "20056      1          1360          1227     0.555921            4   \n",
       "20057      1          1235          1339     0.461434            3   \n",
       "\n",
       "       white_1move  black_1move  \n",
       "0         0.499337     0.493591  \n",
       "1         0.499337     0.563776  \n",
       "2         0.505715     0.515312  \n",
       "3         0.499337     0.493591  \n",
       "4         0.505715     0.515312  \n",
       "...            ...          ...  \n",
       "20053     0.499337     0.506024  \n",
       "20054     0.499337     0.561694  \n",
       "20055     0.499337     0.493591  \n",
       "20056     0.505715     0.561694  \n",
       "20057     0.499337     0.493591  \n",
       "\n",
       "[20040 rows x 7 columns]"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#features\n",
    "x = df.drop('winner', axis = 1)\n",
    "x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "#target\n",
    "y = df.winner"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0        1\n",
       "1        0\n",
       "2        1\n",
       "3        1\n",
       "4        1\n",
       "        ..\n",
       "20053    1\n",
       "20054    0\n",
       "20055    1\n",
       "20056    1\n",
       "20057    0\n",
       "Name: winner, Length: 20040, dtype: int64"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Train Test split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [],
   "source": [
    "x_train, x_test, y_train, y_test = train_test_split(x, y, test_size = 0.3, random_state = 0, stratify = y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Feature Scaling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import StandardScaler"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "StandardScaler()"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "scaler=StandardScaler()\n",
    "scaler.fit(x_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [],
   "source": [
    "x_train_scaled=scaler.transform(x_train)\n",
    "x_test_scaled = scaler.transform(x_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [],
   "source": [
    "#pip install xgboost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [],
   "source": [
    "#pip install catboost"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Models"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Model 1 Logistic Regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LogisticRegression(class_weight='balanced', max_iter=100000.0)"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model1 = LogisticRegression(class_weight = 'balanced',max_iter = 1e5)\n",
    "model1.fit(x_train_scaled,y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred1 = model1.predict(x_test_scaled)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.6606786427145709\n",
      "0.6549821486530347\n",
      "0.6737896494156929\n"
     ]
    }
   ],
   "source": [
    "print(accuracy_score(y_test, y_pred1))\n",
    "print(precision_score(y_test, y_pred1))\n",
    "print(recall_score(y_test, y_pred1))\n",
    "prob_1 = model1.predict_proba(x_test_scaled)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.67      0.65      0.66      3017\n",
      "           1       0.65      0.67      0.66      2995\n",
      "\n",
      "    accuracy                           0.66      6012\n",
      "   macro avg       0.66      0.66      0.66      6012\n",
      "weighted avg       0.66      0.66      0.66      6012\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_test, y_pred1))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Feature Importance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [],
   "source": [
    "#pd.Series(model1.feature_importances_, index=x_train.columns).sort_values().plot.barh()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Model 2 KNN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [],
   "source": [
    "model2= KNeighborsClassifier(2)\n",
    "model2.fit(x_train_scaled, y_train)\n",
    "y_pred2= model2.predict(x_test_scaled)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.5883233532934131\n",
      "0.648741418764302\n",
      "0.37863105175292155\n"
     ]
    }
   ],
   "source": [
    "print(accuracy_score(y_test, y_pred2))\n",
    "print(precision_score(y_test, y_pred2))\n",
    "print(recall_score(y_test, y_pred2))\n",
    "prob_2 = model1.predict_proba(x_test_scaled)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.56      0.80      0.66      3017\n",
      "           1       0.65      0.38      0.48      2995\n",
      "\n",
      "    accuracy                           0.59      6012\n",
      "   macro avg       0.61      0.59      0.57      6012\n",
      "weighted avg       0.61      0.59      0.57      6012\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_test, y_pred2))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Feature importance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [],
   "source": [
    "#importances = CV_RFR_regr.best_estimator_.feature_importances_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Model 3 Gaussian"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [],
   "source": [
    "model3= GaussianNB()\n",
    "model3.fit(x_train_scaled, y_train)\n",
    "y_pred3= model3.predict(x_test_scaled)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[1370 1647]\n",
      " [ 661 2334]]\n",
      "0.6161011310711909\n",
      "0.586284853051997\n",
      "0.7792988313856427\n"
     ]
    }
   ],
   "source": [
    "print(confusion_matrix(y_test, y_pred3))\n",
    "print(accuracy_score(y_test, y_pred3))\n",
    "print(precision_score(y_test, y_pred3))\n",
    "print(recall_score(y_test, y_pred3))\n",
    "prob_3 = model3.predict_proba(x_test_scaled)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.67      0.45      0.54      3017\n",
      "           1       0.59      0.78      0.67      2995\n",
      "\n",
      "    accuracy                           0.62      6012\n",
      "   macro avg       0.63      0.62      0.61      6012\n",
      "weighted avg       0.63      0.62      0.61      6012\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_test, y_pred3))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Feature importance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Model 4 support vector machine"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [],
   "source": [
    "model4= SVC(probability =True)\n",
    "model4.fit(x_train_scaled, y_train)\n",
    "y_pred4= model4.predict(x_test_scaled)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.6586826347305389\n",
      "0.6556619346318917\n",
      "0.6631051752921536\n"
     ]
    }
   ],
   "source": [
    "print(accuracy_score(y_test, y_pred4))\n",
    "print(precision_score(y_test, y_pred4))\n",
    "print(recall_score(y_test, y_pred4))\n",
    "prob_4 = model4.predict_proba(x_test_scaled)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.66      0.65      0.66      3017\n",
      "           1       0.66      0.66      0.66      2995\n",
      "\n",
      "    accuracy                           0.66      6012\n",
      "   macro avg       0.66      0.66      0.66      6012\n",
      "weighted avg       0.66      0.66      0.66      6012\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_test, y_pred4))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Model 5 Decision Tree"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [],
   "source": [
    "model5= DecisionTreeClassifier()\n",
    "model5.fit(x_train_scaled, y_train)\n",
    "y_pred5= model5.predict(x_test_scaled)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.6142714570858283\n",
      "0.6117724867724867\n",
      "0.6176961602671118\n"
     ]
    }
   ],
   "source": [
    "print(accuracy_score(y_test, y_pred5))\n",
    "print(precision_score(y_test, y_pred5))\n",
    "print(recall_score(y_test, y_pred5))\n",
    "prob_5 = model5.predict_proba(x_test_scaled)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.62      0.61      0.61      3017\n",
      "           1       0.61      0.62      0.61      2995\n",
      "\n",
      "    accuracy                           0.61      6012\n",
      "   macro avg       0.61      0.61      0.61      6012\n",
      "weighted avg       0.61      0.61      0.61      6012\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_test, y_pred5))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Feature Importance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7fc9af83d490>"
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1200x800 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "pd.Series(model5.feature_importances_, index=x_train.columns).sort_values().plot.barh()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Model 6 Random forest classifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [],
   "source": [
    "model6=RandomForestClassifier(class_weight='balanced')\n",
    "model6.fit(x_train_scaled, y_train)\n",
    "y_pred6 = model6.predict(x_test_scaled)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.6636726546906188\n",
      "0.6627634660421545\n",
      "0.6614357262103506\n"
     ]
    }
   ],
   "source": [
    "print(accuracy_score(y_test, y_pred6))\n",
    "print(precision_score(y_test, y_pred6))\n",
    "print(recall_score(y_test, y_pred6))\n",
    "prob_6 = model6.predict_proba(x_test_scaled)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.66      0.67      0.67      3017\n",
      "           1       0.66      0.66      0.66      2995\n",
      "\n",
      "    accuracy                           0.66      6012\n",
      "   macro avg       0.66      0.66      0.66      6012\n",
      "weighted avg       0.66      0.66      0.66      6012\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_test, y_pred6))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Feature Importance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7fc9af72c4d0>"
      ]
     },
     "execution_count": 83,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1200x800 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "pd.Series(model6.feature_importances_, index=x_train.columns).sort_values().plot.barh()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Model 7 AdaBoost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [],
   "source": [
    "model7=AdaBoostClassifier()\n",
    "model7.fit(x_train_scaled, y_train)\n",
    "y_pred7 = model7.predict(x_test_scaled)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.6551896207584831\n",
      "0.65244708994709\n",
      "0.6587646076794658\n"
     ]
    }
   ],
   "source": [
    "print(accuracy_score(y_test, y_pred7))\n",
    "print(precision_score(y_test, y_pred7))\n",
    "print(recall_score(y_test, y_pred7))\n",
    "prob_7 = model7.predict_proba(x_test_scaled)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.66      0.65      0.65      3017\n",
      "           1       0.65      0.66      0.66      2995\n",
      "\n",
      "    accuracy                           0.66      6012\n",
      "   macro avg       0.66      0.66      0.66      6012\n",
      "weighted avg       0.66      0.66      0.66      6012\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_test, y_pred7))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Feature Importance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7fc9925b2990>"
      ]
     },
     "execution_count": 87,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1200x800 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "pd.Series(model7.feature_importances_, index=x_train.columns).sort_values().plot.barh()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Modle 8 CatBoost "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Learning rate set to 0.03182\n",
      "0:\tlearn: 0.6899688\ttotal: 68.7ms\tremaining: 1m 8s\n",
      "1:\tlearn: 0.6875355\ttotal: 79ms\tremaining: 39.4s\n",
      "2:\tlearn: 0.6846276\ttotal: 88.8ms\tremaining: 29.5s\n",
      "3:\tlearn: 0.6813912\ttotal: 98.3ms\tremaining: 24.5s\n",
      "4:\tlearn: 0.6785019\ttotal: 107ms\tremaining: 21.3s\n",
      "5:\tlearn: 0.6759879\ttotal: 116ms\tremaining: 19.2s\n",
      "6:\tlearn: 0.6738920\ttotal: 125ms\tremaining: 17.8s\n",
      "7:\tlearn: 0.6716217\ttotal: 135ms\tremaining: 16.8s\n",
      "8:\tlearn: 0.6692691\ttotal: 146ms\tremaining: 16.1s\n",
      "9:\tlearn: 0.6668321\ttotal: 156ms\tremaining: 15.5s\n",
      "10:\tlearn: 0.6648713\ttotal: 167ms\tremaining: 15s\n",
      "11:\tlearn: 0.6630054\ttotal: 177ms\tremaining: 14.5s\n",
      "12:\tlearn: 0.6612359\ttotal: 187ms\tremaining: 14.2s\n",
      "13:\tlearn: 0.6593695\ttotal: 199ms\tremaining: 14s\n",
      "14:\tlearn: 0.6578061\ttotal: 210ms\tremaining: 13.8s\n",
      "15:\tlearn: 0.6561269\ttotal: 220ms\tremaining: 13.5s\n",
      "16:\tlearn: 0.6547581\ttotal: 230ms\tremaining: 13.3s\n",
      "17:\tlearn: 0.6531587\ttotal: 239ms\tremaining: 13s\n",
      "18:\tlearn: 0.6516515\ttotal: 249ms\tremaining: 12.8s\n",
      "19:\tlearn: 0.6503054\ttotal: 259ms\tremaining: 12.7s\n",
      "20:\tlearn: 0.6490130\ttotal: 269ms\tremaining: 12.5s\n",
      "21:\tlearn: 0.6475794\ttotal: 277ms\tremaining: 12.3s\n",
      "22:\tlearn: 0.6461484\ttotal: 287ms\tremaining: 12.2s\n",
      "23:\tlearn: 0.6449299\ttotal: 297ms\tremaining: 12.1s\n",
      "24:\tlearn: 0.6437289\ttotal: 304ms\tremaining: 11.9s\n",
      "25:\tlearn: 0.6425646\ttotal: 312ms\tremaining: 11.7s\n",
      "26:\tlearn: 0.6414925\ttotal: 321ms\tremaining: 11.6s\n",
      "27:\tlearn: 0.6404909\ttotal: 328ms\tremaining: 11.4s\n",
      "28:\tlearn: 0.6396454\ttotal: 336ms\tremaining: 11.3s\n",
      "29:\tlearn: 0.6387517\ttotal: 344ms\tremaining: 11.1s\n",
      "30:\tlearn: 0.6378427\ttotal: 352ms\tremaining: 11s\n",
      "31:\tlearn: 0.6370408\ttotal: 360ms\tremaining: 10.9s\n",
      "32:\tlearn: 0.6363033\ttotal: 367ms\tremaining: 10.8s\n",
      "33:\tlearn: 0.6356212\ttotal: 375ms\tremaining: 10.6s\n",
      "34:\tlearn: 0.6348682\ttotal: 382ms\tremaining: 10.5s\n",
      "35:\tlearn: 0.6340960\ttotal: 390ms\tremaining: 10.4s\n",
      "36:\tlearn: 0.6333039\ttotal: 399ms\tremaining: 10.4s\n",
      "37:\tlearn: 0.6326173\ttotal: 409ms\tremaining: 10.3s\n",
      "38:\tlearn: 0.6320483\ttotal: 417ms\tremaining: 10.3s\n",
      "39:\tlearn: 0.6314441\ttotal: 425ms\tremaining: 10.2s\n",
      "40:\tlearn: 0.6307149\ttotal: 433ms\tremaining: 10.1s\n",
      "41:\tlearn: 0.6300827\ttotal: 442ms\tremaining: 10.1s\n",
      "42:\tlearn: 0.6295599\ttotal: 451ms\tremaining: 10s\n",
      "43:\tlearn: 0.6290090\ttotal: 459ms\tremaining: 9.98s\n",
      "44:\tlearn: 0.6283167\ttotal: 470ms\tremaining: 9.97s\n",
      "45:\tlearn: 0.6276887\ttotal: 480ms\tremaining: 9.95s\n",
      "46:\tlearn: 0.6272637\ttotal: 489ms\tremaining: 9.91s\n",
      "47:\tlearn: 0.6266588\ttotal: 499ms\tremaining: 9.89s\n",
      "48:\tlearn: 0.6261989\ttotal: 508ms\tremaining: 9.86s\n",
      "49:\tlearn: 0.6256577\ttotal: 516ms\tremaining: 9.8s\n",
      "50:\tlearn: 0.6252547\ttotal: 525ms\tremaining: 9.76s\n",
      "51:\tlearn: 0.6246818\ttotal: 535ms\tremaining: 9.74s\n",
      "52:\tlearn: 0.6242074\ttotal: 544ms\tremaining: 9.71s\n",
      "53:\tlearn: 0.6236038\ttotal: 553ms\tremaining: 9.68s\n",
      "54:\tlearn: 0.6229941\ttotal: 561ms\tremaining: 9.64s\n",
      "55:\tlearn: 0.6225268\ttotal: 570ms\tremaining: 9.6s\n",
      "56:\tlearn: 0.6221487\ttotal: 578ms\tremaining: 9.56s\n",
      "57:\tlearn: 0.6218364\ttotal: 586ms\tremaining: 9.51s\n",
      "58:\tlearn: 0.6213448\ttotal: 594ms\tremaining: 9.47s\n",
      "59:\tlearn: 0.6210075\ttotal: 604ms\tremaining: 9.47s\n",
      "60:\tlearn: 0.6205916\ttotal: 612ms\tremaining: 9.42s\n",
      "61:\tlearn: 0.6202555\ttotal: 620ms\tremaining: 9.38s\n",
      "62:\tlearn: 0.6199896\ttotal: 628ms\tremaining: 9.34s\n",
      "63:\tlearn: 0.6196554\ttotal: 636ms\tremaining: 9.3s\n",
      "64:\tlearn: 0.6192962\ttotal: 645ms\tremaining: 9.27s\n",
      "65:\tlearn: 0.6189612\ttotal: 652ms\tremaining: 9.23s\n",
      "66:\tlearn: 0.6186582\ttotal: 660ms\tremaining: 9.19s\n",
      "67:\tlearn: 0.6183584\ttotal: 667ms\tremaining: 9.15s\n",
      "68:\tlearn: 0.6180384\ttotal: 675ms\tremaining: 9.11s\n",
      "69:\tlearn: 0.6177009\ttotal: 682ms\tremaining: 9.07s\n",
      "70:\tlearn: 0.6174208\ttotal: 690ms\tremaining: 9.03s\n",
      "71:\tlearn: 0.6171937\ttotal: 698ms\tremaining: 8.99s\n",
      "72:\tlearn: 0.6168310\ttotal: 706ms\tremaining: 8.96s\n",
      "73:\tlearn: 0.6165892\ttotal: 714ms\tremaining: 8.93s\n",
      "74:\tlearn: 0.6163149\ttotal: 722ms\tremaining: 8.9s\n",
      "75:\tlearn: 0.6161569\ttotal: 729ms\tremaining: 8.87s\n",
      "76:\tlearn: 0.6158623\ttotal: 737ms\tremaining: 8.84s\n",
      "77:\tlearn: 0.6155236\ttotal: 745ms\tremaining: 8.81s\n",
      "78:\tlearn: 0.6153946\ttotal: 753ms\tremaining: 8.78s\n",
      "79:\tlearn: 0.6150969\ttotal: 760ms\tremaining: 8.74s\n",
      "80:\tlearn: 0.6147729\ttotal: 768ms\tremaining: 8.72s\n",
      "81:\tlearn: 0.6145243\ttotal: 776ms\tremaining: 8.69s\n",
      "82:\tlearn: 0.6142856\ttotal: 784ms\tremaining: 8.66s\n",
      "83:\tlearn: 0.6139988\ttotal: 794ms\tremaining: 8.65s\n",
      "84:\tlearn: 0.6137030\ttotal: 803ms\tremaining: 8.64s\n",
      "85:\tlearn: 0.6134590\ttotal: 810ms\tremaining: 8.61s\n",
      "86:\tlearn: 0.6132012\ttotal: 818ms\tremaining: 8.59s\n",
      "87:\tlearn: 0.6129701\ttotal: 826ms\tremaining: 8.56s\n",
      "88:\tlearn: 0.6127452\ttotal: 834ms\tremaining: 8.54s\n",
      "89:\tlearn: 0.6124674\ttotal: 843ms\tremaining: 8.52s\n",
      "90:\tlearn: 0.6122220\ttotal: 851ms\tremaining: 8.5s\n",
      "91:\tlearn: 0.6120198\ttotal: 859ms\tremaining: 8.48s\n",
      "92:\tlearn: 0.6118487\ttotal: 868ms\tremaining: 8.46s\n",
      "93:\tlearn: 0.6115978\ttotal: 876ms\tremaining: 8.44s\n",
      "94:\tlearn: 0.6113355\ttotal: 883ms\tremaining: 8.41s\n",
      "95:\tlearn: 0.6111395\ttotal: 891ms\tremaining: 8.39s\n",
      "96:\tlearn: 0.6109361\ttotal: 900ms\tremaining: 8.38s\n",
      "97:\tlearn: 0.6107741\ttotal: 908ms\tremaining: 8.36s\n",
      "98:\tlearn: 0.6105831\ttotal: 915ms\tremaining: 8.33s\n",
      "99:\tlearn: 0.6104310\ttotal: 923ms\tremaining: 8.3s\n",
      "100:\tlearn: 0.6103025\ttotal: 931ms\tremaining: 8.29s\n",
      "101:\tlearn: 0.6100596\ttotal: 939ms\tremaining: 8.26s\n",
      "102:\tlearn: 0.6098637\ttotal: 946ms\tremaining: 8.24s\n",
      "103:\tlearn: 0.6096973\ttotal: 956ms\tremaining: 8.24s\n",
      "104:\tlearn: 0.6094806\ttotal: 964ms\tremaining: 8.22s\n",
      "105:\tlearn: 0.6092525\ttotal: 972ms\tremaining: 8.2s\n",
      "106:\tlearn: 0.6090471\ttotal: 980ms\tremaining: 8.18s\n",
      "107:\tlearn: 0.6088342\ttotal: 989ms\tremaining: 8.16s\n",
      "108:\tlearn: 0.6086700\ttotal: 1s\tremaining: 8.18s\n",
      "109:\tlearn: 0.6085535\ttotal: 1.01s\tremaining: 8.17s\n",
      "110:\tlearn: 0.6083275\ttotal: 1.02s\tremaining: 8.14s\n",
      "111:\tlearn: 0.6082360\ttotal: 1.02s\tremaining: 8.12s\n",
      "112:\tlearn: 0.6081211\ttotal: 1.03s\tremaining: 8.11s\n",
      "113:\tlearn: 0.6079582\ttotal: 1.04s\tremaining: 8.09s\n",
      "114:\tlearn: 0.6077385\ttotal: 1.05s\tremaining: 8.06s\n",
      "115:\tlearn: 0.6075705\ttotal: 1.05s\tremaining: 8.04s\n",
      "116:\tlearn: 0.6074363\ttotal: 1.06s\tremaining: 8.02s\n",
      "117:\tlearn: 0.6072859\ttotal: 1.07s\tremaining: 8s\n",
      "118:\tlearn: 0.6071306\ttotal: 1.08s\tremaining: 7.98s\n",
      "119:\tlearn: 0.6070156\ttotal: 1.09s\tremaining: 7.96s\n",
      "120:\tlearn: 0.6068252\ttotal: 1.1s\tremaining: 7.99s\n",
      "121:\tlearn: 0.6067045\ttotal: 1.11s\tremaining: 8.02s\n",
      "122:\tlearn: 0.6065804\ttotal: 1.13s\tremaining: 8.06s\n",
      "123:\tlearn: 0.6065224\ttotal: 1.15s\tremaining: 8.11s\n",
      "124:\tlearn: 0.6063716\ttotal: 1.16s\tremaining: 8.15s\n",
      "125:\tlearn: 0.6062202\ttotal: 1.19s\tremaining: 8.26s\n",
      "126:\tlearn: 0.6060557\ttotal: 1.21s\tremaining: 8.29s\n",
      "127:\tlearn: 0.6059247\ttotal: 1.21s\tremaining: 8.27s\n",
      "128:\tlearn: 0.6058106\ttotal: 1.22s\tremaining: 8.27s\n",
      "129:\tlearn: 0.6057097\ttotal: 1.23s\tremaining: 8.26s\n",
      "130:\tlearn: 0.6055741\ttotal: 1.24s\tremaining: 8.25s\n",
      "131:\tlearn: 0.6054366\ttotal: 1.25s\tremaining: 8.24s\n",
      "132:\tlearn: 0.6052711\ttotal: 1.26s\tremaining: 8.22s\n",
      "133:\tlearn: 0.6051239\ttotal: 1.27s\tremaining: 8.21s\n",
      "134:\tlearn: 0.6049980\ttotal: 1.28s\tremaining: 8.23s\n",
      "135:\tlearn: 0.6048915\ttotal: 1.29s\tremaining: 8.23s\n",
      "136:\tlearn: 0.6047717\ttotal: 1.31s\tremaining: 8.23s\n",
      "137:\tlearn: 0.6046016\ttotal: 1.31s\tremaining: 8.22s\n",
      "138:\tlearn: 0.6044717\ttotal: 1.32s\tremaining: 8.2s\n",
      "139:\tlearn: 0.6043557\ttotal: 1.33s\tremaining: 8.19s\n",
      "140:\tlearn: 0.6042616\ttotal: 1.34s\tremaining: 8.17s\n",
      "141:\tlearn: 0.6042155\ttotal: 1.35s\tremaining: 8.16s\n",
      "142:\tlearn: 0.6040726\ttotal: 1.36s\tremaining: 8.14s\n",
      "143:\tlearn: 0.6039355\ttotal: 1.37s\tremaining: 8.13s\n",
      "144:\tlearn: 0.6038119\ttotal: 1.38s\tremaining: 8.12s\n",
      "145:\tlearn: 0.6036460\ttotal: 1.38s\tremaining: 8.1s\n",
      "146:\tlearn: 0.6035485\ttotal: 1.39s\tremaining: 8.09s\n",
      "147:\tlearn: 0.6034508\ttotal: 1.4s\tremaining: 8.08s\n",
      "148:\tlearn: 0.6032398\ttotal: 1.41s\tremaining: 8.06s\n",
      "149:\tlearn: 0.6031503\ttotal: 1.42s\tremaining: 8.05s\n",
      "150:\tlearn: 0.6030239\ttotal: 1.43s\tremaining: 8.04s\n",
      "151:\tlearn: 0.6029362\ttotal: 1.44s\tremaining: 8.02s\n",
      "152:\tlearn: 0.6028759\ttotal: 1.45s\tremaining: 8s\n",
      "153:\tlearn: 0.6027669\ttotal: 1.45s\tremaining: 7.98s\n",
      "154:\tlearn: 0.6026182\ttotal: 1.46s\tremaining: 7.96s\n",
      "155:\tlearn: 0.6024998\ttotal: 1.47s\tremaining: 7.95s\n",
      "156:\tlearn: 0.6023969\ttotal: 1.48s\tremaining: 7.93s\n",
      "157:\tlearn: 0.6022640\ttotal: 1.48s\tremaining: 7.91s\n",
      "158:\tlearn: 0.6021024\ttotal: 1.49s\tremaining: 7.9s\n",
      "159:\tlearn: 0.6020204\ttotal: 1.5s\tremaining: 7.88s\n",
      "160:\tlearn: 0.6019162\ttotal: 1.51s\tremaining: 7.87s\n",
      "161:\tlearn: 0.6018265\ttotal: 1.52s\tremaining: 7.85s\n",
      "162:\tlearn: 0.6017561\ttotal: 1.52s\tremaining: 7.83s\n",
      "163:\tlearn: 0.6016626\ttotal: 1.53s\tremaining: 7.82s\n",
      "164:\tlearn: 0.6015712\ttotal: 1.54s\tremaining: 7.81s\n",
      "165:\tlearn: 0.6014789\ttotal: 1.55s\tremaining: 7.79s\n",
      "166:\tlearn: 0.6013030\ttotal: 1.56s\tremaining: 7.78s\n",
      "167:\tlearn: 0.6011503\ttotal: 1.57s\tremaining: 7.76s\n",
      "168:\tlearn: 0.6010679\ttotal: 1.58s\tremaining: 7.75s\n",
      "169:\tlearn: 0.6009157\ttotal: 1.58s\tremaining: 7.74s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "170:\tlearn: 0.6008544\ttotal: 1.59s\tremaining: 7.73s\n",
      "171:\tlearn: 0.6007681\ttotal: 1.6s\tremaining: 7.72s\n",
      "172:\tlearn: 0.6005498\ttotal: 1.61s\tremaining: 7.71s\n",
      "173:\tlearn: 0.6005081\ttotal: 1.62s\tremaining: 7.7s\n",
      "174:\tlearn: 0.6003787\ttotal: 1.63s\tremaining: 7.69s\n",
      "175:\tlearn: 0.6002430\ttotal: 1.64s\tremaining: 7.68s\n",
      "176:\tlearn: 0.6000991\ttotal: 1.65s\tremaining: 7.67s\n",
      "177:\tlearn: 0.5999947\ttotal: 1.66s\tremaining: 7.65s\n",
      "178:\tlearn: 0.5998673\ttotal: 1.67s\tremaining: 7.64s\n",
      "179:\tlearn: 0.5997630\ttotal: 1.68s\tremaining: 7.63s\n",
      "180:\tlearn: 0.5996748\ttotal: 1.68s\tremaining: 7.62s\n",
      "181:\tlearn: 0.5995961\ttotal: 1.69s\tremaining: 7.61s\n",
      "182:\tlearn: 0.5994488\ttotal: 1.7s\tremaining: 7.59s\n",
      "183:\tlearn: 0.5993549\ttotal: 1.71s\tremaining: 7.58s\n",
      "184:\tlearn: 0.5993092\ttotal: 1.72s\tremaining: 7.57s\n",
      "185:\tlearn: 0.5992571\ttotal: 1.73s\tremaining: 7.58s\n",
      "186:\tlearn: 0.5992025\ttotal: 1.75s\tremaining: 7.59s\n",
      "187:\tlearn: 0.5990741\ttotal: 1.76s\tremaining: 7.59s\n",
      "188:\tlearn: 0.5989591\ttotal: 1.77s\tremaining: 7.58s\n",
      "189:\tlearn: 0.5987783\ttotal: 1.78s\tremaining: 7.59s\n",
      "190:\tlearn: 0.5986853\ttotal: 1.81s\tremaining: 7.68s\n",
      "191:\tlearn: 0.5986233\ttotal: 1.83s\tremaining: 7.7s\n",
      "192:\tlearn: 0.5984817\ttotal: 1.84s\tremaining: 7.71s\n",
      "193:\tlearn: 0.5983512\ttotal: 1.86s\tremaining: 7.71s\n",
      "194:\tlearn: 0.5983157\ttotal: 1.87s\tremaining: 7.71s\n",
      "195:\tlearn: 0.5982591\ttotal: 1.88s\tremaining: 7.7s\n",
      "196:\tlearn: 0.5981098\ttotal: 1.89s\tremaining: 7.69s\n",
      "197:\tlearn: 0.5980605\ttotal: 1.9s\tremaining: 7.69s\n",
      "198:\tlearn: 0.5979999\ttotal: 1.91s\tremaining: 7.68s\n",
      "199:\tlearn: 0.5979340\ttotal: 1.92s\tremaining: 7.67s\n",
      "200:\tlearn: 0.5978230\ttotal: 1.93s\tremaining: 7.66s\n",
      "201:\tlearn: 0.5976153\ttotal: 1.94s\tremaining: 7.66s\n",
      "202:\tlearn: 0.5975136\ttotal: 1.95s\tremaining: 7.65s\n",
      "203:\tlearn: 0.5974302\ttotal: 1.96s\tremaining: 7.64s\n",
      "204:\tlearn: 0.5973341\ttotal: 1.97s\tremaining: 7.63s\n",
      "205:\tlearn: 0.5972390\ttotal: 1.98s\tremaining: 7.62s\n",
      "206:\tlearn: 0.5970787\ttotal: 1.99s\tremaining: 7.61s\n",
      "207:\tlearn: 0.5970137\ttotal: 2s\tremaining: 7.61s\n",
      "208:\tlearn: 0.5969368\ttotal: 2.01s\tremaining: 7.6s\n",
      "209:\tlearn: 0.5968729\ttotal: 2.02s\tremaining: 7.61s\n",
      "210:\tlearn: 0.5968498\ttotal: 2.03s\tremaining: 7.6s\n",
      "211:\tlearn: 0.5967454\ttotal: 2.04s\tremaining: 7.59s\n",
      "212:\tlearn: 0.5966947\ttotal: 2.05s\tremaining: 7.59s\n",
      "213:\tlearn: 0.5966506\ttotal: 2.06s\tremaining: 7.58s\n",
      "214:\tlearn: 0.5965860\ttotal: 2.07s\tremaining: 7.57s\n",
      "215:\tlearn: 0.5964733\ttotal: 2.08s\tremaining: 7.56s\n",
      "216:\tlearn: 0.5964418\ttotal: 2.09s\tremaining: 7.55s\n",
      "217:\tlearn: 0.5963420\ttotal: 2.1s\tremaining: 7.54s\n",
      "218:\tlearn: 0.5963067\ttotal: 2.11s\tremaining: 7.54s\n",
      "219:\tlearn: 0.5962736\ttotal: 2.12s\tremaining: 7.53s\n",
      "220:\tlearn: 0.5961716\ttotal: 2.13s\tremaining: 7.53s\n",
      "221:\tlearn: 0.5960749\ttotal: 2.15s\tremaining: 7.52s\n",
      "222:\tlearn: 0.5959754\ttotal: 2.15s\tremaining: 7.51s\n",
      "223:\tlearn: 0.5958920\ttotal: 2.17s\tremaining: 7.5s\n",
      "224:\tlearn: 0.5958436\ttotal: 2.17s\tremaining: 7.49s\n",
      "225:\tlearn: 0.5957112\ttotal: 2.19s\tremaining: 7.49s\n",
      "226:\tlearn: 0.5956024\ttotal: 2.2s\tremaining: 7.48s\n",
      "227:\tlearn: 0.5955590\ttotal: 2.21s\tremaining: 7.48s\n",
      "228:\tlearn: 0.5954527\ttotal: 2.22s\tremaining: 7.49s\n",
      "229:\tlearn: 0.5953948\ttotal: 2.23s\tremaining: 7.48s\n",
      "230:\tlearn: 0.5953310\ttotal: 2.24s\tremaining: 7.47s\n",
      "231:\tlearn: 0.5951714\ttotal: 2.25s\tremaining: 7.46s\n",
      "232:\tlearn: 0.5951183\ttotal: 2.27s\tremaining: 7.46s\n",
      "233:\tlearn: 0.5950446\ttotal: 2.28s\tremaining: 7.45s\n",
      "234:\tlearn: 0.5949712\ttotal: 2.29s\tremaining: 7.44s\n",
      "235:\tlearn: 0.5949023\ttotal: 2.29s\tremaining: 7.43s\n",
      "236:\tlearn: 0.5948457\ttotal: 2.3s\tremaining: 7.41s\n",
      "237:\tlearn: 0.5947668\ttotal: 2.31s\tremaining: 7.4s\n",
      "238:\tlearn: 0.5947171\ttotal: 2.32s\tremaining: 7.38s\n",
      "239:\tlearn: 0.5946609\ttotal: 2.33s\tremaining: 7.37s\n",
      "240:\tlearn: 0.5946017\ttotal: 2.34s\tremaining: 7.36s\n",
      "241:\tlearn: 0.5945397\ttotal: 2.35s\tremaining: 7.35s\n",
      "242:\tlearn: 0.5944621\ttotal: 2.36s\tremaining: 7.34s\n",
      "243:\tlearn: 0.5943630\ttotal: 2.37s\tremaining: 7.33s\n",
      "244:\tlearn: 0.5942522\ttotal: 2.37s\tremaining: 7.31s\n",
      "245:\tlearn: 0.5942120\ttotal: 2.38s\tremaining: 7.3s\n",
      "246:\tlearn: 0.5941419\ttotal: 2.39s\tremaining: 7.29s\n",
      "247:\tlearn: 0.5940752\ttotal: 2.4s\tremaining: 7.28s\n",
      "248:\tlearn: 0.5940407\ttotal: 2.41s\tremaining: 7.27s\n",
      "249:\tlearn: 0.5939381\ttotal: 2.42s\tremaining: 7.26s\n",
      "250:\tlearn: 0.5938587\ttotal: 2.43s\tremaining: 7.25s\n",
      "251:\tlearn: 0.5937330\ttotal: 2.44s\tremaining: 7.24s\n",
      "252:\tlearn: 0.5936978\ttotal: 2.45s\tremaining: 7.22s\n",
      "253:\tlearn: 0.5936068\ttotal: 2.45s\tremaining: 7.21s\n",
      "254:\tlearn: 0.5935283\ttotal: 2.46s\tremaining: 7.2s\n",
      "255:\tlearn: 0.5934663\ttotal: 2.47s\tremaining: 7.18s\n",
      "256:\tlearn: 0.5934089\ttotal: 2.48s\tremaining: 7.17s\n",
      "257:\tlearn: 0.5933614\ttotal: 2.49s\tremaining: 7.16s\n",
      "258:\tlearn: 0.5932105\ttotal: 2.5s\tremaining: 7.14s\n",
      "259:\tlearn: 0.5931259\ttotal: 2.51s\tremaining: 7.13s\n",
      "260:\tlearn: 0.5930455\ttotal: 2.52s\tremaining: 7.12s\n",
      "261:\tlearn: 0.5930064\ttotal: 2.52s\tremaining: 7.11s\n",
      "262:\tlearn: 0.5929186\ttotal: 2.54s\tremaining: 7.12s\n",
      "263:\tlearn: 0.5928380\ttotal: 2.55s\tremaining: 7.11s\n",
      "264:\tlearn: 0.5927433\ttotal: 2.56s\tremaining: 7.11s\n",
      "265:\tlearn: 0.5926280\ttotal: 2.58s\tremaining: 7.12s\n",
      "266:\tlearn: 0.5925491\ttotal: 2.59s\tremaining: 7.11s\n",
      "267:\tlearn: 0.5924491\ttotal: 2.6s\tremaining: 7.11s\n",
      "268:\tlearn: 0.5923277\ttotal: 2.62s\tremaining: 7.12s\n",
      "269:\tlearn: 0.5922756\ttotal: 2.63s\tremaining: 7.11s\n",
      "270:\tlearn: 0.5922111\ttotal: 2.64s\tremaining: 7.1s\n",
      "271:\tlearn: 0.5920679\ttotal: 2.65s\tremaining: 7.09s\n",
      "272:\tlearn: 0.5919601\ttotal: 2.66s\tremaining: 7.08s\n",
      "273:\tlearn: 0.5918871\ttotal: 2.67s\tremaining: 7.06s\n",
      "274:\tlearn: 0.5918426\ttotal: 2.67s\tremaining: 7.05s\n",
      "275:\tlearn: 0.5917564\ttotal: 2.68s\tremaining: 7.04s\n",
      "276:\tlearn: 0.5916788\ttotal: 2.69s\tremaining: 7.03s\n",
      "277:\tlearn: 0.5915865\ttotal: 2.7s\tremaining: 7.02s\n",
      "278:\tlearn: 0.5915335\ttotal: 2.71s\tremaining: 7s\n",
      "279:\tlearn: 0.5914578\ttotal: 2.72s\tremaining: 6.99s\n",
      "280:\tlearn: 0.5913635\ttotal: 2.73s\tremaining: 6.98s\n",
      "281:\tlearn: 0.5913303\ttotal: 2.73s\tremaining: 6.96s\n",
      "282:\tlearn: 0.5912447\ttotal: 2.74s\tremaining: 6.95s\n",
      "283:\tlearn: 0.5911763\ttotal: 2.75s\tremaining: 6.94s\n",
      "284:\tlearn: 0.5911756\ttotal: 2.76s\tremaining: 6.92s\n",
      "285:\tlearn: 0.5910998\ttotal: 2.77s\tremaining: 6.92s\n",
      "286:\tlearn: 0.5910308\ttotal: 2.78s\tremaining: 6.91s\n",
      "287:\tlearn: 0.5909875\ttotal: 2.79s\tremaining: 6.89s\n",
      "288:\tlearn: 0.5908837\ttotal: 2.8s\tremaining: 6.88s\n",
      "289:\tlearn: 0.5908263\ttotal: 2.81s\tremaining: 6.87s\n",
      "290:\tlearn: 0.5907582\ttotal: 2.82s\tremaining: 6.86s\n",
      "291:\tlearn: 0.5906534\ttotal: 2.83s\tremaining: 6.85s\n",
      "292:\tlearn: 0.5905047\ttotal: 2.84s\tremaining: 6.84s\n",
      "293:\tlearn: 0.5904444\ttotal: 2.85s\tremaining: 6.83s\n",
      "294:\tlearn: 0.5903462\ttotal: 2.85s\tremaining: 6.82s\n",
      "295:\tlearn: 0.5903148\ttotal: 2.86s\tremaining: 6.8s\n",
      "296:\tlearn: 0.5902092\ttotal: 2.87s\tremaining: 6.79s\n",
      "297:\tlearn: 0.5901171\ttotal: 2.88s\tremaining: 6.78s\n",
      "298:\tlearn: 0.5900196\ttotal: 2.89s\tremaining: 6.77s\n",
      "299:\tlearn: 0.5899585\ttotal: 2.9s\tremaining: 6.76s\n",
      "300:\tlearn: 0.5898962\ttotal: 2.9s\tremaining: 6.74s\n",
      "301:\tlearn: 0.5898167\ttotal: 2.91s\tremaining: 6.74s\n",
      "302:\tlearn: 0.5896796\ttotal: 2.92s\tremaining: 6.73s\n",
      "303:\tlearn: 0.5895976\ttotal: 2.93s\tremaining: 6.72s\n",
      "304:\tlearn: 0.5895291\ttotal: 2.94s\tremaining: 6.71s\n",
      "305:\tlearn: 0.5894529\ttotal: 2.95s\tremaining: 6.7s\n",
      "306:\tlearn: 0.5893594\ttotal: 2.96s\tremaining: 6.69s\n",
      "307:\tlearn: 0.5892985\ttotal: 2.97s\tremaining: 6.68s\n",
      "308:\tlearn: 0.5891999\ttotal: 2.98s\tremaining: 6.67s\n",
      "309:\tlearn: 0.5891411\ttotal: 2.99s\tremaining: 6.66s\n",
      "310:\tlearn: 0.5890136\ttotal: 3s\tremaining: 6.65s\n",
      "311:\tlearn: 0.5888989\ttotal: 3.01s\tremaining: 6.64s\n",
      "312:\tlearn: 0.5887710\ttotal: 3.02s\tremaining: 6.64s\n",
      "313:\tlearn: 0.5886881\ttotal: 3.04s\tremaining: 6.63s\n",
      "314:\tlearn: 0.5886329\ttotal: 3.04s\tremaining: 6.62s\n",
      "315:\tlearn: 0.5885498\ttotal: 3.06s\tremaining: 6.62s\n",
      "316:\tlearn: 0.5884424\ttotal: 3.07s\tremaining: 6.62s\n",
      "317:\tlearn: 0.5883572\ttotal: 3.08s\tremaining: 6.61s\n",
      "318:\tlearn: 0.5882914\ttotal: 3.09s\tremaining: 6.6s\n",
      "319:\tlearn: 0.5882107\ttotal: 3.1s\tremaining: 6.59s\n",
      "320:\tlearn: 0.5881366\ttotal: 3.12s\tremaining: 6.59s\n",
      "321:\tlearn: 0.5880563\ttotal: 3.13s\tremaining: 6.58s\n",
      "322:\tlearn: 0.5879439\ttotal: 3.14s\tremaining: 6.58s\n",
      "323:\tlearn: 0.5878913\ttotal: 3.15s\tremaining: 6.57s\n",
      "324:\tlearn: 0.5878596\ttotal: 3.16s\tremaining: 6.57s\n",
      "325:\tlearn: 0.5877829\ttotal: 3.18s\tremaining: 6.57s\n",
      "326:\tlearn: 0.5876127\ttotal: 3.19s\tremaining: 6.56s\n",
      "327:\tlearn: 0.5875250\ttotal: 3.2s\tremaining: 6.56s\n",
      "328:\tlearn: 0.5874510\ttotal: 3.22s\tremaining: 6.57s\n",
      "329:\tlearn: 0.5872875\ttotal: 3.23s\tremaining: 6.57s\n",
      "330:\tlearn: 0.5871889\ttotal: 3.25s\tremaining: 6.56s\n",
      "331:\tlearn: 0.5870825\ttotal: 3.26s\tremaining: 6.55s\n",
      "332:\tlearn: 0.5870039\ttotal: 3.27s\tremaining: 6.55s\n",
      "333:\tlearn: 0.5869240\ttotal: 3.28s\tremaining: 6.54s\n",
      "334:\tlearn: 0.5868086\ttotal: 3.29s\tremaining: 6.53s\n",
      "335:\tlearn: 0.5867688\ttotal: 3.3s\tremaining: 6.53s\n",
      "336:\tlearn: 0.5866906\ttotal: 3.31s\tremaining: 6.52s\n",
      "337:\tlearn: 0.5865941\ttotal: 3.32s\tremaining: 6.51s\n",
      "338:\tlearn: 0.5865577\ttotal: 3.33s\tremaining: 6.5s\n",
      "339:\tlearn: 0.5864339\ttotal: 3.34s\tremaining: 6.49s\n",
      "340:\tlearn: 0.5863553\ttotal: 3.35s\tremaining: 6.48s\n",
      "341:\tlearn: 0.5862979\ttotal: 3.36s\tremaining: 6.47s\n",
      "342:\tlearn: 0.5862807\ttotal: 3.37s\tremaining: 6.45s\n",
      "343:\tlearn: 0.5862293\ttotal: 3.38s\tremaining: 6.44s\n",
      "344:\tlearn: 0.5861454\ttotal: 3.38s\tremaining: 6.43s\n",
      "345:\tlearn: 0.5860927\ttotal: 3.39s\tremaining: 6.42s\n",
      "346:\tlearn: 0.5860282\ttotal: 3.4s\tremaining: 6.4s\n",
      "347:\tlearn: 0.5859319\ttotal: 3.41s\tremaining: 6.39s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "348:\tlearn: 0.5858487\ttotal: 3.42s\tremaining: 6.38s\n",
      "349:\tlearn: 0.5857881\ttotal: 3.43s\tremaining: 6.37s\n",
      "350:\tlearn: 0.5857251\ttotal: 3.44s\tremaining: 6.36s\n",
      "351:\tlearn: 0.5856668\ttotal: 3.45s\tremaining: 6.35s\n",
      "352:\tlearn: 0.5855847\ttotal: 3.46s\tremaining: 6.34s\n",
      "353:\tlearn: 0.5854975\ttotal: 3.47s\tremaining: 6.33s\n",
      "354:\tlearn: 0.5853800\ttotal: 3.48s\tremaining: 6.32s\n",
      "355:\tlearn: 0.5853187\ttotal: 3.49s\tremaining: 6.31s\n",
      "356:\tlearn: 0.5852494\ttotal: 3.5s\tremaining: 6.3s\n",
      "357:\tlearn: 0.5851440\ttotal: 3.51s\tremaining: 6.29s\n",
      "358:\tlearn: 0.5850487\ttotal: 3.52s\tremaining: 6.28s\n",
      "359:\tlearn: 0.5849624\ttotal: 3.53s\tremaining: 6.27s\n",
      "360:\tlearn: 0.5848935\ttotal: 3.54s\tremaining: 6.26s\n",
      "361:\tlearn: 0.5848206\ttotal: 3.55s\tremaining: 6.25s\n",
      "362:\tlearn: 0.5847603\ttotal: 3.56s\tremaining: 6.25s\n",
      "363:\tlearn: 0.5847045\ttotal: 3.57s\tremaining: 6.24s\n",
      "364:\tlearn: 0.5846178\ttotal: 3.58s\tremaining: 6.23s\n",
      "365:\tlearn: 0.5845309\ttotal: 3.59s\tremaining: 6.22s\n",
      "366:\tlearn: 0.5844582\ttotal: 3.6s\tremaining: 6.21s\n",
      "367:\tlearn: 0.5843370\ttotal: 3.61s\tremaining: 6.2s\n",
      "368:\tlearn: 0.5841828\ttotal: 3.62s\tremaining: 6.19s\n",
      "369:\tlearn: 0.5841158\ttotal: 3.63s\tremaining: 6.19s\n",
      "370:\tlearn: 0.5840376\ttotal: 3.65s\tremaining: 6.18s\n",
      "371:\tlearn: 0.5838353\ttotal: 3.66s\tremaining: 6.18s\n",
      "372:\tlearn: 0.5837352\ttotal: 3.67s\tremaining: 6.17s\n",
      "373:\tlearn: 0.5836746\ttotal: 3.69s\tremaining: 6.17s\n",
      "374:\tlearn: 0.5835742\ttotal: 3.7s\tremaining: 6.17s\n",
      "375:\tlearn: 0.5834852\ttotal: 3.71s\tremaining: 6.16s\n",
      "376:\tlearn: 0.5834101\ttotal: 3.72s\tremaining: 6.16s\n",
      "377:\tlearn: 0.5833251\ttotal: 3.74s\tremaining: 6.15s\n",
      "378:\tlearn: 0.5832397\ttotal: 3.75s\tremaining: 6.14s\n",
      "379:\tlearn: 0.5831458\ttotal: 3.76s\tremaining: 6.14s\n",
      "380:\tlearn: 0.5830740\ttotal: 3.77s\tremaining: 6.13s\n",
      "381:\tlearn: 0.5829813\ttotal: 3.79s\tremaining: 6.13s\n",
      "382:\tlearn: 0.5828648\ttotal: 3.8s\tremaining: 6.12s\n",
      "383:\tlearn: 0.5828022\ttotal: 3.81s\tremaining: 6.12s\n",
      "384:\tlearn: 0.5826993\ttotal: 3.83s\tremaining: 6.12s\n",
      "385:\tlearn: 0.5826026\ttotal: 3.85s\tremaining: 6.12s\n",
      "386:\tlearn: 0.5825441\ttotal: 3.85s\tremaining: 6.11s\n",
      "387:\tlearn: 0.5824862\ttotal: 3.87s\tremaining: 6.1s\n",
      "388:\tlearn: 0.5824271\ttotal: 3.87s\tremaining: 6.08s\n",
      "389:\tlearn: 0.5823624\ttotal: 3.88s\tremaining: 6.07s\n",
      "390:\tlearn: 0.5823068\ttotal: 3.89s\tremaining: 6.06s\n",
      "391:\tlearn: 0.5821542\ttotal: 3.9s\tremaining: 6.05s\n",
      "392:\tlearn: 0.5820840\ttotal: 3.91s\tremaining: 6.04s\n",
      "393:\tlearn: 0.5820172\ttotal: 3.92s\tremaining: 6.02s\n",
      "394:\tlearn: 0.5819390\ttotal: 3.92s\tremaining: 6.01s\n",
      "395:\tlearn: 0.5818461\ttotal: 3.93s\tremaining: 6s\n",
      "396:\tlearn: 0.5817551\ttotal: 3.94s\tremaining: 5.99s\n",
      "397:\tlearn: 0.5816497\ttotal: 3.95s\tremaining: 5.97s\n",
      "398:\tlearn: 0.5815715\ttotal: 3.96s\tremaining: 5.96s\n",
      "399:\tlearn: 0.5814886\ttotal: 3.96s\tremaining: 5.95s\n",
      "400:\tlearn: 0.5814251\ttotal: 3.97s\tremaining: 5.94s\n",
      "401:\tlearn: 0.5813107\ttotal: 3.98s\tremaining: 5.93s\n",
      "402:\tlearn: 0.5812210\ttotal: 3.99s\tremaining: 5.92s\n",
      "403:\tlearn: 0.5811799\ttotal: 4s\tremaining: 5.9s\n",
      "404:\tlearn: 0.5810822\ttotal: 4.01s\tremaining: 5.89s\n",
      "405:\tlearn: 0.5810127\ttotal: 4.02s\tremaining: 5.88s\n",
      "406:\tlearn: 0.5809244\ttotal: 4.03s\tremaining: 5.88s\n",
      "407:\tlearn: 0.5807900\ttotal: 4.04s\tremaining: 5.87s\n",
      "408:\tlearn: 0.5807617\ttotal: 4.05s\tremaining: 5.86s\n",
      "409:\tlearn: 0.5807148\ttotal: 4.07s\tremaining: 5.85s\n",
      "410:\tlearn: 0.5806320\ttotal: 4.08s\tremaining: 5.84s\n",
      "411:\tlearn: 0.5805946\ttotal: 4.08s\tremaining: 5.83s\n",
      "412:\tlearn: 0.5804527\ttotal: 4.09s\tremaining: 5.82s\n",
      "413:\tlearn: 0.5803475\ttotal: 4.11s\tremaining: 5.81s\n",
      "414:\tlearn: 0.5802409\ttotal: 4.11s\tremaining: 5.8s\n",
      "415:\tlearn: 0.5801080\ttotal: 4.12s\tremaining: 5.79s\n",
      "416:\tlearn: 0.5800057\ttotal: 4.13s\tremaining: 5.77s\n",
      "417:\tlearn: 0.5799142\ttotal: 4.14s\tremaining: 5.76s\n",
      "418:\tlearn: 0.5798204\ttotal: 4.15s\tremaining: 5.75s\n",
      "419:\tlearn: 0.5797518\ttotal: 4.15s\tremaining: 5.74s\n",
      "420:\tlearn: 0.5796703\ttotal: 4.16s\tremaining: 5.73s\n",
      "421:\tlearn: 0.5796347\ttotal: 4.17s\tremaining: 5.71s\n",
      "422:\tlearn: 0.5795341\ttotal: 4.18s\tremaining: 5.7s\n",
      "423:\tlearn: 0.5794455\ttotal: 4.19s\tremaining: 5.69s\n",
      "424:\tlearn: 0.5793929\ttotal: 4.2s\tremaining: 5.68s\n",
      "425:\tlearn: 0.5793265\ttotal: 4.21s\tremaining: 5.67s\n",
      "426:\tlearn: 0.5791940\ttotal: 4.22s\tremaining: 5.66s\n",
      "427:\tlearn: 0.5791238\ttotal: 4.23s\tremaining: 5.66s\n",
      "428:\tlearn: 0.5790390\ttotal: 4.24s\tremaining: 5.65s\n",
      "429:\tlearn: 0.5790044\ttotal: 4.25s\tremaining: 5.64s\n",
      "430:\tlearn: 0.5789188\ttotal: 4.26s\tremaining: 5.63s\n",
      "431:\tlearn: 0.5788294\ttotal: 4.27s\tremaining: 5.62s\n",
      "432:\tlearn: 0.5787816\ttotal: 4.28s\tremaining: 5.61s\n",
      "433:\tlearn: 0.5786524\ttotal: 4.29s\tremaining: 5.6s\n",
      "434:\tlearn: 0.5785770\ttotal: 4.3s\tremaining: 5.59s\n",
      "435:\tlearn: 0.5784907\ttotal: 4.31s\tremaining: 5.58s\n",
      "436:\tlearn: 0.5783779\ttotal: 4.32s\tremaining: 5.57s\n",
      "437:\tlearn: 0.5782887\ttotal: 4.33s\tremaining: 5.56s\n",
      "438:\tlearn: 0.5781696\ttotal: 4.34s\tremaining: 5.55s\n",
      "439:\tlearn: 0.5780974\ttotal: 4.35s\tremaining: 5.54s\n",
      "440:\tlearn: 0.5780574\ttotal: 4.36s\tremaining: 5.53s\n",
      "441:\tlearn: 0.5779868\ttotal: 4.37s\tremaining: 5.52s\n",
      "442:\tlearn: 0.5779123\ttotal: 4.38s\tremaining: 5.51s\n",
      "443:\tlearn: 0.5778849\ttotal: 4.39s\tremaining: 5.5s\n",
      "444:\tlearn: 0.5778251\ttotal: 4.4s\tremaining: 5.49s\n",
      "445:\tlearn: 0.5777530\ttotal: 4.41s\tremaining: 5.47s\n",
      "446:\tlearn: 0.5776468\ttotal: 4.42s\tremaining: 5.46s\n",
      "447:\tlearn: 0.5775633\ttotal: 4.42s\tremaining: 5.45s\n",
      "448:\tlearn: 0.5774971\ttotal: 4.43s\tremaining: 5.44s\n",
      "449:\tlearn: 0.5773979\ttotal: 4.44s\tremaining: 5.43s\n",
      "450:\tlearn: 0.5773156\ttotal: 4.45s\tremaining: 5.42s\n",
      "451:\tlearn: 0.5772567\ttotal: 4.46s\tremaining: 5.41s\n",
      "452:\tlearn: 0.5772041\ttotal: 4.47s\tremaining: 5.39s\n",
      "453:\tlearn: 0.5771364\ttotal: 4.48s\tremaining: 5.38s\n",
      "454:\tlearn: 0.5770931\ttotal: 4.49s\tremaining: 5.37s\n",
      "455:\tlearn: 0.5770389\ttotal: 4.5s\tremaining: 5.36s\n",
      "456:\tlearn: 0.5769099\ttotal: 4.5s\tremaining: 5.35s\n",
      "457:\tlearn: 0.5768573\ttotal: 4.51s\tremaining: 5.34s\n",
      "458:\tlearn: 0.5767171\ttotal: 4.52s\tremaining: 5.33s\n",
      "459:\tlearn: 0.5766277\ttotal: 4.53s\tremaining: 5.32s\n",
      "460:\tlearn: 0.5765784\ttotal: 4.54s\tremaining: 5.31s\n",
      "461:\tlearn: 0.5764687\ttotal: 4.55s\tremaining: 5.3s\n",
      "462:\tlearn: 0.5764061\ttotal: 4.56s\tremaining: 5.29s\n",
      "463:\tlearn: 0.5763469\ttotal: 4.57s\tremaining: 5.28s\n",
      "464:\tlearn: 0.5762532\ttotal: 4.58s\tremaining: 5.27s\n",
      "465:\tlearn: 0.5761752\ttotal: 4.59s\tremaining: 5.26s\n",
      "466:\tlearn: 0.5760938\ttotal: 4.6s\tremaining: 5.25s\n",
      "467:\tlearn: 0.5759943\ttotal: 4.62s\tremaining: 5.25s\n",
      "468:\tlearn: 0.5759027\ttotal: 4.63s\tremaining: 5.24s\n",
      "469:\tlearn: 0.5758384\ttotal: 4.64s\tremaining: 5.24s\n",
      "470:\tlearn: 0.5757318\ttotal: 4.65s\tremaining: 5.23s\n",
      "471:\tlearn: 0.5756693\ttotal: 4.66s\tremaining: 5.22s\n",
      "472:\tlearn: 0.5755493\ttotal: 4.67s\tremaining: 5.21s\n",
      "473:\tlearn: 0.5754753\ttotal: 4.68s\tremaining: 5.2s\n",
      "474:\tlearn: 0.5754103\ttotal: 4.7s\tremaining: 5.19s\n",
      "475:\tlearn: 0.5753376\ttotal: 4.7s\tremaining: 5.18s\n",
      "476:\tlearn: 0.5752801\ttotal: 4.71s\tremaining: 5.17s\n",
      "477:\tlearn: 0.5752060\ttotal: 4.72s\tremaining: 5.16s\n",
      "478:\tlearn: 0.5751396\ttotal: 4.73s\tremaining: 5.15s\n",
      "479:\tlearn: 0.5750669\ttotal: 4.74s\tremaining: 5.14s\n",
      "480:\tlearn: 0.5750032\ttotal: 4.75s\tremaining: 5.13s\n",
      "481:\tlearn: 0.5748664\ttotal: 4.76s\tremaining: 5.12s\n",
      "482:\tlearn: 0.5748100\ttotal: 4.77s\tremaining: 5.11s\n",
      "483:\tlearn: 0.5747772\ttotal: 4.78s\tremaining: 5.1s\n",
      "484:\tlearn: 0.5746878\ttotal: 4.79s\tremaining: 5.09s\n",
      "485:\tlearn: 0.5746381\ttotal: 4.8s\tremaining: 5.08s\n",
      "486:\tlearn: 0.5745739\ttotal: 4.81s\tremaining: 5.07s\n",
      "487:\tlearn: 0.5745313\ttotal: 4.82s\tremaining: 5.05s\n",
      "488:\tlearn: 0.5744535\ttotal: 4.83s\tremaining: 5.04s\n",
      "489:\tlearn: 0.5743235\ttotal: 4.84s\tremaining: 5.04s\n",
      "490:\tlearn: 0.5742287\ttotal: 4.85s\tremaining: 5.03s\n",
      "491:\tlearn: 0.5741101\ttotal: 4.86s\tremaining: 5.01s\n",
      "492:\tlearn: 0.5740107\ttotal: 4.87s\tremaining: 5s\n",
      "493:\tlearn: 0.5739572\ttotal: 4.88s\tremaining: 5s\n",
      "494:\tlearn: 0.5738580\ttotal: 4.89s\tremaining: 4.99s\n",
      "495:\tlearn: 0.5737408\ttotal: 4.9s\tremaining: 4.97s\n",
      "496:\tlearn: 0.5736935\ttotal: 4.91s\tremaining: 4.96s\n",
      "497:\tlearn: 0.5735819\ttotal: 4.92s\tremaining: 4.96s\n",
      "498:\tlearn: 0.5735096\ttotal: 4.93s\tremaining: 4.95s\n",
      "499:\tlearn: 0.5734427\ttotal: 4.94s\tremaining: 4.94s\n",
      "500:\tlearn: 0.5733293\ttotal: 4.95s\tremaining: 4.93s\n",
      "501:\tlearn: 0.5732589\ttotal: 4.96s\tremaining: 4.92s\n",
      "502:\tlearn: 0.5731993\ttotal: 4.96s\tremaining: 4.91s\n",
      "503:\tlearn: 0.5730776\ttotal: 4.97s\tremaining: 4.9s\n",
      "504:\tlearn: 0.5730048\ttotal: 4.99s\tremaining: 4.89s\n",
      "505:\tlearn: 0.5728895\ttotal: 5s\tremaining: 4.88s\n",
      "506:\tlearn: 0.5728094\ttotal: 5.01s\tremaining: 4.87s\n",
      "507:\tlearn: 0.5727604\ttotal: 5.02s\tremaining: 4.86s\n",
      "508:\tlearn: 0.5726494\ttotal: 5.03s\tremaining: 4.85s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "509:\tlearn: 0.5726009\ttotal: 5.04s\tremaining: 4.84s\n",
      "510:\tlearn: 0.5725320\ttotal: 5.05s\tremaining: 4.83s\n",
      "511:\tlearn: 0.5725039\ttotal: 5.05s\tremaining: 4.82s\n",
      "512:\tlearn: 0.5724246\ttotal: 5.07s\tremaining: 4.81s\n",
      "513:\tlearn: 0.5722866\ttotal: 5.07s\tremaining: 4.8s\n",
      "514:\tlearn: 0.5722437\ttotal: 5.08s\tremaining: 4.79s\n",
      "515:\tlearn: 0.5721588\ttotal: 5.09s\tremaining: 4.78s\n",
      "516:\tlearn: 0.5721023\ttotal: 5.1s\tremaining: 4.76s\n",
      "517:\tlearn: 0.5720274\ttotal: 5.11s\tremaining: 4.75s\n",
      "518:\tlearn: 0.5719358\ttotal: 5.12s\tremaining: 4.74s\n",
      "519:\tlearn: 0.5718491\ttotal: 5.13s\tremaining: 4.73s\n",
      "520:\tlearn: 0.5717838\ttotal: 5.14s\tremaining: 4.72s\n",
      "521:\tlearn: 0.5717163\ttotal: 5.15s\tremaining: 4.71s\n",
      "522:\tlearn: 0.5716434\ttotal: 5.16s\tremaining: 4.7s\n",
      "523:\tlearn: 0.5715778\ttotal: 5.16s\tremaining: 4.69s\n",
      "524:\tlearn: 0.5715211\ttotal: 5.17s\tremaining: 4.68s\n",
      "525:\tlearn: 0.5714128\ttotal: 5.18s\tremaining: 4.67s\n",
      "526:\tlearn: 0.5713545\ttotal: 5.19s\tremaining: 4.66s\n",
      "527:\tlearn: 0.5712413\ttotal: 5.2s\tremaining: 4.65s\n",
      "528:\tlearn: 0.5711676\ttotal: 5.21s\tremaining: 4.64s\n",
      "529:\tlearn: 0.5710661\ttotal: 5.22s\tremaining: 4.63s\n",
      "530:\tlearn: 0.5709782\ttotal: 5.22s\tremaining: 4.62s\n",
      "531:\tlearn: 0.5708728\ttotal: 5.24s\tremaining: 4.61s\n",
      "532:\tlearn: 0.5707941\ttotal: 5.25s\tremaining: 4.59s\n",
      "533:\tlearn: 0.5707302\ttotal: 5.25s\tremaining: 4.58s\n",
      "534:\tlearn: 0.5706423\ttotal: 5.26s\tremaining: 4.57s\n",
      "535:\tlearn: 0.5705019\ttotal: 5.27s\tremaining: 4.56s\n",
      "536:\tlearn: 0.5704111\ttotal: 5.28s\tremaining: 4.55s\n",
      "537:\tlearn: 0.5703582\ttotal: 5.29s\tremaining: 4.54s\n",
      "538:\tlearn: 0.5703328\ttotal: 5.3s\tremaining: 4.53s\n",
      "539:\tlearn: 0.5702391\ttotal: 5.31s\tremaining: 4.52s\n",
      "540:\tlearn: 0.5701786\ttotal: 5.32s\tremaining: 4.51s\n",
      "541:\tlearn: 0.5700993\ttotal: 5.32s\tremaining: 4.5s\n",
      "542:\tlearn: 0.5700179\ttotal: 5.33s\tremaining: 4.49s\n",
      "543:\tlearn: 0.5699817\ttotal: 5.34s\tremaining: 4.48s\n",
      "544:\tlearn: 0.5698956\ttotal: 5.35s\tremaining: 4.47s\n",
      "545:\tlearn: 0.5698613\ttotal: 5.37s\tremaining: 4.46s\n",
      "546:\tlearn: 0.5697819\ttotal: 5.38s\tremaining: 4.45s\n",
      "547:\tlearn: 0.5696855\ttotal: 5.39s\tremaining: 4.45s\n",
      "548:\tlearn: 0.5696017\ttotal: 5.4s\tremaining: 4.44s\n",
      "549:\tlearn: 0.5695297\ttotal: 5.42s\tremaining: 4.43s\n",
      "550:\tlearn: 0.5694383\ttotal: 5.43s\tremaining: 4.42s\n",
      "551:\tlearn: 0.5693776\ttotal: 5.44s\tremaining: 4.41s\n",
      "552:\tlearn: 0.5693025\ttotal: 5.45s\tremaining: 4.4s\n",
      "553:\tlearn: 0.5692448\ttotal: 5.46s\tremaining: 4.39s\n",
      "554:\tlearn: 0.5691430\ttotal: 5.46s\tremaining: 4.38s\n",
      "555:\tlearn: 0.5690643\ttotal: 5.47s\tremaining: 4.37s\n",
      "556:\tlearn: 0.5689991\ttotal: 5.48s\tremaining: 4.36s\n",
      "557:\tlearn: 0.5689498\ttotal: 5.49s\tremaining: 4.35s\n",
      "558:\tlearn: 0.5688914\ttotal: 5.5s\tremaining: 4.34s\n",
      "559:\tlearn: 0.5687996\ttotal: 5.51s\tremaining: 4.33s\n",
      "560:\tlearn: 0.5687002\ttotal: 5.52s\tremaining: 4.32s\n",
      "561:\tlearn: 0.5685672\ttotal: 5.53s\tremaining: 4.31s\n",
      "562:\tlearn: 0.5685401\ttotal: 5.54s\tremaining: 4.3s\n",
      "563:\tlearn: 0.5684992\ttotal: 5.55s\tremaining: 4.29s\n",
      "564:\tlearn: 0.5684229\ttotal: 5.55s\tremaining: 4.28s\n",
      "565:\tlearn: 0.5683904\ttotal: 5.56s\tremaining: 4.27s\n",
      "566:\tlearn: 0.5683375\ttotal: 5.57s\tremaining: 4.25s\n",
      "567:\tlearn: 0.5682724\ttotal: 5.58s\tremaining: 4.24s\n",
      "568:\tlearn: 0.5681774\ttotal: 5.59s\tremaining: 4.23s\n",
      "569:\tlearn: 0.5680791\ttotal: 5.6s\tremaining: 4.22s\n",
      "570:\tlearn: 0.5679761\ttotal: 5.61s\tremaining: 4.21s\n",
      "571:\tlearn: 0.5678866\ttotal: 5.62s\tremaining: 4.2s\n",
      "572:\tlearn: 0.5678170\ttotal: 5.63s\tremaining: 4.19s\n",
      "573:\tlearn: 0.5677904\ttotal: 5.63s\tremaining: 4.18s\n",
      "574:\tlearn: 0.5677001\ttotal: 5.64s\tremaining: 4.17s\n",
      "575:\tlearn: 0.5675998\ttotal: 5.65s\tremaining: 4.16s\n",
      "576:\tlearn: 0.5675687\ttotal: 5.66s\tremaining: 4.15s\n",
      "577:\tlearn: 0.5674938\ttotal: 5.67s\tremaining: 4.14s\n",
      "578:\tlearn: 0.5674307\ttotal: 5.68s\tremaining: 4.13s\n",
      "579:\tlearn: 0.5673907\ttotal: 5.69s\tremaining: 4.12s\n",
      "580:\tlearn: 0.5673276\ttotal: 5.7s\tremaining: 4.11s\n",
      "581:\tlearn: 0.5672870\ttotal: 5.71s\tremaining: 4.1s\n",
      "582:\tlearn: 0.5672315\ttotal: 5.72s\tremaining: 4.09s\n",
      "583:\tlearn: 0.5671868\ttotal: 5.74s\tremaining: 4.08s\n",
      "584:\tlearn: 0.5670776\ttotal: 5.75s\tremaining: 4.08s\n",
      "585:\tlearn: 0.5669561\ttotal: 5.76s\tremaining: 4.07s\n",
      "586:\tlearn: 0.5669224\ttotal: 5.77s\tremaining: 4.06s\n",
      "587:\tlearn: 0.5668489\ttotal: 5.78s\tremaining: 4.05s\n",
      "588:\tlearn: 0.5667869\ttotal: 5.79s\tremaining: 4.04s\n",
      "589:\tlearn: 0.5667414\ttotal: 5.8s\tremaining: 4.03s\n",
      "590:\tlearn: 0.5666644\ttotal: 5.82s\tremaining: 4.02s\n",
      "591:\tlearn: 0.5666041\ttotal: 5.83s\tremaining: 4.02s\n",
      "592:\tlearn: 0.5665718\ttotal: 5.84s\tremaining: 4.01s\n",
      "593:\tlearn: 0.5665014\ttotal: 5.85s\tremaining: 4s\n",
      "594:\tlearn: 0.5664345\ttotal: 5.87s\tremaining: 3.99s\n",
      "595:\tlearn: 0.5663413\ttotal: 5.88s\tremaining: 3.99s\n",
      "596:\tlearn: 0.5662934\ttotal: 5.89s\tremaining: 3.98s\n",
      "597:\tlearn: 0.5661965\ttotal: 5.91s\tremaining: 3.97s\n",
      "598:\tlearn: 0.5661488\ttotal: 5.92s\tremaining: 3.96s\n",
      "599:\tlearn: 0.5660908\ttotal: 5.93s\tremaining: 3.95s\n",
      "600:\tlearn: 0.5660287\ttotal: 5.94s\tremaining: 3.94s\n",
      "601:\tlearn: 0.5659578\ttotal: 5.95s\tremaining: 3.93s\n",
      "602:\tlearn: 0.5658665\ttotal: 5.96s\tremaining: 3.92s\n",
      "603:\tlearn: 0.5658230\ttotal: 5.97s\tremaining: 3.91s\n",
      "604:\tlearn: 0.5657109\ttotal: 5.98s\tremaining: 3.9s\n",
      "605:\tlearn: 0.5655511\ttotal: 5.99s\tremaining: 3.89s\n",
      "606:\tlearn: 0.5654379\ttotal: 6s\tremaining: 3.88s\n",
      "607:\tlearn: 0.5653600\ttotal: 6.01s\tremaining: 3.87s\n",
      "608:\tlearn: 0.5653001\ttotal: 6.01s\tremaining: 3.86s\n",
      "609:\tlearn: 0.5651945\ttotal: 6.02s\tremaining: 3.85s\n",
      "610:\tlearn: 0.5651292\ttotal: 6.03s\tremaining: 3.84s\n",
      "611:\tlearn: 0.5650297\ttotal: 6.04s\tremaining: 3.83s\n",
      "612:\tlearn: 0.5649684\ttotal: 6.05s\tremaining: 3.82s\n",
      "613:\tlearn: 0.5648339\ttotal: 6.06s\tremaining: 3.81s\n",
      "614:\tlearn: 0.5647448\ttotal: 6.07s\tremaining: 3.8s\n",
      "615:\tlearn: 0.5646382\ttotal: 6.07s\tremaining: 3.79s\n",
      "616:\tlearn: 0.5645845\ttotal: 6.08s\tremaining: 3.77s\n",
      "617:\tlearn: 0.5645186\ttotal: 6.09s\tremaining: 3.76s\n",
      "618:\tlearn: 0.5644573\ttotal: 6.1s\tremaining: 3.75s\n",
      "619:\tlearn: 0.5643859\ttotal: 6.11s\tremaining: 3.74s\n",
      "620:\tlearn: 0.5643169\ttotal: 6.11s\tremaining: 3.73s\n",
      "621:\tlearn: 0.5642657\ttotal: 6.12s\tremaining: 3.72s\n",
      "622:\tlearn: 0.5641806\ttotal: 6.13s\tremaining: 3.71s\n",
      "623:\tlearn: 0.5641103\ttotal: 6.14s\tremaining: 3.7s\n",
      "624:\tlearn: 0.5640331\ttotal: 6.14s\tremaining: 3.69s\n",
      "625:\tlearn: 0.5639404\ttotal: 6.15s\tremaining: 3.68s\n",
      "626:\tlearn: 0.5638811\ttotal: 6.16s\tremaining: 3.67s\n",
      "627:\tlearn: 0.5638323\ttotal: 6.17s\tremaining: 3.65s\n",
      "628:\tlearn: 0.5637987\ttotal: 6.18s\tremaining: 3.64s\n",
      "629:\tlearn: 0.5637176\ttotal: 6.18s\tremaining: 3.63s\n",
      "630:\tlearn: 0.5636070\ttotal: 6.19s\tremaining: 3.62s\n",
      "631:\tlearn: 0.5635422\ttotal: 6.2s\tremaining: 3.61s\n",
      "632:\tlearn: 0.5634661\ttotal: 6.21s\tremaining: 3.6s\n",
      "633:\tlearn: 0.5633806\ttotal: 6.22s\tremaining: 3.59s\n",
      "634:\tlearn: 0.5633361\ttotal: 6.22s\tremaining: 3.58s\n",
      "635:\tlearn: 0.5632307\ttotal: 6.23s\tremaining: 3.57s\n",
      "636:\tlearn: 0.5631834\ttotal: 6.24s\tremaining: 3.56s\n",
      "637:\tlearn: 0.5630849\ttotal: 6.25s\tremaining: 3.55s\n",
      "638:\tlearn: 0.5630229\ttotal: 6.26s\tremaining: 3.54s\n",
      "639:\tlearn: 0.5629818\ttotal: 6.27s\tremaining: 3.53s\n",
      "640:\tlearn: 0.5628476\ttotal: 6.28s\tremaining: 3.52s\n",
      "641:\tlearn: 0.5627768\ttotal: 6.29s\tremaining: 3.51s\n",
      "642:\tlearn: 0.5627179\ttotal: 6.3s\tremaining: 3.5s\n",
      "643:\tlearn: 0.5626012\ttotal: 6.3s\tremaining: 3.48s\n",
      "644:\tlearn: 0.5624923\ttotal: 6.32s\tremaining: 3.48s\n",
      "645:\tlearn: 0.5624084\ttotal: 6.33s\tremaining: 3.47s\n",
      "646:\tlearn: 0.5623066\ttotal: 6.33s\tremaining: 3.46s\n",
      "647:\tlearn: 0.5622605\ttotal: 6.34s\tremaining: 3.44s\n",
      "648:\tlearn: 0.5622061\ttotal: 6.35s\tremaining: 3.43s\n",
      "649:\tlearn: 0.5621174\ttotal: 6.36s\tremaining: 3.42s\n",
      "650:\tlearn: 0.5620735\ttotal: 6.36s\tremaining: 3.41s\n",
      "651:\tlearn: 0.5619836\ttotal: 6.37s\tremaining: 3.4s\n",
      "652:\tlearn: 0.5619209\ttotal: 6.38s\tremaining: 3.39s\n",
      "653:\tlearn: 0.5618320\ttotal: 6.39s\tremaining: 3.38s\n",
      "654:\tlearn: 0.5617612\ttotal: 6.39s\tremaining: 3.37s\n",
      "655:\tlearn: 0.5616883\ttotal: 6.4s\tremaining: 3.36s\n",
      "656:\tlearn: 0.5616085\ttotal: 6.41s\tremaining: 3.35s\n",
      "657:\tlearn: 0.5615012\ttotal: 6.42s\tremaining: 3.33s\n",
      "658:\tlearn: 0.5614589\ttotal: 6.42s\tremaining: 3.32s\n",
      "659:\tlearn: 0.5614244\ttotal: 6.43s\tremaining: 3.31s\n",
      "660:\tlearn: 0.5613378\ttotal: 6.44s\tremaining: 3.3s\n",
      "661:\tlearn: 0.5612721\ttotal: 6.45s\tremaining: 3.29s\n",
      "662:\tlearn: 0.5612216\ttotal: 6.46s\tremaining: 3.28s\n",
      "663:\tlearn: 0.5611105\ttotal: 6.47s\tremaining: 3.27s\n",
      "664:\tlearn: 0.5610223\ttotal: 6.47s\tremaining: 3.26s\n",
      "665:\tlearn: 0.5609901\ttotal: 6.48s\tremaining: 3.25s\n",
      "666:\tlearn: 0.5609161\ttotal: 6.49s\tremaining: 3.24s\n",
      "667:\tlearn: 0.5608323\ttotal: 6.5s\tremaining: 3.23s\n",
      "668:\tlearn: 0.5607328\ttotal: 6.51s\tremaining: 3.22s\n",
      "669:\tlearn: 0.5606267\ttotal: 6.51s\tremaining: 3.21s\n",
      "670:\tlearn: 0.5605680\ttotal: 6.52s\tremaining: 3.2s\n",
      "671:\tlearn: 0.5605248\ttotal: 6.54s\tremaining: 3.19s\n",
      "672:\tlearn: 0.5604684\ttotal: 6.57s\tremaining: 3.19s\n",
      "673:\tlearn: 0.5603486\ttotal: 6.59s\tremaining: 3.19s\n",
      "674:\tlearn: 0.5602943\ttotal: 6.6s\tremaining: 3.18s\n",
      "675:\tlearn: 0.5602419\ttotal: 6.61s\tremaining: 3.17s\n",
      "676:\tlearn: 0.5601558\ttotal: 6.62s\tremaining: 3.16s\n",
      "677:\tlearn: 0.5600918\ttotal: 6.62s\tremaining: 3.15s\n",
      "678:\tlearn: 0.5599904\ttotal: 6.63s\tremaining: 3.13s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "679:\tlearn: 0.5599402\ttotal: 6.64s\tremaining: 3.13s\n",
      "680:\tlearn: 0.5598754\ttotal: 6.65s\tremaining: 3.12s\n",
      "681:\tlearn: 0.5598080\ttotal: 6.66s\tremaining: 3.1s\n",
      "682:\tlearn: 0.5597118\ttotal: 6.67s\tremaining: 3.1s\n",
      "683:\tlearn: 0.5596776\ttotal: 6.68s\tremaining: 3.08s\n",
      "684:\tlearn: 0.5596001\ttotal: 6.69s\tremaining: 3.07s\n",
      "685:\tlearn: 0.5595278\ttotal: 6.69s\tremaining: 3.06s\n",
      "686:\tlearn: 0.5594587\ttotal: 6.7s\tremaining: 3.05s\n",
      "687:\tlearn: 0.5594105\ttotal: 6.71s\tremaining: 3.04s\n",
      "688:\tlearn: 0.5593248\ttotal: 6.72s\tremaining: 3.03s\n",
      "689:\tlearn: 0.5592767\ttotal: 6.72s\tremaining: 3.02s\n",
      "690:\tlearn: 0.5591547\ttotal: 6.73s\tremaining: 3.01s\n",
      "691:\tlearn: 0.5590770\ttotal: 6.74s\tremaining: 3s\n",
      "692:\tlearn: 0.5590321\ttotal: 6.75s\tremaining: 2.99s\n",
      "693:\tlearn: 0.5589769\ttotal: 6.75s\tremaining: 2.98s\n",
      "694:\tlearn: 0.5589051\ttotal: 6.76s\tremaining: 2.97s\n",
      "695:\tlearn: 0.5588363\ttotal: 6.77s\tremaining: 2.96s\n",
      "696:\tlearn: 0.5587701\ttotal: 6.78s\tremaining: 2.95s\n",
      "697:\tlearn: 0.5587379\ttotal: 6.79s\tremaining: 2.94s\n",
      "698:\tlearn: 0.5586893\ttotal: 6.8s\tremaining: 2.93s\n",
      "699:\tlearn: 0.5586226\ttotal: 6.81s\tremaining: 2.92s\n",
      "700:\tlearn: 0.5585341\ttotal: 6.82s\tremaining: 2.91s\n",
      "701:\tlearn: 0.5584580\ttotal: 6.83s\tremaining: 2.9s\n",
      "702:\tlearn: 0.5583895\ttotal: 6.84s\tremaining: 2.89s\n",
      "703:\tlearn: 0.5583266\ttotal: 6.85s\tremaining: 2.88s\n",
      "704:\tlearn: 0.5582810\ttotal: 6.86s\tremaining: 2.87s\n",
      "705:\tlearn: 0.5582307\ttotal: 6.87s\tremaining: 2.86s\n",
      "706:\tlearn: 0.5581816\ttotal: 6.88s\tremaining: 2.85s\n",
      "707:\tlearn: 0.5581272\ttotal: 6.89s\tremaining: 2.84s\n",
      "708:\tlearn: 0.5580568\ttotal: 6.9s\tremaining: 2.83s\n",
      "709:\tlearn: 0.5579943\ttotal: 6.91s\tremaining: 2.82s\n",
      "710:\tlearn: 0.5579285\ttotal: 6.92s\tremaining: 2.81s\n",
      "711:\tlearn: 0.5578642\ttotal: 6.93s\tremaining: 2.8s\n",
      "712:\tlearn: 0.5578133\ttotal: 6.94s\tremaining: 2.79s\n",
      "713:\tlearn: 0.5577779\ttotal: 6.95s\tremaining: 2.78s\n",
      "714:\tlearn: 0.5577666\ttotal: 6.96s\tremaining: 2.77s\n",
      "715:\tlearn: 0.5576576\ttotal: 6.97s\tremaining: 2.77s\n",
      "716:\tlearn: 0.5576090\ttotal: 6.98s\tremaining: 2.75s\n",
      "717:\tlearn: 0.5575447\ttotal: 6.99s\tremaining: 2.75s\n",
      "718:\tlearn: 0.5574832\ttotal: 7s\tremaining: 2.74s\n",
      "719:\tlearn: 0.5574065\ttotal: 7.01s\tremaining: 2.73s\n",
      "720:\tlearn: 0.5573777\ttotal: 7.02s\tremaining: 2.72s\n",
      "721:\tlearn: 0.5573463\ttotal: 7.03s\tremaining: 2.71s\n",
      "722:\tlearn: 0.5572652\ttotal: 7.04s\tremaining: 2.7s\n",
      "723:\tlearn: 0.5572120\ttotal: 7.06s\tremaining: 2.69s\n",
      "724:\tlearn: 0.5571492\ttotal: 7.07s\tremaining: 2.68s\n",
      "725:\tlearn: 0.5571128\ttotal: 7.08s\tremaining: 2.67s\n",
      "726:\tlearn: 0.5570574\ttotal: 7.09s\tremaining: 2.66s\n",
      "727:\tlearn: 0.5570099\ttotal: 7.09s\tremaining: 2.65s\n",
      "728:\tlearn: 0.5569562\ttotal: 7.1s\tremaining: 2.64s\n",
      "729:\tlearn: 0.5569081\ttotal: 7.11s\tremaining: 2.63s\n",
      "730:\tlearn: 0.5568026\ttotal: 7.12s\tremaining: 2.62s\n",
      "731:\tlearn: 0.5567179\ttotal: 7.13s\tremaining: 2.61s\n",
      "732:\tlearn: 0.5566783\ttotal: 7.14s\tremaining: 2.6s\n",
      "733:\tlearn: 0.5565903\ttotal: 7.15s\tremaining: 2.59s\n",
      "734:\tlearn: 0.5564638\ttotal: 7.16s\tremaining: 2.58s\n",
      "735:\tlearn: 0.5564216\ttotal: 7.17s\tremaining: 2.57s\n",
      "736:\tlearn: 0.5563950\ttotal: 7.17s\tremaining: 2.56s\n",
      "737:\tlearn: 0.5563366\ttotal: 7.18s\tremaining: 2.55s\n",
      "738:\tlearn: 0.5562515\ttotal: 7.19s\tremaining: 2.54s\n",
      "739:\tlearn: 0.5561520\ttotal: 7.2s\tremaining: 2.53s\n",
      "740:\tlearn: 0.5560902\ttotal: 7.21s\tremaining: 2.52s\n",
      "741:\tlearn: 0.5560347\ttotal: 7.22s\tremaining: 2.51s\n",
      "742:\tlearn: 0.5559726\ttotal: 7.23s\tremaining: 2.5s\n",
      "743:\tlearn: 0.5559117\ttotal: 7.24s\tremaining: 2.49s\n",
      "744:\tlearn: 0.5558580\ttotal: 7.25s\tremaining: 2.48s\n",
      "745:\tlearn: 0.5558228\ttotal: 7.26s\tremaining: 2.47s\n",
      "746:\tlearn: 0.5557485\ttotal: 7.27s\tremaining: 2.46s\n",
      "747:\tlearn: 0.5556862\ttotal: 7.28s\tremaining: 2.45s\n",
      "748:\tlearn: 0.5556264\ttotal: 7.29s\tremaining: 2.44s\n",
      "749:\tlearn: 0.5555533\ttotal: 7.3s\tremaining: 2.43s\n",
      "750:\tlearn: 0.5554739\ttotal: 7.31s\tremaining: 2.42s\n",
      "751:\tlearn: 0.5553611\ttotal: 7.32s\tremaining: 2.41s\n",
      "752:\tlearn: 0.5553048\ttotal: 7.33s\tremaining: 2.4s\n",
      "753:\tlearn: 0.5552049\ttotal: 7.34s\tremaining: 2.39s\n",
      "754:\tlearn: 0.5550854\ttotal: 7.35s\tremaining: 2.38s\n",
      "755:\tlearn: 0.5550384\ttotal: 7.36s\tremaining: 2.38s\n",
      "756:\tlearn: 0.5550003\ttotal: 7.37s\tremaining: 2.37s\n",
      "757:\tlearn: 0.5549519\ttotal: 7.38s\tremaining: 2.36s\n",
      "758:\tlearn: 0.5548796\ttotal: 7.39s\tremaining: 2.35s\n",
      "759:\tlearn: 0.5548375\ttotal: 7.4s\tremaining: 2.33s\n",
      "760:\tlearn: 0.5548051\ttotal: 7.4s\tremaining: 2.33s\n",
      "761:\tlearn: 0.5547536\ttotal: 7.41s\tremaining: 2.31s\n",
      "762:\tlearn: 0.5546285\ttotal: 7.42s\tremaining: 2.3s\n",
      "763:\tlearn: 0.5545848\ttotal: 7.43s\tremaining: 2.29s\n",
      "764:\tlearn: 0.5545279\ttotal: 7.43s\tremaining: 2.28s\n",
      "765:\tlearn: 0.5544665\ttotal: 7.44s\tremaining: 2.27s\n",
      "766:\tlearn: 0.5544040\ttotal: 7.45s\tremaining: 2.26s\n",
      "767:\tlearn: 0.5543508\ttotal: 7.46s\tremaining: 2.25s\n",
      "768:\tlearn: 0.5543010\ttotal: 7.47s\tremaining: 2.24s\n",
      "769:\tlearn: 0.5542337\ttotal: 7.48s\tremaining: 2.23s\n",
      "770:\tlearn: 0.5541680\ttotal: 7.49s\tremaining: 2.22s\n",
      "771:\tlearn: 0.5541345\ttotal: 7.5s\tremaining: 2.21s\n",
      "772:\tlearn: 0.5540468\ttotal: 7.5s\tremaining: 2.2s\n",
      "773:\tlearn: 0.5539784\ttotal: 7.51s\tremaining: 2.19s\n",
      "774:\tlearn: 0.5539237\ttotal: 7.52s\tremaining: 2.18s\n",
      "775:\tlearn: 0.5538620\ttotal: 7.53s\tremaining: 2.17s\n",
      "776:\tlearn: 0.5537956\ttotal: 7.54s\tremaining: 2.16s\n",
      "777:\tlearn: 0.5537615\ttotal: 7.54s\tremaining: 2.15s\n",
      "778:\tlearn: 0.5537324\ttotal: 7.55s\tremaining: 2.14s\n",
      "779:\tlearn: 0.5536710\ttotal: 7.56s\tremaining: 2.13s\n",
      "780:\tlearn: 0.5536004\ttotal: 7.57s\tremaining: 2.12s\n",
      "781:\tlearn: 0.5535248\ttotal: 7.57s\tremaining: 2.11s\n",
      "782:\tlearn: 0.5534692\ttotal: 7.58s\tremaining: 2.1s\n",
      "783:\tlearn: 0.5533939\ttotal: 7.59s\tremaining: 2.09s\n",
      "784:\tlearn: 0.5532969\ttotal: 7.6s\tremaining: 2.08s\n",
      "785:\tlearn: 0.5532148\ttotal: 7.61s\tremaining: 2.07s\n",
      "786:\tlearn: 0.5531509\ttotal: 7.61s\tremaining: 2.06s\n",
      "787:\tlearn: 0.5530933\ttotal: 7.62s\tremaining: 2.05s\n",
      "788:\tlearn: 0.5530156\ttotal: 7.63s\tremaining: 2.04s\n",
      "789:\tlearn: 0.5529251\ttotal: 7.65s\tremaining: 2.03s\n",
      "790:\tlearn: 0.5528644\ttotal: 7.67s\tremaining: 2.02s\n",
      "791:\tlearn: 0.5527704\ttotal: 7.68s\tremaining: 2.02s\n",
      "792:\tlearn: 0.5527359\ttotal: 7.68s\tremaining: 2.01s\n",
      "793:\tlearn: 0.5527005\ttotal: 7.7s\tremaining: 2s\n",
      "794:\tlearn: 0.5526267\ttotal: 7.7s\tremaining: 1.99s\n",
      "795:\tlearn: 0.5525387\ttotal: 7.71s\tremaining: 1.98s\n",
      "796:\tlearn: 0.5524959\ttotal: 7.72s\tremaining: 1.97s\n",
      "797:\tlearn: 0.5524037\ttotal: 7.73s\tremaining: 1.96s\n",
      "798:\tlearn: 0.5523628\ttotal: 7.74s\tremaining: 1.95s\n",
      "799:\tlearn: 0.5522912\ttotal: 7.75s\tremaining: 1.94s\n",
      "800:\tlearn: 0.5522283\ttotal: 7.76s\tremaining: 1.93s\n",
      "801:\tlearn: 0.5521624\ttotal: 7.76s\tremaining: 1.92s\n",
      "802:\tlearn: 0.5520883\ttotal: 7.77s\tremaining: 1.91s\n",
      "803:\tlearn: 0.5519949\ttotal: 7.78s\tremaining: 1.9s\n",
      "804:\tlearn: 0.5519273\ttotal: 7.79s\tremaining: 1.89s\n",
      "805:\tlearn: 0.5518510\ttotal: 7.8s\tremaining: 1.88s\n",
      "806:\tlearn: 0.5517828\ttotal: 7.81s\tremaining: 1.87s\n",
      "807:\tlearn: 0.5517477\ttotal: 7.82s\tremaining: 1.86s\n",
      "808:\tlearn: 0.5516882\ttotal: 7.82s\tremaining: 1.85s\n",
      "809:\tlearn: 0.5516421\ttotal: 7.83s\tremaining: 1.84s\n",
      "810:\tlearn: 0.5515856\ttotal: 7.84s\tremaining: 1.83s\n",
      "811:\tlearn: 0.5515195\ttotal: 7.85s\tremaining: 1.82s\n",
      "812:\tlearn: 0.5514859\ttotal: 7.86s\tremaining: 1.81s\n",
      "813:\tlearn: 0.5514386\ttotal: 7.87s\tremaining: 1.8s\n",
      "814:\tlearn: 0.5513946\ttotal: 7.88s\tremaining: 1.79s\n",
      "815:\tlearn: 0.5513419\ttotal: 7.89s\tremaining: 1.78s\n",
      "816:\tlearn: 0.5513239\ttotal: 7.9s\tremaining: 1.77s\n",
      "817:\tlearn: 0.5512703\ttotal: 7.91s\tremaining: 1.76s\n",
      "818:\tlearn: 0.5512183\ttotal: 7.92s\tremaining: 1.75s\n",
      "819:\tlearn: 0.5511240\ttotal: 7.93s\tremaining: 1.74s\n",
      "820:\tlearn: 0.5510553\ttotal: 7.94s\tremaining: 1.73s\n",
      "821:\tlearn: 0.5509716\ttotal: 7.95s\tremaining: 1.72s\n",
      "822:\tlearn: 0.5509034\ttotal: 7.96s\tremaining: 1.71s\n",
      "823:\tlearn: 0.5508479\ttotal: 7.97s\tremaining: 1.7s\n",
      "824:\tlearn: 0.5507934\ttotal: 7.98s\tremaining: 1.69s\n",
      "825:\tlearn: 0.5507163\ttotal: 7.99s\tremaining: 1.68s\n",
      "826:\tlearn: 0.5506410\ttotal: 8s\tremaining: 1.67s\n",
      "827:\tlearn: 0.5505578\ttotal: 8s\tremaining: 1.66s\n",
      "828:\tlearn: 0.5504892\ttotal: 8.01s\tremaining: 1.65s\n",
      "829:\tlearn: 0.5504078\ttotal: 8.02s\tremaining: 1.64s\n",
      "830:\tlearn: 0.5503551\ttotal: 8.03s\tremaining: 1.63s\n",
      "831:\tlearn: 0.5503096\ttotal: 8.04s\tremaining: 1.62s\n",
      "832:\tlearn: 0.5502188\ttotal: 8.05s\tremaining: 1.61s\n",
      "833:\tlearn: 0.5501385\ttotal: 8.06s\tremaining: 1.6s\n",
      "834:\tlearn: 0.5500739\ttotal: 8.07s\tremaining: 1.59s\n",
      "835:\tlearn: 0.5500308\ttotal: 8.07s\tremaining: 1.58s\n",
      "836:\tlearn: 0.5499816\ttotal: 8.08s\tremaining: 1.57s\n",
      "837:\tlearn: 0.5499306\ttotal: 8.09s\tremaining: 1.56s\n",
      "838:\tlearn: 0.5498696\ttotal: 8.1s\tremaining: 1.55s\n",
      "839:\tlearn: 0.5498240\ttotal: 8.11s\tremaining: 1.54s\n",
      "840:\tlearn: 0.5497798\ttotal: 8.12s\tremaining: 1.53s\n",
      "841:\tlearn: 0.5497350\ttotal: 8.12s\tremaining: 1.52s\n",
      "842:\tlearn: 0.5497005\ttotal: 8.13s\tremaining: 1.51s\n",
      "843:\tlearn: 0.5496369\ttotal: 8.14s\tremaining: 1.5s\n",
      "844:\tlearn: 0.5495723\ttotal: 8.15s\tremaining: 1.49s\n",
      "845:\tlearn: 0.5495269\ttotal: 8.16s\tremaining: 1.48s\n",
      "846:\tlearn: 0.5494651\ttotal: 8.16s\tremaining: 1.47s\n",
      "847:\tlearn: 0.5494128\ttotal: 8.17s\tremaining: 1.46s\n",
      "848:\tlearn: 0.5493397\ttotal: 8.18s\tremaining: 1.46s\n",
      "849:\tlearn: 0.5492972\ttotal: 8.19s\tremaining: 1.45s\n",
      "850:\tlearn: 0.5492423\ttotal: 8.2s\tremaining: 1.44s\n",
      "851:\tlearn: 0.5491161\ttotal: 8.21s\tremaining: 1.43s\n",
      "852:\tlearn: 0.5490666\ttotal: 8.22s\tremaining: 1.42s\n",
      "853:\tlearn: 0.5490016\ttotal: 8.23s\tremaining: 1.41s\n",
      "854:\tlearn: 0.5489084\ttotal: 8.24s\tremaining: 1.4s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "855:\tlearn: 0.5488269\ttotal: 8.25s\tremaining: 1.39s\n",
      "856:\tlearn: 0.5487857\ttotal: 8.26s\tremaining: 1.38s\n",
      "857:\tlearn: 0.5487263\ttotal: 8.27s\tremaining: 1.37s\n",
      "858:\tlearn: 0.5486285\ttotal: 8.28s\tremaining: 1.36s\n",
      "859:\tlearn: 0.5485558\ttotal: 8.29s\tremaining: 1.35s\n",
      "860:\tlearn: 0.5484853\ttotal: 8.3s\tremaining: 1.34s\n",
      "861:\tlearn: 0.5483895\ttotal: 8.31s\tremaining: 1.33s\n",
      "862:\tlearn: 0.5483433\ttotal: 8.32s\tremaining: 1.32s\n",
      "863:\tlearn: 0.5482879\ttotal: 8.32s\tremaining: 1.31s\n",
      "864:\tlearn: 0.5482196\ttotal: 8.33s\tremaining: 1.3s\n",
      "865:\tlearn: 0.5481378\ttotal: 8.34s\tremaining: 1.29s\n",
      "866:\tlearn: 0.5480648\ttotal: 8.35s\tremaining: 1.28s\n",
      "867:\tlearn: 0.5480352\ttotal: 8.36s\tremaining: 1.27s\n",
      "868:\tlearn: 0.5479554\ttotal: 8.36s\tremaining: 1.26s\n",
      "869:\tlearn: 0.5478653\ttotal: 8.37s\tremaining: 1.25s\n",
      "870:\tlearn: 0.5478126\ttotal: 8.38s\tremaining: 1.24s\n",
      "871:\tlearn: 0.5477490\ttotal: 8.39s\tremaining: 1.23s\n",
      "872:\tlearn: 0.5477093\ttotal: 8.39s\tremaining: 1.22s\n",
      "873:\tlearn: 0.5476106\ttotal: 8.4s\tremaining: 1.21s\n",
      "874:\tlearn: 0.5475981\ttotal: 8.41s\tremaining: 1.2s\n",
      "875:\tlearn: 0.5475138\ttotal: 8.42s\tremaining: 1.19s\n",
      "876:\tlearn: 0.5474556\ttotal: 8.43s\tremaining: 1.18s\n",
      "877:\tlearn: 0.5474176\ttotal: 8.43s\tremaining: 1.17s\n",
      "878:\tlearn: 0.5473306\ttotal: 8.44s\tremaining: 1.16s\n",
      "879:\tlearn: 0.5472499\ttotal: 8.45s\tremaining: 1.15s\n",
      "880:\tlearn: 0.5471947\ttotal: 8.46s\tremaining: 1.14s\n",
      "881:\tlearn: 0.5471178\ttotal: 8.47s\tremaining: 1.13s\n",
      "882:\tlearn: 0.5470846\ttotal: 8.48s\tremaining: 1.12s\n",
      "883:\tlearn: 0.5470070\ttotal: 8.48s\tremaining: 1.11s\n",
      "884:\tlearn: 0.5469444\ttotal: 8.49s\tremaining: 1.1s\n",
      "885:\tlearn: 0.5468939\ttotal: 8.5s\tremaining: 1.09s\n",
      "886:\tlearn: 0.5467799\ttotal: 8.51s\tremaining: 1.08s\n",
      "887:\tlearn: 0.5467246\ttotal: 8.52s\tremaining: 1.07s\n",
      "888:\tlearn: 0.5466516\ttotal: 8.53s\tremaining: 1.06s\n",
      "889:\tlearn: 0.5466166\ttotal: 8.54s\tremaining: 1.05s\n",
      "890:\tlearn: 0.5465485\ttotal: 8.55s\tremaining: 1.05s\n",
      "891:\tlearn: 0.5464964\ttotal: 8.56s\tremaining: 1.04s\n",
      "892:\tlearn: 0.5464335\ttotal: 8.57s\tremaining: 1.03s\n",
      "893:\tlearn: 0.5463888\ttotal: 8.58s\tremaining: 1.02s\n",
      "894:\tlearn: 0.5463818\ttotal: 8.59s\tremaining: 1.01s\n",
      "895:\tlearn: 0.5463564\ttotal: 8.6s\tremaining: 998ms\n",
      "896:\tlearn: 0.5463037\ttotal: 8.61s\tremaining: 989ms\n",
      "897:\tlearn: 0.5462760\ttotal: 8.62s\tremaining: 979ms\n",
      "898:\tlearn: 0.5462437\ttotal: 8.63s\tremaining: 970ms\n",
      "899:\tlearn: 0.5461734\ttotal: 8.64s\tremaining: 960ms\n",
      "900:\tlearn: 0.5461320\ttotal: 8.66s\tremaining: 951ms\n",
      "901:\tlearn: 0.5460866\ttotal: 8.67s\tremaining: 942ms\n",
      "902:\tlearn: 0.5460245\ttotal: 8.68s\tremaining: 932ms\n",
      "903:\tlearn: 0.5459427\ttotal: 8.69s\tremaining: 923ms\n",
      "904:\tlearn: 0.5458485\ttotal: 8.7s\tremaining: 913ms\n",
      "905:\tlearn: 0.5458148\ttotal: 8.71s\tremaining: 904ms\n",
      "906:\tlearn: 0.5457601\ttotal: 8.72s\tremaining: 894ms\n",
      "907:\tlearn: 0.5457377\ttotal: 8.73s\tremaining: 884ms\n",
      "908:\tlearn: 0.5456676\ttotal: 8.73s\tremaining: 875ms\n",
      "909:\tlearn: 0.5456153\ttotal: 8.74s\tremaining: 865ms\n",
      "910:\tlearn: 0.5455384\ttotal: 8.75s\tremaining: 855ms\n",
      "911:\tlearn: 0.5455090\ttotal: 8.76s\tremaining: 845ms\n",
      "912:\tlearn: 0.5454463\ttotal: 8.77s\tremaining: 836ms\n",
      "913:\tlearn: 0.5454051\ttotal: 8.78s\tremaining: 826ms\n",
      "914:\tlearn: 0.5453690\ttotal: 8.78s\tremaining: 816ms\n",
      "915:\tlearn: 0.5453164\ttotal: 8.79s\tremaining: 806ms\n",
      "916:\tlearn: 0.5452520\ttotal: 8.8s\tremaining: 797ms\n",
      "917:\tlearn: 0.5452177\ttotal: 8.81s\tremaining: 787ms\n",
      "918:\tlearn: 0.5451207\ttotal: 8.81s\tremaining: 777ms\n",
      "919:\tlearn: 0.5450695\ttotal: 8.82s\tremaining: 767ms\n",
      "920:\tlearn: 0.5450409\ttotal: 8.83s\tremaining: 757ms\n",
      "921:\tlearn: 0.5449916\ttotal: 8.84s\tremaining: 748ms\n",
      "922:\tlearn: 0.5449417\ttotal: 8.85s\tremaining: 738ms\n",
      "923:\tlearn: 0.5447927\ttotal: 8.85s\tremaining: 728ms\n",
      "924:\tlearn: 0.5447179\ttotal: 8.86s\tremaining: 719ms\n",
      "925:\tlearn: 0.5446326\ttotal: 8.87s\tremaining: 709ms\n",
      "926:\tlearn: 0.5444988\ttotal: 8.88s\tremaining: 699ms\n",
      "927:\tlearn: 0.5444554\ttotal: 8.89s\tremaining: 690ms\n",
      "928:\tlearn: 0.5443456\ttotal: 8.9s\tremaining: 680ms\n",
      "929:\tlearn: 0.5442742\ttotal: 8.9s\tremaining: 670ms\n",
      "930:\tlearn: 0.5442332\ttotal: 8.91s\tremaining: 661ms\n",
      "931:\tlearn: 0.5441639\ttotal: 8.92s\tremaining: 651ms\n",
      "932:\tlearn: 0.5441417\ttotal: 8.93s\tremaining: 641ms\n",
      "933:\tlearn: 0.5441011\ttotal: 8.94s\tremaining: 632ms\n",
      "934:\tlearn: 0.5440289\ttotal: 8.95s\tremaining: 622ms\n",
      "935:\tlearn: 0.5439552\ttotal: 8.95s\tremaining: 612ms\n",
      "936:\tlearn: 0.5439272\ttotal: 8.96s\tremaining: 602ms\n",
      "937:\tlearn: 0.5438591\ttotal: 8.97s\tremaining: 593ms\n",
      "938:\tlearn: 0.5438146\ttotal: 8.98s\tremaining: 583ms\n",
      "939:\tlearn: 0.5437590\ttotal: 8.98s\tremaining: 573ms\n",
      "940:\tlearn: 0.5437276\ttotal: 8.99s\tremaining: 564ms\n",
      "941:\tlearn: 0.5436442\ttotal: 9s\tremaining: 554ms\n",
      "942:\tlearn: 0.5436062\ttotal: 9.01s\tremaining: 545ms\n",
      "943:\tlearn: 0.5435467\ttotal: 9.03s\tremaining: 536ms\n",
      "944:\tlearn: 0.5435075\ttotal: 9.04s\tremaining: 526ms\n",
      "945:\tlearn: 0.5434329\ttotal: 9.05s\tremaining: 517ms\n",
      "946:\tlearn: 0.5433981\ttotal: 9.06s\tremaining: 507ms\n",
      "947:\tlearn: 0.5433740\ttotal: 9.07s\tremaining: 498ms\n",
      "948:\tlearn: 0.5433392\ttotal: 9.08s\tremaining: 488ms\n",
      "949:\tlearn: 0.5432849\ttotal: 9.09s\tremaining: 478ms\n",
      "950:\tlearn: 0.5432217\ttotal: 9.1s\tremaining: 469ms\n",
      "951:\tlearn: 0.5431783\ttotal: 9.11s\tremaining: 459ms\n",
      "952:\tlearn: 0.5431514\ttotal: 9.11s\tremaining: 449ms\n",
      "953:\tlearn: 0.5430684\ttotal: 9.12s\tremaining: 440ms\n",
      "954:\tlearn: 0.5429861\ttotal: 9.13s\tremaining: 430ms\n",
      "955:\tlearn: 0.5429519\ttotal: 9.14s\tremaining: 421ms\n",
      "956:\tlearn: 0.5428630\ttotal: 9.14s\tremaining: 411ms\n",
      "957:\tlearn: 0.5428154\ttotal: 9.15s\tremaining: 401ms\n",
      "958:\tlearn: 0.5427562\ttotal: 9.16s\tremaining: 392ms\n",
      "959:\tlearn: 0.5427329\ttotal: 9.17s\tremaining: 382ms\n",
      "960:\tlearn: 0.5427030\ttotal: 9.18s\tremaining: 372ms\n",
      "961:\tlearn: 0.5426568\ttotal: 9.19s\tremaining: 363ms\n",
      "962:\tlearn: 0.5426013\ttotal: 9.19s\tremaining: 353ms\n",
      "963:\tlearn: 0.5424996\ttotal: 9.2s\tremaining: 344ms\n",
      "964:\tlearn: 0.5424242\ttotal: 9.21s\tremaining: 334ms\n",
      "965:\tlearn: 0.5423568\ttotal: 9.22s\tremaining: 324ms\n",
      "966:\tlearn: 0.5422809\ttotal: 9.22s\tremaining: 315ms\n",
      "967:\tlearn: 0.5422573\ttotal: 9.23s\tremaining: 305ms\n",
      "968:\tlearn: 0.5422068\ttotal: 9.24s\tremaining: 296ms\n",
      "969:\tlearn: 0.5421805\ttotal: 9.25s\tremaining: 286ms\n",
      "970:\tlearn: 0.5421300\ttotal: 9.26s\tremaining: 277ms\n",
      "971:\tlearn: 0.5420869\ttotal: 9.27s\tremaining: 267ms\n",
      "972:\tlearn: 0.5419878\ttotal: 9.27s\tremaining: 257ms\n",
      "973:\tlearn: 0.5419467\ttotal: 9.28s\tremaining: 248ms\n",
      "974:\tlearn: 0.5418895\ttotal: 9.29s\tremaining: 238ms\n",
      "975:\tlearn: 0.5418484\ttotal: 9.3s\tremaining: 229ms\n",
      "976:\tlearn: 0.5418131\ttotal: 9.3s\tremaining: 219ms\n",
      "977:\tlearn: 0.5417152\ttotal: 9.31s\tremaining: 210ms\n",
      "978:\tlearn: 0.5416911\ttotal: 9.32s\tremaining: 200ms\n",
      "979:\tlearn: 0.5416398\ttotal: 9.33s\tremaining: 190ms\n",
      "980:\tlearn: 0.5416143\ttotal: 9.34s\tremaining: 181ms\n",
      "981:\tlearn: 0.5415585\ttotal: 9.35s\tremaining: 171ms\n",
      "982:\tlearn: 0.5414903\ttotal: 9.35s\tremaining: 162ms\n",
      "983:\tlearn: 0.5414156\ttotal: 9.36s\tremaining: 152ms\n",
      "984:\tlearn: 0.5413507\ttotal: 9.37s\tremaining: 143ms\n",
      "985:\tlearn: 0.5412961\ttotal: 9.38s\tremaining: 133ms\n",
      "986:\tlearn: 0.5412142\ttotal: 9.39s\tremaining: 124ms\n",
      "987:\tlearn: 0.5411691\ttotal: 9.4s\tremaining: 114ms\n",
      "988:\tlearn: 0.5411366\ttotal: 9.4s\tremaining: 105ms\n",
      "989:\tlearn: 0.5410612\ttotal: 9.41s\tremaining: 95.1ms\n",
      "990:\tlearn: 0.5410476\ttotal: 9.42s\tremaining: 85.6ms\n",
      "991:\tlearn: 0.5410078\ttotal: 9.43s\tremaining: 76.1ms\n",
      "992:\tlearn: 0.5409202\ttotal: 9.44s\tremaining: 66.6ms\n",
      "993:\tlearn: 0.5408648\ttotal: 9.46s\tremaining: 57.1ms\n",
      "994:\tlearn: 0.5408149\ttotal: 9.47s\tremaining: 47.6ms\n",
      "995:\tlearn: 0.5407313\ttotal: 9.48s\tremaining: 38.1ms\n",
      "996:\tlearn: 0.5406848\ttotal: 9.49s\tremaining: 28.5ms\n",
      "997:\tlearn: 0.5406426\ttotal: 9.5s\tremaining: 19ms\n",
      "998:\tlearn: 0.5405862\ttotal: 9.51s\tremaining: 9.52ms\n",
      "999:\tlearn: 0.5405536\ttotal: 9.52s\tremaining: 0us\n"
     ]
    }
   ],
   "source": [
    "model8=CatBoostClassifier()\n",
    "model8.fit(x_train_scaled, y_train.map(int))\n",
    "y_pred8 = model8.predict(x_test_scaled)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.6666666666666666\n",
      "0.6675684815691579\n",
      "0.6590984974958264\n"
     ]
    }
   ],
   "source": [
    "print(accuracy_score(y_test, y_pred8))\n",
    "print(precision_score(y_test, y_pred8))\n",
    "print(recall_score(y_test, y_pred8))\n",
    "prob_8 = model8.predict_proba(x_test_scaled)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.67      0.67      0.67      3017\n",
      "           1       0.67      0.66      0.66      2995\n",
      "\n",
      "    accuracy                           0.67      6012\n",
      "   macro avg       0.67      0.67      0.67      6012\n",
      "weighted avg       0.67      0.67      0.67      6012\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_test, y_pred8))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Feature Importance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7fc996bb1d90>"
      ]
     },
     "execution_count": 91,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1200x800 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "pd.Series(model8.feature_importances_, index=x_train.columns).sort_values().plot.barh()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Model 9 XGBoost "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Library/Frameworks/Python.framework/Versions/3.7/lib/python3.7/site-packages/xgboost/sklearn.py:888: UserWarning: The use of label encoder in XGBClassifier is deprecated and will be removed in a future release. To remove this warning, do the following: 1) Pass option use_label_encoder=False when constructing XGBClassifier object; and 2) Encode your labels (y) as integers starting with 0, i.e. 0, 1, 2, ..., [num_class - 1].\n",
      "  warnings.warn(label_encoder_deprecation_msg, UserWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[17:45:28] WARNING: /Users/travis/build/dmlc/xgboost/src/learner.cc:1061: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.\n"
     ]
    }
   ],
   "source": [
    "model9=XGBClassifier()\n",
    "model9.fit(x_train_scaled, y_train)\n",
    "y_pred9 = model9.predict(x_test_scaled)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.656187624750499\n",
      "0.6546666666666666\n",
      "0.6557595993322204\n"
     ]
    }
   ],
   "source": [
    "print(accuracy_score(y_test, y_pred9))\n",
    "print(precision_score(y_test, y_pred9))\n",
    "print(recall_score(y_test, y_pred9))\n",
    "prob_9 = model9.predict_proba(x_test_scaled)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.66      0.66      0.66      3017\n",
      "           1       0.65      0.66      0.66      2995\n",
      "\n",
      "    accuracy                           0.66      6012\n",
      "   macro avg       0.66      0.66      0.66      6012\n",
      "weighted avg       0.66      0.66      0.66      6012\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_test, y_pred9))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Feature Importance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7fc9af5bf610>"
      ]
     },
     "execution_count": 95,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1200x800 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "pd.Series(model9.feature_importances_, index=x_train.columns).sort_values().plot.barh()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Model 10 LGBM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {},
   "outputs": [],
   "source": [
    "model10=LGBMClassifier()\n",
    "model10.fit(x_train_scaled, y_train)\n",
    "y_pred10 = model10.predict(x_test_scaled)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.664836992681304\n",
      "0.6620370370370371\n",
      "0.6684474123539232\n"
     ]
    }
   ],
   "source": [
    "print(accuracy_score(y_test, y_pred10))\n",
    "print(precision_score(y_test, y_pred10))\n",
    "print(recall_score(y_test, y_pred10))\n",
    "prob_10 = model10.predict_proba(x_test_scaled)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.67      0.66      0.66      3017\n",
      "           1       0.66      0.67      0.67      2995\n",
      "\n",
      "    accuracy                           0.66      6012\n",
      "   macro avg       0.66      0.66      0.66      6012\n",
      "weighted avg       0.66      0.66      0.66      6012\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_test, y_pred10))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Feature Importance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7fc9971dcb90>"
      ]
     },
     "execution_count": 99,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1200x800 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "pd.Series(model10.feature_importances_, index=x_train.columns).sort_values().plot.barh()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Model Comparison"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {},
   "outputs": [],
   "source": [
    "a, p, r, f = [], [], [], []\n",
    "for model in [model1,model2, model3, model4, model5, model6, model7, model8, model9, model10]:\n",
    "    a.append(accuracy_score(y_test, model.predict(x_test_scaled)))\n",
    "    p.append(precision_score(y_test, model.predict(x_test_scaled)))\n",
    "    r.append(recall_score(y_test, model.predict(x_test_scaled)))\n",
    "    f.append(f1_score(y_test, model.predict(x_test_scaled)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_score = pd.DataFrame([a,p,r,f],\n",
    "             columns = ['LR',\n",
    "                        'KNN',\n",
    "                        'Gausssian',\n",
    "                        'SVS',\n",
    "                        'DT',\n",
    "                        'RF',\n",
    "                        'AdaB',\n",
    "                        'CatB',\n",
    "                        'XGB',\n",
    "                        'LGBM']).T.sort_values(by=3, ascending=False).rename(columns={0:'Accuracy', 1: 'Precision', 2: 'Recall', 3: 'F1',}).round(4) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1000x1000 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Create a heatmap of the summary\n",
    "plt.figure(figsize= (10,10))\n",
    "sns.heatmap(model_score, annot=True, cmap = 'coolwarm')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {},
   "outputs": [],
   "source": [
    "#x_test_scaled # I dont know who is going to churn, but I want to know\n",
    "\n",
    "#x_test['prediction']=model10.predict(x_test_scaled)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "outputs": [],
   "source": [
    "#x_test.sample(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [],
   "source": [
    "#model10.predict_proba(x_test_scaled)[:,1] "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## AUC-ROC curve "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import roc_curve"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {},
   "outputs": [],
   "source": [
    "# roc curve for models\n",
    "fpr1, tpr1, thresh1 = roc_curve(y_test, prob_1[:,1], pos_label=1)\n",
    "fpr2, tpr2, thresh2 = roc_curve(y_test, prob_2[:,1], pos_label=1)\n",
    "fpr3, tpr3, thresh3 = roc_curve(y_test, prob_3[:,1], pos_label=1)\n",
    "fpr4, tpr4, thresh4 = roc_curve(y_test, prob_4[:,1], pos_label=1)\n",
    "fpr5, tpr5, thresh5 = roc_curve(y_test, prob_5[:,1], pos_label=1)\n",
    "fpr6, tpr6, thresh6 = roc_curve(y_test, prob_6[:,1], pos_label=1)\n",
    "fpr7, tpr7, thresh7 = roc_curve(y_test, prob_7[:,1], pos_label=1)\n",
    "fpr8, tpr8, thresh8 = roc_curve(y_test, prob_8[:,1], pos_label=1)\n",
    "fpr9, tpr9, thresh9 = roc_curve(y_test, prob_9[:,1], pos_label=1)\n",
    "fpr10, tpr10, thresh10 = roc_curve(y_test, prob_10[:,1], pos_label=1)\n",
    "\n",
    "# roc curve for tpr = fpr \n",
    "random_probs = [0 for i in range(len(y_test))]\n",
    "p_fpr, p_tpr, _ = roc_curve(y_test, random_probs, pos_label=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.7245614306907491 0.7245614306907491 0.6720511979141017 0.7220967107370975 0.6144523271854594 0.7255057180152757 0.7166697561896057 0.7308646661682852 0.7220688773632775 0.7296366776358564\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import roc_auc_score\n",
    "\n",
    "# auc scores\n",
    "auc_score1 = roc_auc_score(y_test, prob_1[:,1])\n",
    "auc_score2 = roc_auc_score(y_test, prob_2[:,1])\n",
    "auc_score3 = roc_auc_score(y_test, prob_3[:,1])\n",
    "auc_score4 = roc_auc_score(y_test, prob_4[:,1])\n",
    "auc_score5 = roc_auc_score(y_test, prob_5[:,1])\n",
    "auc_score6 = roc_auc_score(y_test, prob_6[:,1])\n",
    "auc_score7 = roc_auc_score(y_test, prob_7[:,1])\n",
    "auc_score8 = roc_auc_score(y_test, prob_8[:,1])\n",
    "auc_score9 = roc_auc_score(y_test, prob_9[:,1])\n",
    "auc_score10 = roc_auc_score(y_test, prob_10[:,1])\n",
    "\n",
    "print(auc_score1, auc_score2,auc_score3, auc_score4, auc_score5, auc_score6, auc_score7, auc_score8, auc_score9, auc_score10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 800x550 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# matplotlib\n",
    "import matplotlib.pyplot as plt\n",
    "plt.style.use('seaborn')\n",
    "\n",
    "# plot roc curves\n",
    "plt.plot(fpr1, tpr1, linestyle='--',color='orange', label='Logistic Regression')\n",
    "plt.plot(fpr2, tpr2, linestyle='--',color='green', label='KNN')\n",
    "plt.plot(fpr3, tpr3, linestyle='--',color='pink', label='Gausssian')\n",
    "plt.plot(fpr4, tpr4, linestyle='--',color='black', label='SVC')\n",
    "plt.plot(fpr5, tpr5, linestyle='--',color='red', label='DT')\n",
    "plt.plot(fpr6, tpr6, linestyle='--',color='yellow', label='RF')\n",
    "plt.plot(fpr7, tpr7, linestyle='--',color='deeppink', label='AdaB')\n",
    "plt.plot(fpr8, tpr8, linestyle='--',color='magenta', label='CatB')\n",
    "plt.plot(fpr9, tpr9, linestyle='--',color='violet', label='XGB')\n",
    "plt.plot(fpr10, tpr10, linestyle='--',color='firebrick', label='LGBM')\n",
    "plt.plot(p_fpr, p_tpr, linestyle='--', color='blue')\n",
    "# title\n",
    "plt.title('ROC curve')\n",
    "# x label\n",
    "plt.xlabel('False Positive Rate')\n",
    "# y label\n",
    "plt.ylabel('True Positive rate')\n",
    "\n",
    "plt.legend(loc='best')\n",
    "plt.savefig('ROC',dpi=300)\n",
    "plt.show();"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.7308646661682852\n"
     ]
    }
   ],
   "source": [
    "print(max(auc_score1, auc_score2,auc_score3, auc_score4, auc_score5, auc_score6, auc_score7, auc_score8, auc_score9, auc_score10))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "CatBoost model AUC score is the maximum score so, I will consider catBoost as a best predictor between all the other 9 models "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "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
}