model_kaggle.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Adult Income Prediction notebook\n",
    "This data was extracted from the 1994 Census bureau database by Ronny Kohavi and Barry Becker (Data Mining and Visualization, Silicon Graphics). A set of reasonably clean records was extracted using the following conditions: ((AAGE>16) && (AGI>100) && (AFNLWGT>1) && (HRSWK>0)). *The prediction task is to determine whether a person makes over  USD 50K a year.*  \n",
    "This is a binary classification problem where we have to predict whether a person earns over $50k per year or not. The scoring function to optimize is accuracy. The notebook follows the following steps to analyse the data and build predictive models.\n",
    "- Data cleaning and preprocessing\n",
    "- Exploratory data analysis\n",
    "- Modelling: I have tried out different classification algorithms.\n",
    "  - Random forest\n",
    "  - XGBoost\n",
    "  - Naive Bayes\n",
    "  - Logistic regression\n",
    "  - CatBoost  \n",
    "These models were then optimized by tuning the hyper-parameters through Grid Search, keeping a close check on the cross-validation scores to prevent overfitting. Thereafter I also tried out stacking different models together to improve the accuracy but it didn't improve the accuracy sigificantly.\n",
    "- Finally I explored oversampling techniques including SMOTE (in progress).\n",
    "\n",
    "Please feel free to suggest/comment.\n",
    "\n",
    "<img src=\"money.jpg\" style=\"width: 700px;\"/>\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import seaborn as sns\n",
    "os.environ['KMP_DUPLICATE_LIB_OK']='True'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "sns.set() ##set defaults"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = pd.read_csv(\"adult.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(32561, 15)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "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>age</th>\n",
       "      <th>workclass</th>\n",
       "      <th>fnlwgt</th>\n",
       "      <th>education</th>\n",
       "      <th>education.num</th>\n",
       "      <th>marital.status</th>\n",
       "      <th>occupation</th>\n",
       "      <th>relationship</th>\n",
       "      <th>race</th>\n",
       "      <th>sex</th>\n",
       "      <th>capital.gain</th>\n",
       "      <th>capital.loss</th>\n",
       "      <th>hours.per.week</th>\n",
       "      <th>native.country</th>\n",
       "      <th>income</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>90</td>\n",
       "      <td>?</td>\n",
       "      <td>77053</td>\n",
       "      <td>HS-grad</td>\n",
       "      <td>9</td>\n",
       "      <td>Widowed</td>\n",
       "      <td>?</td>\n",
       "      <td>Not-in-family</td>\n",
       "      <td>White</td>\n",
       "      <td>Female</td>\n",
       "      <td>0</td>\n",
       "      <td>4356</td>\n",
       "      <td>40</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>82</td>\n",
       "      <td>Private</td>\n",
       "      <td>132870</td>\n",
       "      <td>HS-grad</td>\n",
       "      <td>9</td>\n",
       "      <td>Widowed</td>\n",
       "      <td>Exec-managerial</td>\n",
       "      <td>Not-in-family</td>\n",
       "      <td>White</td>\n",
       "      <td>Female</td>\n",
       "      <td>0</td>\n",
       "      <td>4356</td>\n",
       "      <td>18</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>66</td>\n",
       "      <td>?</td>\n",
       "      <td>186061</td>\n",
       "      <td>Some-college</td>\n",
       "      <td>10</td>\n",
       "      <td>Widowed</td>\n",
       "      <td>?</td>\n",
       "      <td>Unmarried</td>\n",
       "      <td>Black</td>\n",
       "      <td>Female</td>\n",
       "      <td>0</td>\n",
       "      <td>4356</td>\n",
       "      <td>40</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>54</td>\n",
       "      <td>Private</td>\n",
       "      <td>140359</td>\n",
       "      <td>7th-8th</td>\n",
       "      <td>4</td>\n",
       "      <td>Divorced</td>\n",
       "      <td>Machine-op-inspct</td>\n",
       "      <td>Unmarried</td>\n",
       "      <td>White</td>\n",
       "      <td>Female</td>\n",
       "      <td>0</td>\n",
       "      <td>3900</td>\n",
       "      <td>40</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>41</td>\n",
       "      <td>Private</td>\n",
       "      <td>264663</td>\n",
       "      <td>Some-college</td>\n",
       "      <td>10</td>\n",
       "      <td>Separated</td>\n",
       "      <td>Prof-specialty</td>\n",
       "      <td>Own-child</td>\n",
       "      <td>White</td>\n",
       "      <td>Female</td>\n",
       "      <td>0</td>\n",
       "      <td>3900</td>\n",
       "      <td>40</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   age workclass  fnlwgt     education  education.num marital.status  \\\n",
       "0   90         ?   77053       HS-grad              9        Widowed   \n",
       "1   82   Private  132870       HS-grad              9        Widowed   \n",
       "2   66         ?  186061  Some-college             10        Widowed   \n",
       "3   54   Private  140359       7th-8th              4       Divorced   \n",
       "4   41   Private  264663  Some-college             10      Separated   \n",
       "\n",
       "          occupation   relationship   race     sex  capital.gain  \\\n",
       "0                  ?  Not-in-family  White  Female             0   \n",
       "1    Exec-managerial  Not-in-family  White  Female             0   \n",
       "2                  ?      Unmarried  Black  Female             0   \n",
       "3  Machine-op-inspct      Unmarried  White  Female             0   \n",
       "4     Prof-specialty      Own-child  White  Female             0   \n",
       "\n",
       "   capital.loss  hours.per.week native.country income  \n",
       "0          4356              40  United-States  <=50K  \n",
       "1          4356              18  United-States  <=50K  \n",
       "2          4356              40  United-States  <=50K  \n",
       "3          3900              40  United-States  <=50K  \n",
       "4          3900              40  United-States  <=50K  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 32561 entries, 0 to 32560\n",
      "Data columns (total 15 columns):\n",
      "age               32561 non-null int64\n",
      "workclass         32561 non-null object\n",
      "fnlwgt            32561 non-null int64\n",
      "education         32561 non-null object\n",
      "education.num     32561 non-null int64\n",
      "marital.status    32561 non-null object\n",
      "occupation        32561 non-null object\n",
      "relationship      32561 non-null object\n",
      "race              32561 non-null object\n",
      "sex               32561 non-null object\n",
      "capital.gain      32561 non-null int64\n",
      "capital.loss      32561 non-null int64\n",
      "hours.per.week    32561 non-null int64\n",
      "native.country    32561 non-null object\n",
      "income            32561 non-null object\n",
      "dtypes: int64(6), object(9)\n",
      "memory usage: 3.7+ MB\n"
     ]
    }
   ],
   "source": [
    "data.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Private             22696\n",
       "Self-emp-not-inc     2541\n",
       "Local-gov            2093\n",
       "?                    1836\n",
       "State-gov            1298\n",
       "Self-emp-inc         1116\n",
       "Federal-gov           960\n",
       "Without-pay            14\n",
       "Never-worked            7\n",
       "Name: workclass, dtype: int64"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data['workclass'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<=50K    24720\n",
       ">50K      7841\n",
       "Name: income, dtype: int64"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data['income'].value_counts()    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Encode the target variable to binary"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "data['income'] = data['income'].apply(lambda inc: 0 if inc == \"<=50K\" else 1) # Binary encoding of the target variable"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Exploratory analysis"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x1a23ae94a8>"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(10,5))\n",
    "sns.countplot(data['income'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As one can see, there is considerable class imbalance in the target variable, i.e. income. This is also intuitively obvious as one expects fewer 'rich' people (earning>50k/annum) than 'not-so-rich' people (earning <50k/annum). Therefore we might need to consider over-sampling techniques in our ML model to improve our accuracy."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x1a21ae5f98>"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(14,6))\n",
    "sns.countplot(data['marital.status'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Those with `Never-married` and `Married-civ-spouse` labels dominate the dataset."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Married-civ-spouse       14976\n",
       "Never-married            10683\n",
       "Divorced                  4443\n",
       "Separated                 1025\n",
       "Widowed                    993\n",
       "Married-spouse-absent      418\n",
       "Married-AF-spouse           23\n",
       "Name: marital.status, dtype: int64"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA4EAAAF8CAYAAABxOb4iAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvnQurowAAIABJREFUeJzs3XucVXW9+P/XzHATQUIiEbPUsHdaliHa5Whm3orCMrUMLc1MPZlll69dtFLLLnY5dko7JppWUnlJk9Iy75opIWqe7Pc+lqgpYIKoiIAyM78/1h7cIMysgVnsubyej8c81qzbZ733Xnuvvd/781mfT1N7ezuSJEmSpIGhudEBSJIkSZI2HJNASZIkSRpATAIlSZIkaQAxCZQkSZKkAcQkUJIkSZIGEJNASZIkSRpATAIlSZIkaQAxCZQkSZKkAcQkUJIkSZIGEJNASZIkSRpATAIlSZIkaQAZ1OgA1sNQYGdgHtDa4FgkSZIkaUNrATYH/gIsL7tTX04CdwZubnQQkiRJktRguwG3lN24LyeB8wAWLVpCW1t7o2ORJEmSpA2qubmJ0aM3hlpuVFZfTgJbAdra2k0CJUmSJA1k3bo9rvIkMCKmAicBg4EzMvPM1dZPBM4GhgD/Ag7NzCeqjkuSJEmSBqJKeweNiC2A04BdgR2BoyJi+9U2+z7w5cx8HZDAZ6uMSZIkSZIGsqqHiNgLuC4zH8/MJcAlwIGrbdMCbFL7fziwtOKYJEmSJGnAqro56HhWvUlxHrDLatt8Grg6Is4AlgBvWN+DtrauYNGix1ix4tn1LarfGzRoCKNHj6WlpS/fHipJkiSprKq/+TcD9b22NAFtHTMRsRFwLrBXZs6MiE8DPwXeWfYAY8aMeMGy+++/n4033pgRI8bT1NS0rrH3e+3t7Sxe/CTPPLOIbbbZptHhSJIkSdoAqk4CH6YYs6LDOGBu3fxrgKWZObM2fzbw1e4cYOHCp1/QO+iSJc+w2WYvprW1nVVzUK1uo41G8uiji3jsscWNDkWSJElSNzQ3N62xUqzL/SqIpd41wJ4RMTYihgMHAL+vW/8PYMuIiNr8uylGu19v1gCW4/MkSZIkDSyVJoGZ+QhwInA9cBcwvdbs88qImJSZi4DDgYsi4q/AEcCHq4xJkiRJkgaypvb2Pttccitgzpqag86f/yDjxr28IUH1RT5fkiRJUt9T1xx0a+CB0vtVFdBANH/+PP7zPz/CYYcdzAknHM/ixYv59a8v5ogjDuGwww7m8ssvBeDTnz6O3/3uCtrb2/n4x4/illtuanDkkiRJkgYKxwXoQb/73RVMnjyFKVPew0UXTef++//BrFkzmTbtZ6xYsYLjjjuaSZN24TOf+RzHH38szzzzDJtuOoZdd31Lo0OXJEmSNECYBPag179+J77ylS9yzz13s8cee3Hffcm99/4vRxxxKABLlizhgQfuZ9ddd2fy5Hdxzjk/4pe//HWDo5YkSZI0kJgE9qCJEyfxk59M5+abb+Ccc85it93eyrvf/V4+/OGPArBo0SJGjhwJwMMPP8TQoUN55JGH2XTTMY0MW5IkSVIJs2fPYsaMy5gyZX8mTpzU6HDWmfcE9qCzzvo+d9wxk/33P5CDDz6U2bNnce21f+SZZ57hmWeWcPTRhzNv3lxmz57FI488wre/fQbf/e63WLFiRaNDlyRJktSFiy+ezt///jcuvnh6o0NZL9YE9qD3vOdATjnlJH71qwsZPnxjTjzxZG644TqOOupw2tpaef/7D2HcuM35whc+y0knncyrXrU9r3xlcNFF05k69UONDl+SJElSJ5YuXbbKtK9yiAj5fEmSJEklHH/8x5g/fy7jxo3njDPOanQ4DhEhSZIkSeqaSaAkSZIkDSAmgZIkSZI0gJgESpIkSdIAYhIoSZIkSQOISaAkSZIkDSAmgZIkSZI0gAyIweJHbjKMYUMH93i5y5Y/x+Knuh4oct68uXzgA+9lq622WWX5t771PTbbbFyPxjRv3lyOO+5oLrlkRo+WK0mSJKl/GBBJ4LChg5l6woU9Xu700w9hMV0ngQAvfvFYzj9/eo/HIEmSJEndMSCSwN7q8ccX8u1vf51HH32U5uZmjj76WHbe+Q2ce+7ZPProfP71r4d44olFfOhDR3DHHX/h3nv/lwkTXskpp3yd1tZWvvvdb3L//f/k8ccfZ8KECZx88mmlypckSZI0cJkEbiALFjzG4YdPXTm/zz5vJ/PvvPOd+7HrrruzYMECPvaxj6ysLbz//n/yP/9zHvfcczef/OR/csEFv2TLLV/GoYcexD/+cR9LljzNoEGDOfvsn9DW1sYnPnEMf/7zn4jYbuUxvv/976yx/OHDN97gj1+SJElS72ASuIGsqTnoO9+5Jw8++CDTpp0NwIoVK3jkkYcB2HnnNzBo0CDGjducMWNezNZbb7OynMWLn2LixElssskoLr30Ih566AEefvhfLF26dJXyZ82aucbyt902qn64kiRJknopk8AGam1t47//+0dssskoABYsWMDo0aO56aYbGDTo+VPT0tLygn1vueVGpk07m4MOOpjJk/fjiSeeoL29vVT5kiRJkgYuh4hooJ12msSvf30xAHPm3M+HPvR+li8v19HMrFkzedvb9uKd79yPESNGcOedd9DW1tpj5UuSJEnqnwZETeCy5c8x/fRDKil3fXzqUydw+umncdhhB9Pe3s6XvnRq6fv1pkzZn1NOOZFrrvkDgwYNZocdXsvcuXPZaaeeKV+SJElS/9S0ehPCPmQrYM7ChU/T1rbqY5g//0HGjXt5Q4Lqi3y+JEmSpK4df/zHmD9/LuPGjeeMM85qdDg0NzcxZswIgK2BB0rvV1VAkiRJkqTexyRQkiRJkgYQk0BJkiRJGkBMAiVJkiRpADEJlCRJkqQBxCRQkiRJkgaQATFO4OhRQxg0ZGiPl7vi2eUsevLZLrebN28uBx20H/vttz8nnHDiyuX33Zd8+MOH8MUvfoXJk6escd8DD5zCD35wNptvPr7H4pYkSZI0cA2IJHDQkKHccfqRPV7uTidMA7pOAgFGjRrF7bf/mdbWVlpaWgC49to/8qIXje7xuCRJkiRpbSpPAiNiKnASMBg4IzPPrFu3I3B+3eZjgUWZ+Zqq49rQNtpoONtu+0ruvvtOJk6cBMDMmbcxadIuAFx66a/4/e+vZNmypQwePJiTTz6Nl71sq5X7t7a2ctZZ3+fOO++gtbWNyZPfxfvff0gjHookSZKkPqzSewIjYgvgNGBXYEfgqIjYvmN9Zt6VmTtm5o7Am4FFwDFVxtRIe+yxN9dffy0Af//735gwYVsGDx7MkiVLuOmmG/nhD8/mZz+7iDe/eTcuvfSiVfadMeMyAM4770LOOecCbr75Ru6++84N/hgkSZIk9W1V1wTuBVyXmY8DRMQlwIHAqWvY9gvAjZl5S8UxNcyuu76Fc875EW1tbVx77R9529v25tprr2bjjTfm5JO/xjXXXM2//vUQt99+K9tuG6vsO2vWTO677/+4445ZACxd+gz//Oc/eN3rXt+IhyJJkiSpj6o6CRwPzKubnwfssvpGETEKOArYoeJ4Gmr48OFMmLAtf/3rXcye/ReOOebjXHvt1fz7349y9NEf5oAD3scb3/hmNt10DPfdl6vs29raxsc+9gl23/1tADzxxBNstNFGjXgYkiRJkvqwqpPAZqC9br4JaFvDdocCl2fmv7t7gDFjRrxg2b//3cygQRtm9Isyx2lpaV657d5778PZZ/+Q7bbbnmHDhtDU1MTw4cPZcsstOeSQD7Js2TLOPfdsNttss5Vlt7Q0s/POuzBjxuXsvvvuPPvscxx77JGccMIX2WmnSev9GJqbmxk7duR6lyNJkiT1Zy0tTSunffn7c9VJ4MPAbnXz44C5a9juPcDX1+UACxc+TVtb+yrL2traWLFiTblmzytznNbWtpXbvvGNu3LaaafwkY8cw4oVbbS3t9PS0kJraxsHH3wA7e3t7LjjRO6//58ry25tbWO//d7LQw89yAc/+AFaW1uZPHkKr3vdxB55nG1tbTz22OL1LkeSJEnqz1pb21dOe8P35+bmpjVWinWlqb29veut1lGtY5hbKJqALgFuBY7KzJl12zQBC4HxmbmsG8VvBcxZUxI4f/6DjBv38pXzjR4nsLdb/fmSJEmS9ELHH/8x5s+fy7hx4znjjLMaHU59Erg18EDZ/SqtCczMRyLiROB6YAgwLTNnRsSVwJczcxbFsBDPdjMB7JYiUev7yZokSZIkra/KxwnMzOnA9NWWTa77/98UzUQlSZIkSRXbML2nSJIkSZJ6hX6bBFZ5r2N/4vMkSZIkDSz9MgkcNGgIS5Y8ZYLThfb2dpYseYpBg4Y0OhRJkiRJG0jl9wQ2wujRY1m06DGefvqJRofS6w0aNITRo8c2OgxJkiRJG0i/TAJbWgbx4hdv3ugwJEmSJKnX6ZfNQSVJkiRJa2YSKEmSJEkDiEmgJEmSJA0gJoGSJEmSNICYBEqSJEnSALLOSWBEjOrJQCRJkiRJ1SuVBEbEN1ab3xu4p5KIJEmSJEmVKTtO4Jsj4rPAD4BvAwcCx1YWlSRJkiSpEmWbg04G3gMksBmwQ2ZeVllUkiRJkqRKdJoERsTEiJgIBPBlYDBwPfDy2nJJkiRJUh/SVXPQS1ebfxb4XO3/dmCbHo9IkiRJklSZTpPAzNx6QwUiSZIkqf+bPXsWM2ZcxpQp+zNx4qRGhzMgddkxTEQcCBwKvBRoBR4GLsvMn1ccmyRJkqR+5uKLpzNnzv0sW7bUJLBBOk0CI+IrwB7ATyiSvyZgPHBEROyQmZ/rbH9JkiRJqrd06bJVptrwuqoJPAR4TWY+W78wIn4F/JXn7w+UJEmSJPUBXQ0R0bqWbVqA53o+HEmSJElSlbqqCfwp8KeImA78i6JH0PEUNYTeEyhJkiRJfUxXvYN+IyLuAvYH9qKoFXwIODUzf7sB4pMkSZIk9aAuewfNzKuAqwAi4uXAc5k5t+rAJEmSJEk9r9N7AiPiF7XplhFxB3A38PeIuDkiXrohApQkSZIk9ZyuOoaJ2vS/gF9k5ouAFwG/AM6tMjBJkiRJUs/rKgns8MrM/A5AZrZn5lkUg8dLkiRJkvqQrpLAMRHxRuAfEbFdx8LavYFtlUYmSZIkSepxXXUMcwnwPeD1wBDgXRFxKPB94LiKY5MkSZIk9bCuhoj4DEBEDKEYHxCKzmHekpl/qzg2SZIkSVIP63KIiIjYKDOXAg9ExDhgAmACKEmSJEl90FqTwIgYClwJzAb+X0S8BriRoiZw+4j4fGae39UBImIqcBIwGDgjM89cbX0AZwOjgfnAwZm5aN0ejiRJkiSpM511DDMVeAz4XG3+c8Apmfk2YKe65WsVEVsApwG7AjsCR0XE9nXrm4ArgG9m5uuAO4HPr8PjkCRJkiSV0Flz0E8ATwDTasna/gARcR7QBIyLiPMy84hOytgLuC4zH6/tewlwIHBqbf1EYElm/r42/3WKcQglSZIkSRXoLAn8LjAF+ChF4rZ1Zn4QICJ2Ad7cRQIIRWcy8+rm5wG71M1PAOZHxLkUPZD+HXsdlSRJkqTKdJYE/go4CFgKPAXsDRARn6SosTusRPnNQHvdfBOrji84CHgrRW+jsyLiqxRDUhxeLnwYM2ZE2U0lSZIkNVhLS9PK6dixIxscTff05djrrTUJzMzngHdHxIuAZzLz2dqqS4FfZeb8EuU/DOxWNz8OmFs3Px+4LzNn1eZ/QTE2YWkLFz5NW1t71xtKkiRJarjW1vaV08ceW9zgaLqnt8Xe3Ny0TpViXQ4RkZlPrDb/cDfKvwY4OSLGAkuAA4Cj6tbfCoyNiNdl5t0UzU/v6Eb5kiRJkqRu6Kx30PWWmY8AJwLXA3cB0zNzZkRcGRGTauMP7g+cExF/A94GfKbKmCRJkiRpIOuyJnB9ZeZ0YPpqyybX/X87q3YWI0mSJEmqSOmawIh4Sa1ZpyRJkiSpj+pOTeAvgOeAt1cUiyRJkiSpYqVqAiNiAjAKGB0R21QbkiRJkiSpKmWbgx4JnAf8hFV795QkSZIk9SFdJoERMQh4H/Bzig5e3ldbJkmSJEnqY8rUBO4P/Ckzn8rMp4AbKcb7kyRJkiT1MWWSwOeAb9bNn15bJkmSJEnqY7ps1pmZl682/3fg75VFJEmSJEmqTOlxAiVJkiRJfZ9JoCRJkiQNICaBkiRJkjSAlBrqISLGAecC2wK7AT8FDs/MeRXGJkmSJEnqYWVrAs8CLgeWAo8DdwHTqgpKkiRJklSNskngVpl5DtCWmc9l5ueAl1UYlyRJkiSpAmWTwLaIWLltRIzsxr6SJEmSpF6ibCL3a+BCYFREHA1cB1xUWVSSJEmSpEqUSgIz8+vAVcBfgL2BHwOnVhiXJEmSJKkCpXoHBcjMn1L0CipJkiRJ6qPKDhHxfoqav9H1yzPzJVUEJUmSJEmqRtmawNOB44B/VhiLJEmSJKliZZPABzLzikojkSRJkiRVrmwSeEFEfBv4PfBcx8LMvKmSqCRJkiRJlSibBO4BTAbeXresHXhtj0ckSZIkSapM2SRwIrBFZi6rMhhJkiRJUrXKDhY/n24MJyFJkiRJ6p3KJnaPAHdHxDXA8o6FmfmJSqKSJEmSJFWibBL4TxweQpIkSZL6vFJJYGaeEhEjgJ2AwcDtmbm40sgkSZIkST2u1D2BEbEz8H/AGcD3gAcj4s1VBiZJkiRJ6nllO4b5LnBIZr4+M18LHEiRDEqSJEmS+pCySeDIzLy+YyYzrwOGVxOSJEmSJKkqZTuGaY+Il2fmgwARsRXQWmbHiJgKnERxL+EZmXnmauu/AhwBLKotOmf1bSRJkiRJPaNsEngqcFttiIh2YF/gY13tFBFbAKdRdCizHLg1Iq7PzHvrNpsEHJyZf+5W5JIkSZKkbivVHDQzLwfeCtwKzAR2z8xLS+y6F3BdZj6emUuASyjuJ6w3CfhiRPw1In4YEcNKRy9JkiRJ6payvYO+FvheZv4IuBH4ZUREiV3HA/Pq5ucBL60rdwRwJ/D/gInAi4AvlQtdkiRJktRdZZuD/gj4DkBm3hMRJwNnU9QOdqaZovlohyagrWMmM58GJnfMR8R3gfOAE0vGxZgxI8puKkmSJKnBWlqaVk7Hjh3Z4Gi6py/HXq9sErhxZl7WMZOZl0fEl0vs9zCwW938OGBux0xEvAzYKzPPqy1qAp4rGRMACxc+TVtbe9cbSpIkSWq41tb2ldPHHlvc4Gi6p7fF3tzctE6VYmWHiGivNQkFICK2o1zvoNcAe0bE2IgYDhwA/L5u/VLg9IjYOiKagGOBy9ZQjiRJkiSpB5StCfwScGNE3FObfxVwSFc7ZeYjEXEicD0wBJiWmTMj4krgy5k5KyKOBmbU1t9CMTC9JEmSJKkCpZLAzPxtrSOY/wBWALdn5r9L7jsdmL7assl1/18KlOlpVJIkSZK0nso2BwXYrjYdDOwaEe+tIB5JkiRJUoVK1QRGxDnAO4B/8Hxvn+3AryuKS5IkSZJUgbL3BO4JbJeZje8CR5IkDXizZ89ixozLmDJlfyZOnNTocCSpTymbBP7LBFCSJPUWF188nTlz7mfZsqUmgZLUTWWTwD9FxC8pevFc2rEwM20OKkmSNrilS5etMpUklVc2CXxTbXpk3TLvCZQkSZKkPqbsEBF7VB2IJEmSJKl6nSaBEXFGZh4fETN4vlfQlTJzv8oikyRJkiT1uK5qAq+tTS+pOhBJkiRJUvU6TQIzc0ZtesGGCUeSJEmSVKXmRgcgSZIkSdpwTAIlSZIkaQAxCZQkSdrAZs+exSmnnMjs2bMaHYqkAajUEBEREcBngZcATR3L7R1UkiStzchNhjFs6OBKym5paVo5HTt2ZCXHWLb8ORY/Vc1g9BdfPJ05c+5n2bKlTJw4qZJjSNLalB0sfjpwM3AZaxgqQpIkaXXDhg5m6gkXVlL2ggWLAZi/YHFlx5h++iEsppokcOnSZatMJWlDKpsEDs7M4yuNRJIkSZJUubJJ4EMRsXVmzqk0GkmSpF6ibcVzlTU13RDNWVc8u5xFTz5bSdmS+rZOk8CImEHR/HMcMCsiZgLPdaz3nkBJktRfNQ8azB2nH1lJ2csXPbpyWtUxdjphGmASqIFn9KghDBoytJKy+8sPOF3VBF5S6dElSZIkqQcNGjLUH3C60GkSmJkXAETEVzPzS/XrIuL7wAUVxiZJkiRJ6mFdNQc9BRgNvD8iRtWtGgzsC3yywtgkSZL6paGDmleZStKG1NWV53ZgIdBWm3b8PQwcUm1okiRJa9bUMniVaV+zz4TRbDN6GPtMGN3oUCQNQF01B70SuDIirsrMmRsoJkmSpE6NGD+RJfPvYeNxOzQ6lHWy3djhbDd2eKPDkDRAddUc9Iza+IBfiogXDBJv76CSJKkRho7akqGjtmx0GFK/NXKTYQwbWk1N+4boYVOd66p30GtrU3sJlSRJkgaIYUMHM/WECyspe8GCxQDMX7C4kmNMP9271rrSVXPQGbV/30GRCF6Zmc9UHpUkSZIkqRJlu6SaARwE3B8Rl0bE1Iiw7laSJEmS+phSSWBmXpiZ7we2BC4FvgH8u8rAJEmSJEk9r6t7AgGIiN2BvWt/44HrgD9UGJckSZIkqQKlkkDgemAe8DXgnMxcUV1IkiRJkqSqlL0ncAvgJGB34P8i4jcRcWx1YUmSJEmSqlD2nsB5mfkT4DjgVGBbilrBLtU6kbk3Iu7rLHGMiHdGxJwyZUqSJEmS1k2pJDAiTo2ImcDdwJuBzwAvKbHfFsBpwK7AjsBREbH9GrbbDPgO0FQ+dEmSJElSd5W9J3AkReJ3S2a2d6P8vYDrMvNxgIi4BDiQojax3jTgFOCb3ShbkiRJktRNpZLAzPzUOpY/nqJDmQ7zgF3qN4iITwCzgdvW8RiSJEmSpJLK1gSuq2agvuawCWjrmImI1wAHAHsCL12XA4wZM2J94pMkSeq3xo4d2egQJK2Dqt+7VSeBDwO71c2PA+bWzR8EbA7MAoYA4yPi5sys36dTCxc+TVtbd1qoSpKkDcEEpPEee2xxo0NQH+X7t7HKvnebm5vWqVKsdBIYES8FXksxSPwWmflQid2uAU6OiLHAEopav6M6VmbmV4Cv1MrfCrihOwmgJEmSJKl7yvYO+k7gVuBMil5B742Id3e1X2Y+ApxIMdj8XcD0zJwZEVdGxKR1D1uSJEmStC7K1gR+GXgDcGVmzouIXYELgN90tWNmTgemr7Zs8hq2ewDYqmQ8kiRJkqR1UKomEGjJzJW9fGbmXaza4YskSZIkqQ8omwQ+ExEvo5b4RcRuwLLKopIkSZIkVaJsc9DPA1cDm0fEn4FtKTp5kSRJkiT1IWUHi781It4IvAloAW7LzAWVRiZJkiRJ6nGdJoER8d61rHpLRJCZv64gJkmSJElSRbqqCTyuk3XtgEmgJEmSJPUhnSaBmbnHhgpEkiRJklS9UvcERsQ44L+BfYFW4ArgU5m5qMLYJEmSJEk9rOwQERcA/wReTzFo/ALgx1UFJUmSJEmqRtkhIl6amfvWzX82Iu6tIiBJkiRJUnXK1gQ+GBGv6JiJiM2BudWEJEmSJEmqStmawDbgzoi4GlgB7Ak8HBFXAGTmfhXFJ0mSJEnqQWWTwItrfx2uqiAWSZIkSVLFSiWBmXlBRLwceCswGLghM/9RZWCSJEmSpJ5X6p7AiNgXmAW8B9gP+EtEvLvKwCRJkiRJPa9sc9CvArtn5r0AEfFq4OfAb6oKTJIkSZLU88r2DjqkIwEEyMy/AS3VhCRJkiRJqkrZJHBpREzqmKn9/0w1IUmSJEmSqlK2OegJwG8j4j6gHXgVcFBlUUmSJEmSKlG2d9CbI2J74A0UzUD/nJkLK41MkiRJktTjyvYO2gwcARwM3AQcFRHeEyhJkiRJfUzZ5qDfBsYCO9fm3w5sDnyiiqAkSZIkSdUo2zHMnsDhwLLMfArYB9i7qqAkSZIkSdUomwQ+l5ltHTOZuRxYUU1IkiRJkqSqlG0O+r8RcSzQEhEBfBq4q7qwJEmSJElVKFsT+ElgIrAZ8CdgBHB8VUFJkiRJkqpRdoiIp4CPVByLJEmSJKlipZLAiBgH/DewL9AKXAF8KjMXVRibJEmSJKmHlW0OegHwT+D1FAPGLwB+XFVQkiRJkqRqlO0Y5qWZuW/d/Gcj4t4qApIkSZIkVadsTeCDEfGKjpmI2ByYW01IkiRJkqSqlK0JbAPujIirKcYH3BN4OCKuAMjM/da2Y0RMBU4CBgNnZOaZq63fHzgFaAH+AhyVmc9294FIkiRJkrpWNgm8uPbX4aoyO0XEFsBpwE7AcuDWiLg+M++trd8Y+CEwMTMfjYhfAofj/YaSJEmSVImyQ0RcsI7l7wVcl5mPA0TEJcCBwKm1cpdExFaZ+VxEDAdeAtjjqCRJkiRVpGxN4LoaD8yrm58H7FK/QS0BfAfwc+AR4OruHGDMmBHrG6OA2267jYsuuoj3ve99vPGNb2x0OJIkqQeMHTuy0SFIWgdVv3erTgKbgfa6+SaK+wtXkZlXAWMi4uvAj4CpZQ+wcOHTtLW1d72hOjVt2rnMmXM/Tz21mFe84tWNDkeS1A+YgDTeY48tbnQI6qN8/zZW2fduc3PTOlWKddo7aERMq00P6nbJhYeBzevmx1HXq2hEbBoR+9StvxB47ToeS+th6dJlq0wlSZIk9U9d1QTuHRFvBk6NiH9Q1OStlJmzu9j/GuDkiBgLLAEOAI6qW98E/DwiJmXmQ8BBwC3deQCSJEmSpPK6SgJ/DPwMeCnw69XWtQPbdLZzZj4SEScC1wNDgGmZOTMirgS+nJmzIuIo4LcR0Q7cCxyzDo9DkiRJklRCp0lgZp4GnBYRv8zMg9flAJk5HZi+2rLJdf9fDly+LmVLkiRJkrqn7BARB0fE+4B3UAz6fnVm/rTSyCRJkiRJPa7TjmE6RMRngC8CdwOzgU9HxElVBiZJkiRJ6nllh4g4DNg1M58CiIhzgduAr1UVmCRJkiSp55WqCQToSABr/z8JPFdJRJIkSZKkypStCXwgIj4JnFWbPxZ4qJqQJEn4bzZ4AAAgAElEQVSSJElVKVsT+J/A/sAztb8DgI9VFZQkSZIkqRplewd9BHhrRAwHmjPz6WrDkiRJkiRVoWxzUAAy85mqApEkSZIkVa90xzCSJEmSpL7PJFCSJEmSBpDSzUEjYndgU6CpY1lm/rqKoCRJkiRJ1SiVBEbEOcA7gH8A7bXF7YBJoCRJkiT1IWVrAvcEtsvMxVUGI0mSJEmqVtl7Av9lAihJkiRJfV/ZmsA/RcQvgRnA0o6F3hMoSZIkSX1L2STwTbXpkXXLvCdQkiRJkvqYUklgZu4BEBGDgKbMfK7SqCRJkiRJlSh1T2BEvCQirgKWAMsi4rqIGF9taJIkSZKknla2Y5gfArcBmwEvAW4GflRVUJIkSZKkapS9J/CVmfm+uvmvRMTfqghIkiRJklSdsjWBgyNiWMdMRAzn+UHjJUmSJEl9RNmawF8C10TETyiSvyOASyqLSpIkSZJUibK9g341Ih4G3g60AOcD51YYl9Zg9KghDBoytJKyW1qaVk7Hjh1ZyTFWPLucRU8+W0nZkiRJksrpNAmMiE0y86mI2BT4Te2vw2jg8SqD06oGDRnKHacf2fWG62D5okdXTqs6xk4nTANMAiVJkqRG6qom8AZgIrCAVe8BbKrNt1QTliRJkiSpCp0mgZk5sWO7zGyrX1erHZQkSZIk9SFlewedtYZlN/VkIJIkSZKk6nV1T+C1wM7A8Ih4qm5VC/CXKgOTJEmSJPW8ru4J3B/YFDgP+HDd8hXAvKqCkiRJkiRVo9PmoJn5VGY+ALwHmJqZD9ZW/T9go4pjkyRJkiT1sLKDxf8EmFP7/wmKnkHPAaZ2tWNETAVOAgYDZ2TmmautfzdwCkWPo3OAD2fmopJxSZIkSZK6oWzHMNtm5mcBMvPJzPwU8OqudoqILYDTgF2BHYGjImL7uvWbAD8C3pmZrwP+CpzcrUcgSZIkSSqtbBI4uJawARARIyhq7rqyF3BdZj6emUuAS4AD68sFjs3MR2rzfwVeVjImSZIkSVI3lW0O+lPg9oi4mKIp6Hspmoh2ZTyrdiAzD9ilYyYzFwKXAUTERsDngR+UjEmSJEmS1E2lksDM/EZE/A3Yk6Jn0BMy86oSuzZTJI0dmoC21TeKiFEUyeDdmXlBmZg6jBkzojubq8HGjh3Z6BAkSRow/NyV+qaq37tlawLJzCuAKwAioikits3M+7rY7WFgt7r5ccDc+g0iYnPgD8B1wKfKxtNh4cKnaWtr73rDfqA/XMgfe2xxo0OQJG0g/eFzq6/zc1fryvdvY5V97zY3N61TpVipJDAijgFOBzauj40iqevMNcDJETEWWAIcABxVV24LMAO4KDO/1o24JUmSJEnroGxN4OeAvYETKYZ7mAK8tKudMvORiDgRuB4YAkzLzJkRcSXwZWBLYCIwKCI6OoyZlZlHdu9hSJIkSZLKKJsEPp6Zt0fEXcBmmXlaRNxbZsfMnA5MX23Z5Nq/syjfQ6kkSZIkaT2VTcCei4jRwH0837tnSzUhSZIkSZKqUrYm8MfAbymagd4VEfsDf68sKkmSJElSJcrWBN4F7JOZjwNvAr4KfKCyqCRJkiRJlShbE3hhZm4HRWcvwCPVhSSpJ82ePYsZMy5jypT9mThxUqPDkSRJUoOVTQL/GhFTgVuApzsW1moG1Q8MHdS8ylT9x8UXT2fOnPtZtmypSaAkSZJKJ4HvBg5abVk7dg7Tb+wzYTQ3znmS3bce1ehQ1MOWLl22ylSSJEkDW6kkMDOHVR2IGmu7scPZbuzwRochSZIkqWKdtv2LiB/X/f/i6sORJEmSJFWpqxvA6m8gurrKQCRJkiRJ1esqCWxay/+SJEmSpD6oO11BtlcWhSRpncyePYtTTjmR2bNnNToUSZLUR3TVMUxzRIymqAVsqfsfcIgIqSeNHjWEQUOG9ni5LS1NK6djx47s8fIBVjy7nEVPPltJ2eqcQ4BIkqTu6ioJ3AFYwPOJ38K6dQ4RIfWgQUOGcsfpR/Z4ucsXPbpyWkX5ADudMA0wCWwEhwCRJEnd1WkSmJmOHC5JkiRJ/YhJniRJktSDvF9bvV2pweIlSZIkleP92urtrAmUJEmSepD3a6u3MwmUJEmSpAHEJFDq54YOal5lKkmS1EhNLYNXmWrD81uh1M/tM2E024wexj4TRjc6FEmSJEaMn8jgEeMYMX5io0MZsOwYRurnths7nO3GDm90GJIkSQAMHbUlQ0dt2egwBjRrAiVJkiRpADEJlCRJkqQBxCRQkiRJkgYQk0BJkiRJGkBMAiVJkiRpALF3UEmq2OhRQxg0ZGglZbe0NK2cjh07spJjrHh2OYuefLaSsiVJ0oZnEihJFRs0ZCh3nH5kJWUvX/ToymlVx9jphGmASaAkSf2FzUElSZIkaQAxCZQkSZKkAcQkUJIkSZIGkMrvCYyIqcBJwGDgjMw8cy3b/RS4LjPPrzomSZIkSequoYOaV5n2VZUmgRGxBXAasBOwHLg1Iq7PzHvrthkPnA3sCVxXZTySJEkS2HOz1s0+E0Zz45wn2X3rUY0OZb1UXRO4F0Xt3uMAEXEJcCBwat02hwC/ARZWHIskSZIE2HOz1s12Y4ez3djhjQ5jvVWdBI4H5tXNzwN2qd8gM78NEBG7VhyLJEmSJA14VSeBzUB73XwT0NaTBxgzZkRPFqeKVdUkQo3nue3fPL9S3+R7t3/z/PZfVZ/bqpPAh4Hd6ubHAXN78gALFz5NW1t71xv2A/3hjf7YY4sbHUKv1dfPr+d27fr6uQXPr9ZNf3jt93W+d9euP7w+qzy//eH56cvKntvm5qZ1qhSrOgm8Bjg5IsYCS4ADgKMqPqYkSZIkaS0q7ds0Mx8BTgSuB+4CpmfmzIi4MiImVXlsSZIkSdILVT5OYGZOB6avtmzyGrY7vOpYJKm/6S/jFUmSpA3Hbw2S1IftM2E024wexj4TRjc6FEmS1EdUXhMoSapOfxmvSJLqzZ49ixkzLmPKlP2ZONE7iKSeZhIoSZKkXuXii6czZ879LFu21CRQqoDNQSVJktSrLF26bJWppJ5lEiiVMHv2LE455URmz57V6FAkSZKk9WJzUKkEm6VIkiSpv7AmsIdYU9S/2SxFkiRJ/YU1gT3EmiJJkiSBY7iq9zMJ7CHWFDXeyE2GMWzo4ErKbmlpWjkdO3ZkJceQJEn9wz4TRnPjnCfZfetRjQ5FWiOTQPUbw4YOZuoJF1ZS9oIFiwGYv2BxZceYfvohlZQrSZI2LMdwVW9nHbUkSb2U95tLkqpgTaAkSb2U95tLkqpgTaBUQlPL4FWmkrQheL+5JKkKA6om0I5DtK5GjJ/Ikvn3sPG4HRodiqReZvSoIQwaMrSSsjfEZ8uKZ5ez6MlnKylbktQ7Dagk0I5DtK6GjtqSoaO2bHQYknqhQUOGcsfpR1ZS9vJFj66cVnWMnU6YBpgEqvv8cV3quwZUEihJkqSe4Y/rUt/lPYGSJPVSDjgtSaqCnyqSJPVS+0wYzTajh7HPhNGNDkWS1I+YBEoa0ByHrX/r6+d3u7HDOWaXzR10WpLUo7wnsIc4hIDUNzkOW//m+ZUk6YWsCewhI8ZPZPCIcYwYP7HRoUjqBsdh6988v5IkvZA1gT3EIQQkad3YzbwkSRuWSaAkqaHsZl6SpA3L5qCSpH7L+7UlSXohawIl9Xo2F9S6GjF+Ikvm38PG43ZodCiSJPUaJoGSej2bC2pdeb+2JEkvZHNQSZIk9So25ZaqZRIoSZKkXsWht6Rq2RxUkiRJvYpNuaVqWRMoSZIkSQOISaCkAc37TiRJ0kBTeXPQiJgKnAQMBs7IzDNXW78jMA3YBLgJOCYzV1QdlySBQwhIkqSBp9KawIjYAjgN2BXYETgqIrZfbbOfAx/PzFcCTcBHq4xJkuoNHbUlm8Zk7z2RJEkDRtXNQfcCrsvMxzNzCXAJcGDHyoh4ObBRZt5WW3Q+cFDFMUmSJEnSgFV1c9DxwLy6+XnALl2sf2nJslsAmpubuhXQi0dv3K3te5shm4xpdAjrpbvnq7s8v43jue1cXz634Pntiud37Ty3jeV7t3Oe38715fM7UM5t3XYt3Sm/qb29vZshlRcRJwLDMvNLtfmPAjtl5jG1+f8AvpmZu9XmtwVmZOarShS/K3BzNZFLkiRJUp+xG3BL2Y2rrgl8mCKgDuOAuaut37yT9Z35S63seUDresQoSZIkSX1RC0U+9Zfu7FR1EngNcHJEjAWWAAcAR3WszMwHI2JZRPxHZv4J+CBwVcmyl9ONbFeSJEmS+qF/dneHSjuGycxHgBOB64G7gOmZOTMiroyISbXNDgH+KyL+P2AE8N9VxiRJkiRJA1ml9wRKkiRJknqXqoeIkCRJkiT1IiaBkiRJkjSAmARKkiRJ0gBiEihJkiRJA4hJoCRJkiQNIFWPEzjgRMQfgTMz8/La/HeAY4BNM/PZ2rK5FENmHJmZc1fb/wbg5My8oeI4DwfempmHV3mc/ioitgL+D7i3tmgj4Fbg88CWwDGZeWRjolsZ3w2ZuVWjYuhLIuJA4AsU18Rm4KeZ+e0GxPEu4JWZ+b1u7HMDG+Ca0dvUXuNzgH0y8491yx+guLY90JDAekBEjAemZebkbuzzAL34cdedrx9n5tF1y3cE7gQ+nJnnr0O5+wGTMvPL3Yij9LWxu+X3N331vPV2VV+3I+IU4JrMvLkHytqKPvIaiIhZwLzMnFK37HDge8BDdZs+mpn7djfm/sYksOddB/wHcHltfi/gz8CuwHURMQFY0p0Pd/VaczNzR4CIaAK+DlySmbsBDUsA1T0RsQXwXWBiZi6MiBHAjRGRmXnFBg5nUtebqM5zwDkRsUNmLm50MD2l9uNgf/yMWAi8PSJaMrO1tuz9wGPrWmDtPVrZ+7Tq8vuIPnfexO4UY3T3lF7/GoiI1wLLgddFxJaZ+a+61VdY6fFCJoE971rgDFj5a+5y4BJgX4oEcTfg6o5fbYF5wDSKL38PAC/uKCgivggcCrQCVwMnUCSXZ2XmVRHxdeD1mfmOiNgc+GNmviYiPgQcT1GjcQdwbGYui4gPAicBTwEPAk9X9iwMMJnZHhFfAR6NiE8A7wU+AVyYmTsARMQUitrfd6/l3G4J/B5YACwFpgBnUvyA8Bzw1cz8VUTsDPwXMLy27dGZOSciXg+cWwvp7g3xuPuJFwODKZ7PhZn5dEQcBizr5Lm+gaI2/y3AMOD4zLw6Il4D/AAYAbwE+EZm/k9EnAy8EXhZbf29wGm1cl8EfAq4j6LVABHxIHAxxfl/DdACfCszfxERQ1nLNWMAmgv8kSKJP6p+RUR8HngfxXP3B+Bzte0eyczv1ra5FPg5RS3+2RTvwTbgC5l5zernLTN/tNox5lNck98AzAfOo3jfvxQ4PDNvjIjdWe1cZ+ZvIuJ8YAwwgeL9/wPgdmBH4IPARZm5VURstpbYNq3FviXF62nYej2TG8bTPP++6fiCug9wDUBEfJzisW8MPAt8IDOz9nlZ/9z8lOevkxdSa9WyvtfGiNiL4jXSTPEZOZXiWv5W4NfARztqGCLiOGBCZn5ytTKmUpzPVorak0MpXkNforiObw3MpPgsWB4RHwY+A7RTfF5/vHYNas/MplqZh9c9xu8Ae1O8Fi7PzFNqP1y94Fqx9tPQbb39vK3pOTkZeDmwHcU18uzM/HZENFN8R9uT4jn/WWZ+KyLeSlEz99ZamecDN1Cc918A42qHOyUzr6j9oP8jivfwM8BxmXnnGmI7iOL8bgQMBY7IzFtrq4+KiP+q/f+pzLwhIvYETq/Ftqj2XC7o5HvdPIrvmLsCKyiuebtRfD5Mi4j9M/OetT133dCrXwM1H6b4PBgDfBToVu19RGzCms/1Daz5836zWmwvo3juv5iZv6+99sjMk2vlPkBxDdkE+DFF7rWMogb1voh4O3AqxfeQORTXmYXdiX1deU9gz7sDeEVEDKNI/K6u/XVUO7+F4kXa4TiAzNyO4svDKwAi4h3AfhRv5NdTfFE4BvgdxcULijf6dhHRArwd+F1EvJrixf/mWi3Vv4HP1hLS02vHfxMwsscf+QBXa+57H8WXQTLzr0BbLTEAOBj4eSfnFiCAQzNzb4rXxgiKD7G9gC9HxBCKBGBqZk6k+MJyTm3fnwKfqy2/v8rH2p9k5t3Ab4D7I2JmRHyL4ovUQ6z9uQbYpLZ8KnBB7dwcCXwtM3cG9gDqm5QOy8zta4nEcRRfAifW7XMv8D/A/2TmTyh+sLkjM3eieN+eGBHbsJZrxgD2GWDfiNi7btnbgZ2AnSneY1sAhwA/Az4AEBEjKa6FvwO+D5xXe673A86urYdVz9vqNgOuyszXU3w52L/WEuBkii9ssIZzXbf/wszcLjNn1OavysyguG53WFtspwKzaz8ynVmLpS+4CDgQoPbF768UXxo3Ad5D8aXwNcBvgY/X7Vf/3NRfJ6mVtV7XxtqPKxcCh9We03uAw+qPD+wUEaNr8wdTJOGr+xpFE+WdKL7Qvaq2/M3AJ2vzw4BjI2IH4ERg99oxlwBfWVN8tRhfDrwjM19H0eJo+9p3jbVdK3pSbz1va3tOoLgG7FWbHh0REyk+a7cEXgvsAhwQEe/s5HHvDzxQe24/QvG9C+AC4IRabEcBv1xDbM21472rFt/pFLcddHi6du04jOK7wVCKc3lMZk6i+K44cW3f62pljAOurZVzE8WPCD8FZlFcd3oiAezQK18DtTIGU1zjLwJ+BXwkIuoruvaLiLvq/vZYQzFrO9ew5s/7HwDXZeZra8/LebXEcG0+BXy3dm7PAd4YEWOBbwL71s7hH4BvdVJGjzIJ7GFZVJPfRvEFf1/g6sycAwyvfXi8iaJGsMNbKV60ZOZ9FL9IQ5Ho/SIzn8nMFRS/MO9JLQms+4JyNzAReAfFG28PYFvgtoi4C3g3xYfOm4FbM/PRWnlr+vDS+mun+IWrw8+BgyNiI4rmGTNY+7kF+Hc+f0/P7hQ1iW2ZOT8zXw28kuJL/xW18/stYJuIeDEwPp+/N+r8yh5hP5SZ/wlsRfHL7ssp3sOfZw3Pdd1u59T2vYuiRv+1FAnJsIj4AsWXwRF1299e9/+hwGsi4ku1feq367AXcEzt2DdR/ML6atZ+zRiQMvMpii9I59RdF/eiqJ27A5hNcT1+de2X+mG1X/H3B2bUfrzZCzi19lxfRfGLbEdyXX/e1uSq2vRBnr+2Pwh0JAudnevVy17TsdYW21spvuyQmTfRd374uQJ4R+0L8vupPQaKFipTKa6X36BoCbG256r+Otlhfa+NO1DUEt8FkJlfyMwfdKzMzOeAyyiShpcBYzLzL2soZwbwp4g4Hbi0ozzgpiy0U/wY8TZqnwl1v/r/mOc/C9bkEWBpRPyJIqH8XGYuY+3Xip7UW8/b2p4TKD5nn87MJ2vxv632d35mtmbmMxSJf2fP+a3AeyLicooflb5aq3ndGfhJLebpwIiIGFO/Y2a2UVxn9o2IU4HDV3tuzq1t91eKBOlVtTgvi4gfAndm5tWs/Xtdh9/Xpv8LbNrJY1lfvfU1APAuinsB7wX+RFErPKVu/RWZuWPd35qayr7gXNetW9Pn/dt4/hzeX3ucb+gkxt8BP4yIc4EnKV43b6CoSby+9tg/TnGuNwibg1aj477AXSi+AEBRZf5uYEFmPhURHdu2A011+66oTVdP0JuAQZn5r9ob8ACKF/qjFBewnShewBMpmhF9AqB2sRpU22ZNx1EPqf0yFBTNADtcSPF6uBv4QxbNN9Z4bmv/1yeQz1G8PjrKn0BRQ3V/Pn8vYgtFDcDaXkfqQu1X4BGZ+SvgJxQf7P9/e/caY1dVhnH8XxoqIHghWNKiRJT0MWBiaVpLEaOpIRpJuRlTkApeWhANiEEuIgIlxiGIaRPlZvzAtaSlBLCFKsJYoFqMY6F0CrwpKIgghtLIzYIUxw/v2p09h3OmZ+gUOj3P79OcM2tfzlrn7HV7195zyEqtWV5X6nm8U3m9iAzhWUKODB9XS1Mv2/vIkJrlZAj5gianNpocMV1Vjr83sIEcdXZZ10SG5lRhoZB5Nz/KDXYkvY/+fLqebMAcQo7AVumnR8SGkn4c2Sg7ilJuJZrijpL+mSjruksnstKsLAYr640NaRtfD3ZuI/I3HxnquJoMX5tODrYcS87OrAR+QXZ2nyVncSsbW/xdGfK1sTS6Kicy8Hr7Xt4cMXMd2TB8P3ltR1IVmg058/Ld0sg7nJzduRD4B82vF4PVBUgaVTqNOwNExCZJU8nO4xeBlcpw41bXimGzvZZbRExskScDtmHLed54vCrP10n6GBldMIMcyJkKvFqdczmnDwIbGr5Th5Khv9eTnfOHGDhD1nh+r0fEPElLyE7NJZIWk6GYzdp1VR5Und7GzzCstufvABkKuq8y9BJydvJkcuCmKUmTyRlIgJ6ImN1Y1pIOaDw27X2X6v+rvkuLJa0ky/Z75DViKbAiIo4o57QLzQeFtwnPBG4bdwMnAGvKTA/ktP4ZDAwFhewcHi9ppxLWcEh5vxs4TtKuZUr76/THYS8jQwaWl3SnAveXWcjlwNGSxipvVnIFGZa0ApgmaZ/aKI4Nk5Knc8kZpMer9yNv8PAUGQJSzb4OVrZ19wIzJY2SNBa4h1wDtqekKkzhG8CCMpL8ZC2s5SvD+fl2cP8BupR3Hqtu8jORLMs35XVtu2NL+slko3ANuS7l/Ii4jZydryq0zZRruSaQ6xWWkYNDVZpN9Ffu3cApZZtxZANiX1pfMzrdGWT0xTgy774qaffyG7uVEsZENt5nkmHYK8p73cC3AUql30uuTdksIp6pjSK3ddOWLZR1u1qd213kGpsqNGv/Ie73nbSI7ID31OrIV4DHImIe8GdyBmUoefUoQ7w21mcGgADG1hp9Z9Efpl+lvx8YT+b7DeW92bV9PChpHTnY20WGslWN4UNr9e8J5PdhORmmVs3ezKG/LlgPHFiuR1UD8SCyHrg3Ir5PrgUVra8Vw227K7dB8gSyLfQuZRTWDHJpTjdwoqTRknYjQwh/T+b3RyTtUsrj07B5rdvciLiJ/B2OJRv76yTNKmkOI+vrxu/UBLJD8JNyjGMa8ub4sv1kcsBhnaQ/AXtExHxyjdwkWrfrBlOvS4bT9vgd2Jusez8eER+OvHvoQWTUXMuw6Ijoqe1ndouyfk9J3qy+7ybDRinH+RTZGV5PmYmX9EmyTkLSQmBKRFxFrhGeRM4eTpM0oRznR8ClQ8i7reJO4DYQEb3kwtQ7a293k9P3jZ3Ay8mp9EfI6ebeso+l5AhBD7CWXJ9UhabcToasrSAv9mNK2mp909xyvLXkD/HiiPgX2Vm8ixyZenG4Pm8HG68SX07O9O3DwJmfynXAB8iKaktlW3c5eXFdTZbbqSWs5cvAzyQ9RI5ef7OknwVcIOkBvE6sbSUsZC6wVFKQFdIb5LquVnkN2WBYRYZwzSyDMBcCKyQ9TDYiniBvBFE/3gYyhGQt+bvfgwwXfzfZkDheedOJucCuknrJ3/NZEfE4La4ZnS76w0LHkDOxN5MVbC+5qP+aku4pspJeXGZZIK+NB5dyXkjOqmz13Ua3UNbtanVuF5Drz9eSI/IjJRwUsnwm0h9OBrm2aKfy21lF/g73a7JtUxHxGltxbSyzKbOAa8v2B9A/U1y3EHiphH817mMT2eH/nfJW9QfTv77nGbJT+DAZwvirEgbYRd6N+FHyxkHnlfTnkPXESrKDSmQ480qgt1x7HiY7k62uFcNteyy3VnkCOet0X/l/VwkVvIqcmV1NPt5gSUTcEhFrybbVWvKmXNWjFa4FJGlNee/MiPg32YGbXc65i6wDNs8kF6vJa8+jZb/PkW23yu7ls11Jrod7HTgXuFrSX0o+nNOqXTd4zvIb4EpJwz1IuN19B8hBmTsi4unaPv9Khq+e1GKbZlqVNTSv708Dppf0t5KRAP8ko4D2LPlxKvk9gxwM+GHZzyXAKRHxLNnpXVT2M4kc0HxbjOrra/zOmpnZYNShz+Yzs6FTw50nbdtTwx0azd6qHbm+90ygmZmZmZlZB/FMoJmZmZmZWQfxTKCZmZmZmVkHcSfQzMzMzMysg7gTaGZmZmZm1kHcCTQzs44n6SJJJ5S/z5d0ZBvb9EnaawjHaHe/baUzMzN7q7bFgyTNzMxGlIg4v/ZyOvm8seHW7n631fHNzMwAdwLNzGwEKs9e6wL+Dgh4hXyA8mnl9c3kQ3fnkQ/t3gMYRT7Q9w+Srgb2JB9AvBTYm3yw/EZgMvBTSW+QD2e+rGw/jnz488zycPFW5zYXOJp8iPLzwNeAY9rZL/mQ5Hq6I4HeiLi07Pvq6rWkU4BvleO8CpxcHohtZmY2KIeDmpnZSDUFuDgiJgIvAj8ADgcmAd8BpgHjgWkRcQBwDXBObfvdIuLAiDi7eiMiLgN6gDMj4hZgDnBNRBwM7A/sV47RlKQPAacDUyJiMnAnMLXd/TZJ1+o4o4H5wBciYgrwS+DQdjLNzMzMM4FmZjZS/S0iHih/Pw68EBH/BdZLehF4ATgPOFnSR4HPAi/Vtl/RxjHOBg6TdBYwgexU7j5I+qeB1cAqScuAZRFx9zDsd4CIeEPSTcAfJd0O/BZY0O72ZmbW2TwTaGZmI9VrDa9fb3j9OeD28vdtwJVkSGjl5TaOcSNwEvAkGVq6qmEfA0TE/4DPkCGgzwPzJF2yFfvta3h/TO1Ys4AZwGPkDOeNbXweMzMzdwLNzGyHNQNYEhFXkCGWRwGj29huE7Bz+fvzwEURsbC8njrYPiR9glxb+EhEdJEdvClD3G893XPkGkEkjSc7mEjaS9JTwPMRMZ+c8ayOY2ZmNiiHg5qZ2Y7qdOA6SWvI+u5O4EuStjQA+mugS9IY4FzgFkmvkOGl95Br+AaQ9CB505keSYuAHkkvkzeaOW2I+62n+zlwg6QAngC6ASJivaQfA/Z4ekMAAABzSURBVHdL2kh2HOcMLXvMzKxTjerr63unz8HMzMzMzMzeJg4HNTMzMzMz6yDuBJqZmZmZmXUQdwLNzMzMzMw6iDuBZmZmZmZmHcSdQDMzMzMzsw7iTqCZmZmZmVkHcSfQzMzMzMysg7gTaGZmZmZm1kH+DxyT7gGtD90PAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 1080x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(15,6))\n",
    "ax=sns.barplot(x='marital.status',y='income',data=data,hue='sex')\n",
    "ax.set(ylabel='Fraction of people with income > $50k')\n",
    "data['marital.status'].value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The above plot shows the the fraction of people earning more than $50k per annum, grouped by their marital status and gender. The data shows that married people have a higher %age of high-earners, compared to those who either never married or are widowed/divorced/separated. The black lines indicate 2 standard deviations (or 95\\% confidence interval) in the data set. The married spouses of armed forces personnel have a much higher variation in their income compared to civil spouses because of low-number statistics."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x10dcc30b8>"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12,6))\n",
    "sns.countplot(data['workclass'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[Text(0, 0.5, 'Fraction of people with income > $50k')]"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12,6))\n",
    "ax=sns.barplot('workclass', y='income', data=data, hue='sex')\n",
    "ax.set(ylabel='Fraction of people with income > $50k')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'Heatmap showing correlations between numerical data')"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x576 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(10,8))  \n",
    "sns.heatmap(data.corr(),cmap='Accent',annot=True)\n",
    "#data.corr()\n",
    "plt.title('Heatmap showing correlations between numerical data')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x1a23a9b518>"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12,6))\n",
    "sns.boxplot(x=\"income\", y=\"age\", data=data, hue='sex')\n",
    "#data[data['income']==0]['age'].mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The mean age of people earning more than 50k per annum is around 44 whereas the mean age of of those earning less than 50k per annum is 36."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x1a2321a358>"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "norm_fnl = (data[\"fnlwgt\"] - data['fnlwgt'].mean())/data['fnlwgt'].std()\n",
    "plt.figure(figsize=(8,6))\n",
    "sns.boxplot(x=\"income\", y=norm_fnl, data=data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As evident from the plot above, there are many outliers in the `fnlwgt` column and this feature is uncorrelated with `income`, our target variable. The correlation coefficient (which one can read from the heatmap) is -0.0095. The number of outliers, i.e. the number of records which are more than 2 s.d's away from the mean, is 1249."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1249, 15)"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data[norm_fnl>2].shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[Text(0, 0.5, 'Fraction of people with income > $50k')]"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(10,5))\n",
    "ax = sns.barplot(x='sex',y='income',data=data)\n",
    "ax.set(ylabel='Fraction of people with income > $50k')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The fraction of rich among men is significantly higher than that among women."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x1a23f06160>"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12,6))\n",
    "sns.boxplot(x='income',y ='hours.per.week', hue='sex',data=data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "On the basis of the above plot, we can only conclude that males put in slightly more hours per week than women on an average."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[Text(0, 0.5, 'mean hours per week')]"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(15,6))\n",
    "ax = sns.barplot(x='marital.status',y='hours.per.week',data=data,hue='sex')\n",
    "ax.set(ylabel='mean hours per week')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[Text(0, 0.5, 'Mean education')]"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(10,6))\n",
    "ax = sns.barplot(x='income', y='education.num',hue='sex', data=data)\n",
    "ax.set(ylabel='Mean education')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The `education.num` is label encoded such that a higher number corresponds to a higher level of education. As on would naïvely expect, people who earn more (>50k per annum) are also highly educated. The mean education level for `income=1` class is between 11 (Assoc-voc) and 12 (Assoc-acdm) whereas that for the `income=0` class is between 9 (HS-grad) and 10 (Some-college)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "White                 27816\n",
      "Black                  3124\n",
      "Asian-Pac-Islander     1039\n",
      "Amer-Indian-Eskimo      311\n",
      "Other                   271\n",
      "Name: race, dtype: int64\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[Text(0, 0.5, 'Fraction of people with income > $50k')]"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "print(data['race'].value_counts())\n",
    "plt.figure(figsize=(12,6))\n",
    "ax=sns.barplot(x='race',y='income',data=data)\n",
    "ax.set(ylabel='Fraction of people with income > $50k')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<seaborn.axisgrid.JointGrid at 0x1a235b7898>"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 864x432 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x432 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12,6))\n",
    "sns.jointplot(x=data['capital.gain'], y=data['capital.loss'])\n",
    "#print(data[((data['capital.gain']!=0) & (data['capital.loss']!=0))].shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x1a249c9940>"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAs4AAAHoCAYAAABZ8WmaAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvnQurowAAIABJREFUeJzt3XuUZVddJ/BvdVfTD7ojpKmQxJgoxN4EXSQYE1RAUMIjioJLIouESNAEMsrIrOEhShCGEQcZVsC45DGEGLBlQIMykdCIhAwJYCDyiArJhqUhEtNATwelO5WuVHXV/HFPh7Kp7tq3c289P5+1slJ33985d9/69en61ul9zxmZmZkJAABweGsWewIAALAcCM4AANBAcAYAgAaCMwAANBCcAQCggeAMAAANBGcAAGggOAMAQAPBGQAAGgjOAADQYHSxJzDL+iRnJNmZZP8izwUAgJVrbZLjktyUZKJ1o6UUnM9IcsNiTwIAgFXj8Uk+0Vq8lILzziT51rfuzvT0zIK+8Natm7N7994FfU0Wnj6vDvq8OujzyqfHq8Ni9XnNmpE8+MEPTLr82WopBef9STI9PbPgwfnA67Ly6fPqoM+rgz6vfHq8Oixyn/taHuzDgQAA0EBwBgCABoIzAAA0EJwBAKCB4AwAAA0EZwAAaCA4AwBAA8EZAAAaCM4AANBAcAYAgAaCMwAANBCcAQCggeAMAAANBGcAAGggOAMAQAPBGQAAGgjOAADQQHAGAIAGo/MVlFIuTPKiWUM/kORPknwgyaVJNiZ5X631kq7+tCSXJzkqyfVJLq61Tg143otuajqZmGx/W+vXjWbUrykAAMvWvMG51np5ekE4pZQfSi8w/36STyZ5QpKvJbmmlHJ2rXVHku1JLqy13lhKeWeSi5K8dUjzXzQTk1O56ZZvNNefccpDM7p+3m83AABLVL/nQN+a5LeTPCzJV2qtt3Vnk7cnOaeUclKSjbXWG7v6K5OcM6jJAgDAYmk+BVpKOSu9UPznpZTnJNk56+mdSU5Icvwhxptt3bq5n/KBGRvb0lf9zF3j2bJ5Q3P9pk3rM3b0pn6nxYD122eWJ31eHfR55dPj1WE59bmftQMvTG9Nc9I7Uz0z67mRJNOHGW+2e/feTE/PzF84QGNjW7Jr156+thmfmMqevfva68cnsmv//n6nxgAdSZ9ZfvR5ddDnlU+PV4fF6vOaNSNHdLK2aalGKeUB6a1nvrobuiPJcbNKjk1y52HGAQBgWWtd4/yoJF+utd7dPf50klJKObmUsjbJuUl21FpvT7KvlPLYru78JDsGOmMAAFgErcH5YemdTU6S1Fr3JbkgyfuTfCnJrUmu6p4+L8mbSim3Jtmc5LJBTRYAABZL0xrnWuufJfmzg8auTXLqHLU3JzlzILMDAIAlwi05AACggeAMAAANBGcAAGggOAMAQAPBGQAAGgjOAADQQHAGAIAGgjMAADQQnAEAoIHgDAAADQRnAABoIDgDAEADwRkAABoIzgAA0EBwBgCABoIzAAA0EJwBAKCB4AwAAA0EZwAAaCA4AwBAA8EZAAAaCM4AANBAcAYAgAaCMwAANBCcAQCggeAMAAANBGcAAGggOAMAQAPBGQAAGgjOAADQQHAGAIAGgjMAADQQnAEAoIHgDAAADQRnAABoIDgDAEADwRkAABoIzgAA0EBwBgCABoIzAAA0EJwBAKCB4AwAAA0EZwAAaCA4AwBAA8EZAAAaCM4AANBAcAYAgAajiz0B5jY1nUxMTjXXr183mlG/BgEADI3gvERNTE7lplu+0Vx/xikPzeh67QQAGBbnKAEAoEHTKcpSys8leXWSByb5SK31xaWUs5JcmmRjkvfVWi/pak9LcnmSo5Jcn+TiWmv7mgMAAFiC5j3jXEp5WJK3JXlmkkcl+ZFSytlJrkjyjCSnJDmjG0uS7UleVGvdlmQkyUXDmDgAACyklqUav5DeGeU7aq2TSZ6dZDzJV2qtt3Vnk7cnOaeUclKSjbXWG7ttr0xyzhDmDQAAC6plqcbJSe4tpVyd5MQkH0zyxSQ7Z9XsTHJCkuMPMd5s69bN/ZQPzNjYlr7qZ+4az5bNG5rrN21an7GjNy2Z/a9W/faZ5UmfVwd9Xvn0eHVYTn1uCc6jSX4yyROT7E1ydZJ7kszMqhlJMp3eGey5xpvt3r0309Mz8xcO0NjYluzataevbcYnprJn7772+vGJ7Nq/f8nsfzU6kj6z/Ojz6qDPK58erw6L1ec1a0aO6GRtS3D+epKP1lp3JUkp5S/TW34xO6Udm+TOJHckOW6OcQAAWNZa1jh/MMlTSykPKqWsTXJ2kquSlFLKyd3YuUl21FpvT7KvlPLYbtvzk+wYxsQBAGAhzRuca62fTvKGJJ9I8qUktyd5a5ILkry/G7s1vTCdJOcleVMp5dYkm5NcNvBZAwDAAmu6jnOt9Yr0Lj8327VJTp2j9uYkZ97/qQEAwNLhzoEAANBAcAYAgAaCMwAANBCcAQCggeAMAAANBGcAAGggOAMAQAPBGQAAGgjOAADQQHAGAIAGgjMAADQQnAEAoIHgDAAADQRnAABoIDgDAEADwRkAABoIzgAA0EBwBgCABoIzAAA0EJwBAKCB4AwAAA0EZwAAaCA4AwBAA8EZAAAaCM4AANBAcAYAgAaCMwAANBCcAQCggeAMAAANBGcAAGggOAMAQAPBGQAAGgjOAADQQHAGAIAGgjMAADQQnAEAoIHgDAAADQRnAABoIDgDAEADwRkAABoIzgAA0EBwBgCABoIzAAA0EJwBAKCB4AwAAA0EZwAAaCA4AwBAA8EZAAAaCM4AANBgtKWolHJdkmOSTHZDL0zy8CSXJFmX5M211j/qas9KcmmSjUneV2u9ZNCTBgCAhTZvcC6ljCTZluSkWutUN/a9Sd6b5PQkE0k+1YXr25JckeQJSb6W5JpSytm11h1Dmj8AACyIljPOpfv/R0opW5O8I8meJB+rtd6VJKWUq5I8K8nHk3yl1npbN749yTlJBGcAAJa1ljXOD05ybZJfSPKkJBcnOTHJzlk1O5OckOT4Q4wDAMCyNu8Z51rr3yb52wOPSynvTG8N8+/OKhtJMp1eEJ+ZY7zZ1q2b+ykfmLGxLX3Vz9w1ni2bNzTXb9q0PmNHb1oy+1+t+u0zy5M+rw76vPLp8eqwnPrcssb5cUnW11qv7YZGknw1yXGzyo5NcmeSOw4x3mz37r2Znp6Zv3CAxsa2ZNeuPX1tMz4xlT1797XXj09k1/79S2b/q9GR9JnlR59XB31e+fR4dVisPq9ZM3JEJ2tb1jg/KMlrSyk/kd4VNJ6X5LlJtpdSxpLcneQXk7wgyd8nKaWUk9P7oOC56X1YEAAAlrV51zjXWj+Y5Jokn0/y2SRX1Fo/meSVSa5L8oUk76m1fqbWui/JBUnen+RLSW5NctVwpg4AAAun6TrOtdZXJXnVQWPvSfKeOWqvTXLqQGYHAABLhDsHAgBAA8EZAAAaCM4AANBAcAYAgAaCMwAANBCcAQCggeAMAAANBGcAAGggOAMAQAPBGQAAGgjOAADQQHAGAIAGgjMAADQQnAEAoIHgDAAADQRnAABoIDgDAEADwRkAABoIzgAA0EBwBgCABoIzAAA0EJwBAKCB4AwAAA0EZwAAaCA4AwBAA8EZAAAaCM4AANBAcAYAgAaCMwAANBCcAQCggeAMAAANBGcAAGggOAMAQAPBGQAAGgjOAADQQHAGAIAGgjMAADQQnAEAoIHgDAAADQRnAABoIDgDAEADwRkAABoIzgAA0EBwBgCABoIzAAA0EJwBAKCB4AwAAA0EZwAAaCA4AwBAA8EZAAAajLYWllLemOQhtdYLSimnJbk8yVFJrk9yca11qpRyYpLtSY5JUpOcV2vdO4R5AwDAgmo641xKeVKS580a2p7kRbXWbUlGklzUjb8lyVtqrY9I8ndJXjXAuQIAwKKZNziXUo5O8rokv9c9PinJxlrrjV3JlUnOKaWsS/KTSa6aPT7g+QIAwKJoWarx9iSvTPJ93ePjk+yc9fzOJCckeUiSb9dapw4a78vWrZv73WQgxsa29FU/c9d4tmze0Fy/adP6jB29acnsf7Xqt88sT/q8OujzyqfHq8Ny6vNhg3Mp5cIkX6u1XltKuaAbXpNkZlbZSJLpOcbTjfdl9+69mZ4+eDfDNTa2Jbt27elrm/GJqezZu6+9fnwiu/bvXzL7X42OpM8sP/q8OujzyqfHq8Ni9XnNmpEjOlk73xnnZyc5rpTyhSRHJ9mcXjg+blbNsUnuTPLNJN9TSllba93f1dzZ94wAAGAJOuwa51rrk2utP1xrPS3J7yS5utb6/CT7SimP7crOT7Kj1jqZ5Ib0wnaS/HKSHUOaNwAALKgjvY7zeUneVEq5Nb2z0Jd147+W5AWllC8leXySS+7/FAEAYPE1X8e51nplelfKSK315iRnzlFze5InDmZqAACwdLhzIAAANBCcAQCggeAMAAANBGcAAGjQ/OFA6NfUdDIxOTV/YWf9utGM+lUOAFiiBGeGZmJyKjfd8o3m+jNOeWhG1/sjCQAsTc7vAQBAA8EZAAAaCM4AANBAcAYAgAY+ibVARtaM5O6J9itMTM8McTIAAPRNcF4gE5P7c/OXdzXXn7ptbIizAQCgX5ZqAABAA8EZAAAaCM4AANBAcAYAgAaCMwAANBCcAQCggeAMAAANBGcAAGggOAMAQAPBGQAAGgjOAADQQHAGAIAGgjMAADQQnAEAoIHgDAAADQRnAABoIDgDAEADwRkAABoIzgAA0EBwBgCABoIzAAA0EJwBAKCB4AwAAA0EZwAAaCA4AwBAA8EZAAAaCM4AANBAcAYAgAaCMwAANBCcAQCggeAMAAANBGcAAGggOAMAQAPBGQAAGgjOAADQQHAGAIAGgjMAADQQnAEAoMFoS1Ep5bVJnpVkJsk7a62XllLOSnJpko1J3ldrvaSrPS3J5UmOSnJ9kotrrVPDmDwAACyUec84l1KekOSnkzwqyY8m+c+llFOTXJHkGUlOSXJGKeXsbpPtSV5Ua92WZCTJRcOYOAAALKR5g3Ot9eNJfqo7a3xMemepH5TkK7XW27rx7UnOKaWclGRjrfXGbvMrk5wzlJkDAMACalqqUWudLKX8tyQvTfLnSY5PsnNWyc4kJxxmvNnWrZv7KR+YsbEtfdXP3DWeLZs3NNevWzc61PpNm9Zn7OhNzfULod/v0UK8h377zPKkz6uDPq98erw6LKc+NwXnJKm1vrqU8vtJ/irJtvTWOx8wkmQ6vTPYc4032717b6anZ+YvHKCxsS3ZtWtPX9uMT0xlz959zfWTk8OtHx+fyK79+5vrF0K/36Nhv4cj6TPLjz6vDvq88unx6rBYfV6zZuSITta2rHF+RPeBv9Rax5P8RZInJjluVtmxSe5McschxgEAYFlruRzdw5K8o5SyvpTygPQ+EPj2JKWUcnIpZW2Sc5PsqLXenmRfKeWx3bbnJ9kxjIkDAMBCavlw4IeSXJPk80k+m+RTtdb3JrkgyfuTfCnJrUmu6jY5L8mbSim3Jtmc5LLBTxsAABZW64cDX5PkNQeNXZvk1Dlqb05y5gDmBgAAS4Y7BwIAQAPBGQAAGgjOAADQQHAGAIAGgjMAADQQnAEAoIHgDAAADQRnAABoIDgDAEADwRkAABoIzgAA0EBwBgCABoIzAAA0EJwBAKCB4AwAAA0EZwAAaCA4AwBAA8EZAAAaCM4AANBAcAYAgAaCMwAANBCcAQCggeAMAAANBGcAAGggOAMAQAPBGQAAGgjOAADQQHAGAIAGgjMAADQQnAEAoIHgDAAADQRnAABoIDgDAEADwRkAABoIzgAA0EBwBgCABoIzAAA0EJwBAKCB4AwAAA0EZwAAaCA4AwBAA8EZAAAaCM4AANBAcAYAgAaCMwAANBCcAQCggeAMAAANBGcAAGggOAMAQAPBGQAAGoy2FJVSXp3kl7qH19RaX15KOSvJpUk2JnlfrfWSrva0JJcnOSrJ9UkurrVODXzmAACwgOY949wF5KckeXSS05KcXkp5TpIrkjwjySlJziilnN1tsj3Ji2qt25KMJLloGBMHAICF1LJUY2eSl9Ra7621Tia5Jcm2JF+ptd7WnU3enuScUspJSTbWWm/str0yyTlDmDcAACyoeZdq1Fq/eODrUsoPprdk4w/TC9QH7ExyQpLjDzHebOvWzf2UD8zY2Ja+6mfuGs+WzRua69etGx1q/aZN6zN29Kbm+oXQ7/doId5Dv31medLn1UGfVz49Xh2WU5+b1jgnSSnlh5Jck+RlSabSO+t8wEiS6fTOYM/MMd5s9+69mZ6emb9wgMbGtmTXrj19bTM+MZU9e/c1109ODrd+fHwiu/bvb65fCP1+j4b9Ho6kzyw/+rw66PPKp8erw2L1ec2akSM6Wdt0VY1SymOTXJvkFbXWdyW5I8lxs0qOTXLnYcYBAGBZa/lw4Pcl+UCSc2ut7+2GP917qpxcSlmb5NwkO2qttyfZ1wXtJDk/yY4hzBsAABZUy1KNlybZkOTSUsqBsbcluSDJ+7vnPpTkqu6585K8o5RyVJLPJblsgPMFAIBF0fLhwBcnefEhnj51jvqbk5x5P+cFAABLijsHAgBAA8EZAAAaCM4AANBAcAYAgAaCMwAANBCcAQCggeAMAAANBGcAAGggOAMAQAPBGQAAGsx7y+3VYmo6mZicaq6fnhniZAAAWHIE587E5FRuuuUbzfWnbhsb4mwAAFhqLNUAAIAGgjMAADQQnAEAoIHgDAAADQRnAABoIDgDAEADl6MDAGAg+r0vxobxe4c4m8ETnAEAGIh+74vxhNNPzMgQ5zNolmoAAEADwRkAABoIzgAA0EBwBgCABoIzAAA0EJwBAKCB4AwAAA0EZwAAaCA4AwBAA8EZAAAaCM4AANBAcAYAgAaCMwAANBCcAQCggeAMAAANBGcAAGggOAMAQAPBGQAAGgjOAADQQHAGAIAGgjMAADQQnAEAoIHgDAAADQRnAABoIDgDAEADwRkAABoIzgAA0EBwBgCABoIzAAA0GG0tLKUcleRTSZ5ea/1qKeWsJJcm2ZjkfbXWS7q605JcnuSoJNcnubjWOjXwmQMAwAJqOuNcSnlMkk8k2dY93pjkiiTPSHJKkjNKKWd35duTvKjWui3JSJKLBj1pAABYaK1LNS5K8utJ7uwen5nkK7XW27qzyduTnFNKOSnJxlrrjV3dlUnOGeB8AQBgUTQt1ai1XpgkpZQDQ8cn2TmrZGeSEw4zDgAAy1rzGueDrEkyM+vxSJLpw4w327p18xFO6f7ZtGl9tmze0Fy/bt3okqrftGl9xo7e1Fy/EGbuGl9y72FsbMtQ98/SoM+rgz6vfHq8/PT7sz9ZXn0+0uB8R5LjZj0+Nr1lHIcab7Z7995MT8/MXzhAY2NbMj4+kT179zVvMzk5taTqx8cnsmv//ub6hTA+sbTew9jYluzatWdo+2dp0OfVQZ9XPj1envr92Z9kUfq8Zs3IEZ2sPdLL0X06SSmlnFxKWZvk3CQ7aq23J9lXSnlsV3d+kh1H+BoAALBkHFFwrrXuS3JBkvcn+VKSW5Nc1T19XpI3lVJuTbI5yWX3f5oAALC4+lqqUWv9/llfX5vk1Dlqbk7vqhsAALBiuHMgAAA0EJwBAKCB4AwAAA0EZwAAaCA4AwBAA8EZAAAaHOmdA1liRtaM5O6Jqb62Wb9uNKN+dQIAaCI4rxATk/tz85d39bXNGac8NKPr/REAAGjhfCMAADQQnAEAoIHgDAAADQRnAABoIDgDAEADl1Sg2dR0MjHZfsm76ZkhTgYAYIEJzjSbmJzKTbd8o7n+1G1jQ5wNAMDCslQDAAAaOOO8ivV7t0FLLwCA1UxwXsX6vdugpRcAwGpmqQYAADQQnAEAoIHgDAAADQRnAABoIDgDAEADwRkAABoIzgAA0EBwBgCABoIzAAA0EJwBAKCB4AwAAA0EZwAAaDC62BOAIzU1nUxMTjXXbxi/d4izAQBWOsGZZWticio33fKN5vonnH5iRoY4HwBgZROcO7/xpusz9uCNedpjTsy7P1zzy08refeHa5Lc9/WjHr41f/9Pu3sbdM81O2ifsx0YHxlJNq4fzbOe+PD89Y2359ijN31X/Vzz2rRhNF+/azzHHr0pSe6b/2zv/nDN6NqRnPvkbd/1+rP3913T/vS/5GmPObH5bc5+7Zdc9onv2u+f/HXNzMx/HJtrvv1qmeevvu5vcsUrfrp5nx+44Z/zzMc/7H7N64Bfu/Tj2Xfv/r5e/2C/8vqP3a/th+WiN1yXd7z8pxZ7Gs1e/6efzSvOO32xpwFA+v/ZvNisce7MJPnmt+45bM19oXlYc5hJxvf1lh585DNfa95ufN9UU/3U/pm+5zTf96RfM/1Pocmg55kkV3/yqwPb17579885PjWd3D0x1fTfgfqlZv/0kJo6JF/+2r8v9hQAWKaccYZF1O9yk4nJqYyud9gCwGJwxhkAABoIzgAA0EBwBgCABoIzAAA0EJwBAKCB4AwAAA0EZwAAaCA4AwBAA8EZAAAauAUZDNDUdO/ufody4NbZByyzu1UDwKomOMMAzXcL7YOfO3Xb2LCnBAAMiKUaAADQQHAGAIAGgjMAADQQnAEAoMFQPhxYSjk3ySVJ1iV5c631j4bxOsDScPDVQg5n/brRjPqVHYBlaODBuZTyvUlel+T0JBNJPlVKua7W+qVBvxawNBzuSiIHO+OUh2Z0vQv6ALD8DOOn11lJPlZrvStJSilXJXlWktfOs93aJFmzZmQIU5rfMQ/emCTZtGFdjnnwxvv+P3vs/jrUfg4en2sOc9XOtd2B93Lg67ne3wGja9d81/4ONntfB+oPZ3b9XPs91NimDesyunZNX/2fPZ+53vNcc+tn//3WHzyng/eVfPf3uOV7Onsf/X6PFkLL9362Yb+Hvffcm4mp6UM+f8yDN/6H5x8wujZrnQFflpbascDg6fHy08/PteTIftYOwqzXXNvPdiMzM4O9A0Mp5beSPLDWekn3+MIkZ9ZaXzDPpo9LcsNAJwMAAIf2+CSfaC0exhnnNUlmp/GRJIc+/fMdN6U3+Z1J9g9hXgAAkPTONB+XXv5sNozgfEd6AfiAY5Pc2bDdRPpI/AAAcD/8U78bDCM4fzTJa0opY0nuTvKLSeZbpgEAAEvawD8SU2v91ySvTHJdki8keU+t9TODfh0AAFhIA/9wIAAArEQuwgQAAA0EZwAAaCA4AwBAA8EZAAAaCM4AANBgGNdxXjZKKecmuSTJuiRvrrX+0SJPiQallFcn+aXu4TW11peXUs5KcmmSjUneN+uW76cluTzJUUmuT3JxrXWqlHJiku1JjklSk5xXa91bSnlQkj9N8rAku5L8Uq316wv49jhIKeWNSR5Sa71gUP0spTwgyTuT/GiSe5KcW2u9dcHfHCml/FySVyd5YJKP1Fpf7HheeUopz03yW93DHbXWlzqeV4ZSylFJPpXk6bXWrw77+F3sfq/aM86llO9N8rokj0tyWpIXlFIeubizYj7dAfmUJI9Or2+nl1Kek+SKJM9IckqSM0opZ3ebbE/yolrrtvRu/35RN/6WJG+ptT4iyd8leVU3/rtJbqi1npLkHUn+YPjvikMppTwpyfNmDQ2qn7+R5O5u/L8kuXKY74O5lVIeluRtSZ6Z5FFJfqQ7dh3PK0gpZVOSy5I8IcmpSR7f/V3ueF7mSimPSe+uz9u6xxsz/ON3Ufu9aoNzkrOSfKzWelet9e4kVyV51iLPifntTPKSWuu9tdbJJLekd8B+pdZ6W611Kr2D85xSyklJNtZab+y2vbIbX5fkJ9Pr+X3j3dc/m95vuEnyv5Oc3dWzwEopR6f3y+3vdY8H2c/7xmut1ycZ6856sLB+Ib0zUnd0x/Ozk4zH8bzSrE0vbzwwvX/hXZdkMo7nleCiJL+e5M7u8ZkZ/vG7qP1ezcH5+PRC2AE7k5ywSHOhUa31iwcOvFLKD6a3ZGM6c/fyUD1+SJJvdwf17PHM3qZ7/ttJxobyZpjP29O7C+m3useD7Kfjf2k4OcnaUsrVpZQvJPm1HLo3judlqta6J72ziLcmuSPJV5PcG8fzsldrvbDWesOsoYU4fhe136s5OK9JMvu2iSPpBTCWgVLKDyX5myQvS/LPmbuXh+rxwePJd3o/ctC4PxeLoJRyYZKv1VqvnTU8yH46/peG0fT+9e9Xk/x4ksekt57R8byClFIeleRXkpyUXujZn96SO8fzytN6nC7bfq/m4HxHkuNmPT423/mnBpawUspjk1yb5BW11nfl0L081Pg3k3xPKWVtN35cvtP7f+3qUkoZTbIlye7hvBMO49lJntKdhXxtkp9PcmEG10/H/9Lw9SQfrbXuqrXek+Qv0wvSjueV5alJrq21frPWOpHeP8c/MY7nlWghfh4var9Xc3D+aJInlVLGug8u/GKSDy/ynJhHKeX7knwgvU/Rvrcb/nTvqXJyd/Cdm96ntm9Psq8L2klyfjc+meSG9MJZkvxykh3d1x/qHqd7/oaungVUa31yrfWHa62nJfmdJFfXWp+fwfXzvvFSyuOS7Ku1/suw3xff5YNJnlpKeVB37J6d3lpHx/PKcnOSs0opDyyljCT5uSQfj+N5JVqIn8eL2u9Vezm6Wuu/llJemeS6JA9Icnmt9TOLPC3m99IkG5JcWko5MPa2JBckeX/33IfynQ8anJfkHd3lcj6X3ie7k95ayneVUi5J8i9JntONvyrJlaWULyb5t257lo5B9fMPk7y9G59I7y9xFlit9dOllDek96n8dektv3premthHc8rRK31I6WURyf5bHofCvxMkten9y8MjucVpNa6r5RyQYZ7/C5qv0dmZg5eWgIAABxsNS/VAACAZoIzAAA0EJwBAKCB4AwAAA0EZwAAaCA4AyxBpZQvdNc3/p5Sysca6i8opXxwQK99fCnlU4PYF8BKsmqv4wywlHU3f0kp5fuTnLnAr31nkp9YyNcEWA4EZ4ABKqX8SpKXJNmf5P8leX6S/5rkx9K7ZexIkgtrrZ8spVyZ5J4kpyU5JslHkvxGrXWylDK9vbowAAAC4UlEQVSTZCzJHyfZ2N1+/PQkz0vywvRu3HR0ktfXWt96mPmsTfI/07tt+b+nd2evR9Zan1hK+bEkb0iyPr1b2P5NrfVXu7D+j7XWzaWU1yT5/u75k9K7De5za6077/93C2B5sVQDYEBKKacm+f0kT6u1PirJ1Un+V5Ljk/x4rfWRSd6V5BWzNntMkicneWT33wsP2u3zk9zTnYHemOSiJD9Ta310erehfcM807owvcD9w0l+PMnDZz334iS/U2t9TPfaP19KOX2OfTw+yTm11kckuTvJxfO8JsCKJDgDDM6Tkvx1rfVrSVJrfXOt9alJLknywlLKG5M8K8nmWdtcWWvdW2udSPLuJE891M5rrXuTPD3Jz5ZS/nuSVx60r7n8TJJ311r31VrvTfL2Wc89L8mDSim/neQt6QXzufb3f2ut3+6+/nx6Z7oBVh3BGWBwppLMHHhQStlYSvn1JNd0Q/8nydvSW64xe5sD1qS3xGNOpZQTknwhvSUTn0gvkLfMafbrzd7/9ekF61uTvDa9ZRizaw+4Z9bXM4eoAVjxBGeAwbkuyVmllOO6xy9M7wzyX3XrkP8uyTOTrJ21zbNLKetLKRvSOwP8VwftcyrJ2lLKSJIfTbIrye+mtx766cl965gP5Zokz+1eYzTJBUlmSikPSnJGkt+stf5FkhOSnHzQ3ACYRXAGGJBa6z8keVmSD5dSbk7ytCSvT/LEUso/JPlckn9K8gOllAN//44nuSHJP3T//+ODdrszyWeSfDHJTUnuSFKT3JLkxPSC9MmzNyil/Hwp5UPdwyvT+0Dg55N8Ksm9ScZrrf+W5H8k+Vwp5R/TW3f9yYP3BcB3jMzMzMxfBcDAdVfV+Mda6xuH+BpPSXJMrXV79/gPkuyrtf7msF4TYKVyOTqAle2LSV5WSnl5esswbk7ynxZ3SgDLkzPOAADQwBpnAABoIDgDAEADwRkAABoIzgAA0EBwBgCABv8faJZdkKqtiRQAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 864x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12,8))\n",
    "sns.distplot(data[(data['capital.gain']!=0)]['capital.gain'],kde=False, rug=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x1a24bab898>"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAs4AAAHoCAYAAABZ8WmaAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvnQurowAAHeBJREFUeJzt3XuQ5Wdd5/FP9/QwmTgTIKEDQS5qxXmIukwsTNRFFCVeWCnZKomuyaLBTYQC1KoFLS2jIKsWumxQLFAWZIMVXdkNFosLwUtACaYCQU1WQ/LAQmS5zOKYsJLJOE33dO8f53Q4NH35nun7zOtVlcqc33nO7zzd/fTpd//6d86ZWFhYCAAAsLrJ7Z4AAADsBsIZAAAKhDMAABQIZwAAKBDOAABQIJwBAKBAOAMAQIFwBgCAAuEMAAAFwhkAAAqmtnsCI/YluSTJkSQnt3kuAACcvvYkuSDJ7UlmqjfaSeF8SZJbtnsSAACcMZ6W5H3VwTspnI8kyWc/+2Dm5xeWHXDeeQdy333HtnRSnF6sIdbLGmK9rCHWyxpav8nJiTzykV+WDPuzaieF88kkmZ9fWDGcF6+H9bCGWC9riPWyhlgva2jDjHV6sCcHAgBAgXAGAIAC4QwAAAXCGQAACoQzAAAUCGcAACgQzgAAUCCcAQCgQDgDAECBcAYAgALhDAAABcIZAAAKhDMAABQIZwAAKBDOAABQIJwBAKBAOAMAQIFwBgCAgqntngDARpqbT2Zm58rj9+2dypRDCAAUCGfgtDIzO5fb7/5MefwlFz06U/s8FAKwNsdZAACgQDgDAECBcAYAgALhDAAABcIZAAAKhDMAABQIZwAAKBDOAABQIJwBAKBAOAMAQIFwBgCAAuEMAAAFwhkAAAqEMwAAFAhnAAAoEM4AAFAgnAEAoEA4AwBAgXAGAIAC4QwAAAXCGQAACoQzAAAUCGcAACgQzgAAUCCcAQCgQDgDAECBcAYAgALhDAAABcIZAAAKhDMAABQIZwAAKBDOAABQIJwBAKBAOAMAQIFwBgCAAuEMAAAFwhkAAAqEMwAAFAhnAAAoEM4AAFAgnAEAoEA4AwBAgXAGAICCqerA1tqrkjyq935Va+3iJG9Mck6S9yZ5Qe99rrX2hCQ3JDk/SU9yZe/92CbMGwAAtlTpiHNr7RlJfmRk0w1JXtx7P5RkIsk1w+2vS/K63vuTknwwyc9v4FwBAGDbrBnOrbVzk/xykl8ZXn5ikv2999uGQ65PcnlrbW+Sb01y4+j2DZ4vAABsi8qpGq9P8nNJHj+8/NgkR0auP5LkcUkeleRzvfe5JdvHct55B1a9fnr64Li7hC9iDZ3eFu4/noMHziqPP/vsfZk+9+yx7sMaYr2sIdbLGtoeq4Zza+3qJJ/ovd/cWrtquHkyycLIsIkk88tsz3D7WO6771jm55fuZmB6+mCOHn1g3F3CQ6yh09/xmbk8cOxEffzxmRw9ebI83hpivawh1ssaWr/JyYk1D9YuZ60jzj+Y5ILW2h1Jzk1yIIM4vmBkzGOSfDrJPyR5eGttT+/95HDMp8eeEQAA7ECrnuPce//O3vvX9d4vTvILSd7ee39ekhOttacOhz03yU2999kkt2QQ20nyw0lu2qR5AwDAljrV13G+MsmrW2v3ZHAU+jXD7S9M8mOttQ8leVqSa9c/RQAA2H7l13HuvV+fwStlpPd+Z5JLlxnz8SRP35ipAQDAzuGdAwEAoEA4AwBAgXAGAIAC4QwAAAXCGQAACoQzAAAUCGcAACgQzgAAUCCcAQCgQDgDAECBcAYAgALhDAAABcIZAAAKhDMAABQIZwAAKBDOAABQIJwBAKBAOAMAQIFwBgCAAuEMAAAFwhkAAAqEMwAAFAhnAAAoEM4AAFAgnAEAoEA4AwBAgXAGAIAC4QwAAAXCGQAACoQzAAAUCGcAACgQzgAAUCCcAQCgQDgDAECBcAYAgALhDAAABcIZAAAKhDMAABQIZwAAKBDOAABQIJwBAKBAOAMAQIFwBgCAAuEMAAAFwhkAAAqEMwAAFAhnAAAoEM4AAFAgnAEAoEA4AwBAgXAGAIAC4QwAAAXCGQAACoQzAAAUCGcAACgQzgAAUCCcAQCgQDgDAECBcAYAgALhDAAABcIZAAAKhDMAABQIZwAAKBDOAABQIJwBAKBAOAMAQIFwBgCAAuEMAAAFwhkAAAqEMwAAFAhnAAAoEM4AAFAgnAEAoEA4AwBAgXAGAIAC4QwAAAXCGQAACoQzAAAUCGcAACgQzgAAUCCcAQCgQDgDAEDBVGVQa+0VSZ6TZCHJ7/Ter2utXZbkuiT7k7yl937tcOzFSd6Y5Jwk703ygt773GZMHgAAtsqaR5xba9+W5DuSPDnJNyT58dba4SRvSvLsJBcluaS19szhTW5I8uLe+6EkE0mu2YyJAwDAVloznHvvf5Hk24dHjc/P4Cj1I5J8pPd+73D7DUkub609Mcn+3vttw5tfn+TyTZk5AABsodKpGr332dbaLyZ5aZL/nuSxSY6MDDmS5HGrbC8777wDq14/PX1wnN3Bl7CGTm8L9x/PwQNnlcefffa+TJ979lj3YQ2xXtYQ62UNbY9SOCdJ7/1lrbVfTfJHSQ5lcL7zookk8xkcwV5ue9l99x3L/PzCstdNTx/M0aMPjLM7+CLW0Onv+MxcHjh2oj7++EyOnjxZHm8NsV7WEOtlDa3f5OTEmgdrl73dWgNaa08aPuEvvffjSf4wydOTXDAy7DFJPp3kkytsBwCAXa3ycnRfleQNrbV9rbWHZfCEwNcnaa21C1tre5JckeSm3vvHk5xorT11eNvnJrlpMyYOAABbqfLkwHcmeUeSv0nyV0lu7b3/QZKrkrw1yYeS3JPkxuFNrkzy6tbaPUkOJHnNxk8bAAC2VvXJgS9P8vIl225OcniZsXcmuXQD5gYAADuGdw4EAIAC4QwAAAXCGQAACoQzAAAUCGcAACgQzgAAUCCcAQCgQDgDAECBcAYAgALhDAAABcIZAAAKhDMAABQIZwAAKBDOAABQIJwBAKBAOAMAQIFwBgCAAuEMAAAFwhkAAAqEMwAAFAhnAAAoEM4AAFAgnAEAoEA4AwBAgXAGAIAC4QwAAAXCGQAACoQzAAAUCGcAACgQzgAAUCCcAQCgQDgDAECBcAYAgALhDAAABcIZAAAKhDMAABRMbfcEALbTxOREHpyZK48/6/jnN3E2AOxkwhk4o83MnsydHz5aHv9tT3lCJjZxPgDsXE7VAACAAuEMAAAFwhkAAAqEMwAAFAhnAAAoEM4AAFAgnAEAoEA4AwBAgXAGAIAC4QwAAAXCGQAACoQzAAAUCGcAACgQzgAAUCCcAQCgQDgDAECBcAYAgALhDAAABcIZAAAKhDMAABQIZwAAKBDOAABQIJwBAKBAOAMAQIFwBgCAAuEMAAAFwhkAAAqEMwAAFAhnAAAoEM4AAFAgnAEAoEA4AwBAgXAGAIAC4QwAAAXCGQAACoQzAAAUCGcAACgQzgAAUCCcAQCgQDgDAECBcAYAgALhDAAABcIZAAAKhDMAABQIZwAAKJiqDGqtvSzJDwwvvqP3/tOttcuSXJdkf5K39N6vHY69OMkbk5yT5L1JXtB7n9vwmQMAwBZa84jzMJC/K8nXJ7k4yVNaaz+U5E1Jnp3koiSXtNaeObzJDUle3Hs/lGQiyTWbMXEAANhKlVM1jiR5Se/987332SR3JzmU5CO993uHR5NvSHJ5a+2JSfb33m8b3vb6JJdvwrwBAGBLrXmqRu/9rsV/t9a+OoNTNn4zg6BedCTJ45I8doXtZeedd2DV66enD46zO/gS1tDpbeH+4zl44Kzy+L17p8Yan1hDrJ81xHpZQ9ujdI5zkrTWvjbJO5L8VJK5DI46L5pIMp/BEeyFZbaX3XffsczPLyx73fT0wRw9+sA4u4MvYg2d/o7PzOWBYyfK42dnxxufxBpiXTwOsV7W0PpNTk6sebB22dtVBrXWnprk5iQ/03t/c5JPJrlgZMhjknx6le0AALCrVZ4c+Pgkb0tyRe/9D4ab3z+4ql3YWtuT5IokN/XeP57kxDC0k+S5SW7ahHkDAMCWqpyq8dIkZyW5rrW2uO23k1yV5K3D696Z5MbhdVcmeUNr7Zwkf53kNRs4XwAA2BaVJwf+ZJKfXOHqw8uMvzPJpeucFwAA7CjeORAAAAqEMwAAFAhnAAAoEM4AAFAgnAEAoEA4AwBAgXAGAIAC4QwAAAXCGQAACoQzAAAUCGcAACgQzgAAUCCcAQCgQDgDAECBcAYAgALhDAAABcIZAAAKhDMAABRMbfcEgDPL3HwyMztXHr9v71Sm/IoPwA4gnIEtNTM7l9vv/kx5/CUXPTpT+zxUAbD9HMcBAIAC4QwAAAXCGQAACoQzAAAUCGcAACgQzgAAUCCcAQCgQDgDAECBcAYAgALhDAAABcIZAAAKhDMAABQIZwAAKBDOAABQIJwBAKBAOAMAQIFwBgCAAuEMAAAFwhkAAAqEMwAAFAhnAAAoEM4AAFAgnAEAoEA4AwBAgXAGAIAC4QwAAAXCGQAACoQzAAAUCGcAACgQzgAAUCCcAQCgQDgDAECBcAYAgALhDAAABcIZAAAKhDMAABQIZwAAKBDOAABQMLXdEwB2t7n5ZGZ2rjx+fmETJwMAm0g4A+syMzuX2+/+THn84UPTmzgbANg8TtUAAIAC4QwAAAXCGQAACoQzAAAUCGcAACgQzgAAUCCcAQCgQDgDAECBcAYAgALhDAAABcIZAAAKhDMAABQIZwAAKBDOAABQIJwBAKBAOAMAQIFwBgCAAuEMAAAFwhkAAAqEMwAAFAhnAAAoEM4AAFAwVR3YWjsnya1JntV7//vW2mVJrkuyP8lbeu/XDsddnOSNSc5J8t4kL+i9z234zAEAYAuVjji31r4xyfuSHBpe3p/kTUmeneSiJJe01p45HH5Dkhf33g8lmUhyzUZPGgAAtlr1VI1rkrwoyaeHly9N8pHe+73Do8k3JLm8tfbEJPt777cNx12f5PINnC8AAGyL0qkavferk6S1trjpsUmOjAw5kuRxq2wHAIBdrXyO8xKTSRZGLk8kmV9le9l55x1Y9frp6YPj7A6+hDW0sRbuP56DB84qj9+7d2qs8WefvS/T5569Y+aTWEOsnzXEellD2+NUw/mTSS4YufyYDE7jWGl72X33Hcv8/MKy101PH8zRow+MN1MYYQ1tvOMzc3ng2Iny+NnZ8cYfPz6ToydP7pj5JLGGWBePQ6yXNbR+k5MTax6sXfZ2p3h/70/SWmsXttb2JLkiyU29948nOdFae+pw3HOT3HSK9wEAADvGKYVz7/1EkquSvDXJh5Lck+TG4dVXJnl1a+2eJAeSvGb90wQAgO011qkavfevGPn3zUkOLzPmzgxedQMAAE4b3jkQAAAKhDMAABQIZwAAKBDOAABQIJwBAKBAOAMAQIFwBgCAAuEMAAAFwhkAAAqEMwAAFAhnAAAoEM4AAFAgnAEAoGBquycAwNaam09mZufK4/ftncqUwywAwhngTDMzO5fb7/5MefwlFz06U/v8uABwDAEAAAqEMwAAFAhnAAAoEM4AAFAgnAEAoEA4AwBAgXAGAIACL8wJPGTcN8ZIkvmFTZoMAOwwwhl4yLhvjJEkhw9Nb9JsAGBncaoGAAAUCGcAACgQzgAAUCCcAQCgQDgDAECBcAYAgALhDAAABcIZAAAKhDMAABQIZwAAKBDOAABQIJwBAKBAOAMAQIFwBgCAAuEMAAAFwhkAAAqmtnsCwOaZm09mZufK4+cXNnEyALDLCWc4jc3MzuX2uz9THn/40PQmzoaqcX/h2bd3KlP+fgiw6YQzwA4z7i88l1z06Ezt83AOsNkcowAAgAKHKAB2uYnJiTw441x2gM0mnAF2uZnZk7nzw0fL453LDnBqnKoBAAAFwhkAAAqEMwAAFAhnAAAoEM4AAFAgnAEAoMDL0QGMYe7kfD4/xmsmJ94SG+B0IZxhF5mbH7wdc5U3uth4M7Mn88Ex3g478ZbYAKcLj+Swi8zMzuX2MaLNG10AwMbxx0MAACgQzgAAUCCcAQCgQDgDAECBcAYAgALhDAAABcIZAAAKhDMAABQIZwAAKBDOAABQIJyH3nbLx1a9/oXX/cW69v/S1/3llt5uvX70le/esnm88vf+qjTu3y2Z00buezVLPxcbvf9x19bvvquXtv/uu3r++LaPj73P5fZ/x0f+8UvGL/43enn09ivt8yWved+q81h6+9/4b3es+jVIvvB1WPxc3vGRf1z287HS/Eat9LEu+on/9Oerzn3pfpfub9To99Xix/C2Wz6WH3/1e79oP0s/nhv//KN51/v/z0P3u/g5vfHPP7rmfSZf+Bos7mP0Yxj9/9J/J1/4flic++j3QPVxdKWv5zW/9p5Vb3+qqt+na62zjfKzr135e4CtNe5j+EY85m/Gvnaz7eqcUyWch97+l3+/6vUnPn9yXfu//3MzW3q7jbaZ8/jwJ/6pNG5hE/d9qjZi/+tdW6v5kw98YkP2878+et+G7OdU/O9PfS5J8uDM3Ir/ffgT/5QHZ+Zy4vMnM7+wvvlu9Me62v5Gv68W19Jyj0VL93H8xFz+4bP//CXjjp+YW/M+Ry23j6rFuY9+D6z3cfTk/Kl8l69tsx8HxvV3H9u+7ye+2LhrYyPX0k5bl9tlp3RO1dR2TwCg4va7P1O6fm5+fiumA8AZyBFnAAAoEM4AAFDgVA0AVjUxOZEHZ+Yeurz476X/H/33vr1TmXJoBjjNCGcAVjUzezJ3fvjoQ5cXzydf+v/Rf19y0aMztc+PGOD04lENgA239Cj1ouWOUieDI9QAO51HqjPY3HwyM/ulP9gWLf3BNjcff3oFSpYepV603FHqZHCEGmCnE85nsJnZuVVf4mvpdTOzc/70CgCcsRw/BACAAuEMAAAF/u4OwGlvred0LPJyesBqhDOsYm747s3LvTrAotHr/LCFnWmt53Qs8nJ6wGo8KsAqFo9QVZ9E6YctAJy+HBsDAIACh8bYMSrnIDotgt1opTcDSZZ/u2oAdibhPGKtH1qibXNVzkF0WgS70UpvBpIs/3bVZ6KJyYkk9V8ePAYD22FTqqO1dkWSa5PsTfLrvffXbsb9bDTRxlarPPlw1PzCJk4GttHM7Mkk9V8ePAYD22HDH3Vaa1+e5JeTPCXJTJJbW2vv6b1/aKPvazut9qfXjTB62kLlfnbi0Zfqyz8lg49x3Cisfg1GX15qp6k8+XDU4UPTmzkd2DXGfQze7F86x3m8SwaPSzvxcXu3G/frcCrGWXd7p6YyO1d7VaZkZ/4sH9dWfA2202aUxGVJ3t17vz9JWms3JnlOklescbs9STI5/HPdSta6/lSd/8j9OfusveXrT84v5O577x9r/+PM/eTcydx17/05/5H7c1fhfg5f+Kg8bGpPef9JMrVncsWPeenHe/4j92dqz+QpfQxrWfwYL/rKczf8azD6+Tt84aOSjLeGpvZMrro2ll437udorf2PM365r1mS0r4Xb7vS/lfa9+icFu9rcezobVb693LzWG4/a30co2OXzme5MaP/X/p9sNbncbXP/6LlvnfW+pot/XiX3udKl9f63K/1+Vrtc7Tc/Ff6PC79eJb7mFeb37jfC+M8Bp//yP3JxNrfC0vnvxmPd4v3c9e995/S4zarG+frkIz/s7n6M3nRRV957orrdLl9jbMmNquH1muzvwYbZeQ+x/omnFhY2Nhfw1trP5vky3rv1w4vX53k0t77j61x029JcsuGTgYAAFb2tCTvqw7ejCPOk0lGa3wiyXzhdrdnMPkjSU5uwrwAACAZHGm+IIP+LNuMcP5kBgG86DFJPl243UzGKH4AAFiHj457g80I5z9L8vLW2nSSB5N8f5K1TtMAAIAdbcOfu9l7/1SSn0vyniR3JPn93vsHNvp+AABgK234kwMBAOB0tMtfLRAAALaGcAYAgALhDAAABcIZAAAKhDMAABRsxus4b4rW2hVJrk2yN8mv995fu81TYgdprZ2T5NYkz+q9/31r7bIk1yXZn+QtI28Bf3GSNyY5J8l7k7yg9z7XWntCkhuSnJ+kJ7my935sGz4UtkFr7WVJfmB48R2995+2hhhHa+0VSZ6TwTvn/k7v/TpriFPRWntVkkf13q8ad6201h6R5PeSfFWSo0l+oPf+f7flAzlN7Yojzq21L0/yy0m+JcnFSX6stfY12zsrdorW2jdm8K6Th4aX9yd5U5JnJ7koySWttWcOh9+Q5MW990MZvB38NcPtr0vyut77k5J8MMnPb91HwHYaxs13Jfn6DB5fntJa+6FYQxS11r4tyXckeXKSb0jy4621w7GGGFNr7RlJfmRk07hr5ZeS3NJ7vyjJG5L8xpZM/AyyK8I5yWVJ3t17v7/3/mCSGzP4zR6SwQPJi/KFt3a/NMlHeu/39t7nMnjguby19sQk+3vvtw3HXT/cvjfJt2awrh7avkVzZ/sdSfKS3vvne++zSe7O4Jcwa4iS3vtfJPn24Vo5P4O/5j4i1hBjaK2dm8FBwl8ZXj6VtfK9GRxxTpL/muSZw/FskN0Szo/N4IfboiNJHrdNc2GH6b1f3Xu/ZWTTSutlpe2PSvK54Q+30e2cAXrvdy3+YGqtfXUGp2zMxxpiDL332dbaLyb5UJKb43GI8b0+g3de/uzw8qmslYduM7z+c0mmN3faZ5bdEs6TGZw3tmgigx9ssJyV1kt1e2J9nXFaa1+b5E+T/FSSj8UaYky995dlECmPz+CvFtYQJa21q5N8ovd+88jmU1krE0u266UNtlvC+ZNJLhi5/Jh84c/ysNRK62Wl7f+Q5OGttT3D7RfE+jqjtNaemsFRwp/pvb851hBjaK09afgkrvTejyf5wyRPjzVE3Q8m+a7W2h1JXpHk+5JcnfHXyqeG49Jam0pyMMl9mz77M8huCec/S/KM1tp0a+3sJN+f5F3bPCd2rvcnaa21C4cPLFckuan3/vEkJ4aRlCTPHW6fTXJLBg9cSfLDSW7a6kmzPVprj0/ytiRX9N7/YLjZGmIcX5XkDa21fa21h2XwhMDXxxqiqPf+nb33r+u9X5zkF5K8vff+vIy/Vt45vJzh9bcMx7NBdkU4994/lcF5P+9JckeS3++9f2B7Z8VO1Xs/keSqJG/N4HzDe/KFJ1FcmeTVrbV7khxI8prh9hdm8GotH0rytAxe+pAzw0uTnJXkutbaHcMjPlfFGqKo9/7OJO9I8jdJ/irJrcNfwq6KNcT6jLtWfj7JN7XW7hqOedEWz/e0N7GwsPQ0GQAAYKldccQZAAC2m3AGAIAC4QwAAAXCGQAACoQzAAAUCGeAHWr48niPaK09vLX27sL4q1pr/7O6HYDxTG33BABY3vDNENJa+4okl27vbAAQzgAbrLX2o0lekuRkkn9M8rwk/z7JN2XwFrgTSa7uvf9la+36JP+c5OIk5yf5kyQ/0Xufba0tJJlO8l+S7B++OctTkvxIkucneViSc5O8svf+W8W5PS7JbyX5iuE83tx7/4/Dt+f9zSRPTTKb5GPDeZ9Ybnvv/dipfn4AdiunagBsoNba4SS/muR7eu9PTvL2JP85yWOTfHPv/WuSvDnJz4zc7BuTfGeSrxn+9/wlu31ekn8eHoHen+SaJP+q9/71Gbyt7q+NMcXfS/Ke3vu/yCCG/21r7d8k+eYkT09yuPf+lAwC+cmrbAc44whngI31jCR/3Hv/RJL03n+99/7dGbwl7vNba69K8pwM3j530fW992O995kkv5vku1fa+fBI77OSfG9r7T8k+bkl+1pRa+3LMojl1w739U9Jrk/yzCR/m8ER8vcP9/vW3vutq2wHOOMIZ4CNNZdkYfFCa21/a+1FSd4x3PQ/kvx2BqdJjN5m0WQGobqs4akWdyR5YpL3ZRDkVZNL7ndx297e+/9LcjjJS4f3/5bW2gtX2j7GfQKcNoQzwMZ6T5LLWmsXDC8/P4MjyH80PA/5g0n+dZI9I7f5wdbavtbaWRmcv/xHS/Y5l2RPa20iyTckOZrklzI4H/pZSdJa25M19N4fSHJbkhcNb/PwJD+c5E9ba89KcnOSW3vvL8/gyPclK20vfzYATiPCGWAD9d7/NslPJXlXa+3OJN+T5JVJnt5a+9skf53ko0m+srW2+Bh8PMktGZwWcUsGTwYcdSTJB5LcleT2JJ9M0pPcneQJGYT0haM3aK19X2vtnctM8cokzxjO5QNJ/jCD0zVuGu7/71prH0zyL5P84irbAc44EwsLC2uPAmBTDF9V4+9676/a7rkAsDpHnAEAoMARZwAAKHDEGQAACoQzAAAUCGcAACgQzgAAUCCcAQCg4P8DB8yPp8/XXiQAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 864x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12,8))\n",
    "sns.distplot(data[(data['capital.loss']!=0)]['capital.loss'], kde=False,rug=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[Text(0, 0.5, 'Fraction of people with income > $50k')]"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(20,6))\n",
    "ax=sns.barplot(x='occupation', y='income', data=data)\n",
    "ax.set(ylabel='Fraction of people with income > $50k')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "United-States                 29170\n",
      "Mexico                          643\n",
      "?                               583\n",
      "Philippines                     198\n",
      "Germany                         137\n",
      "Canada                          121\n",
      "Puerto-Rico                     114\n",
      "El-Salvador                     106\n",
      "India                           100\n",
      "Cuba                             95\n",
      "England                          90\n",
      "Jamaica                          81\n",
      "South                            80\n",
      "China                            75\n",
      "Italy                            73\n",
      "Dominican-Republic               70\n",
      "Vietnam                          67\n",
      "Guatemala                        64\n",
      "Japan                            62\n",
      "Poland                           60\n",
      "Columbia                         59\n",
      "Taiwan                           51\n",
      "Haiti                            44\n",
      "Iran                             43\n",
      "Portugal                         37\n",
      "Nicaragua                        34\n",
      "Peru                             31\n",
      "Greece                           29\n",
      "France                           29\n",
      "Ecuador                          28\n",
      "Ireland                          24\n",
      "Hong                             20\n",
      "Cambodia                         19\n",
      "Trinadad&Tobago                  19\n",
      "Thailand                         18\n",
      "Laos                             18\n",
      "Yugoslavia                       16\n",
      "Outlying-US(Guam-USVI-etc)       14\n",
      "Honduras                         13\n",
      "Hungary                          13\n",
      "Scotland                         12\n",
      "Holand-Netherlands                1\n",
      "Name: native.country, dtype: int64\n",
      "3391 people are not from the United States\n"
     ]
    }
   ],
   "source": [
    "print(data['native.country'].value_counts())\n",
    "not_from_US = np.sum(data['native.country']!='United-States')\n",
    "print(not_from_US, 'people are not from the United States')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Convert the `native.country` feature to binary since there is a huge imbalance in this feature"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "data['native.country'] = (data['native.country']=='United-States')*1\n",
    "#data['US_or_not']=np.where(data['native.country']=='United-States',1,0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "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>workclass</th>\n",
       "      <th>education</th>\n",
       "      <th>marital.status</th>\n",
       "      <th>occupation</th>\n",
       "      <th>relationship</th>\n",
       "      <th>race</th>\n",
       "      <th>sex</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>?</td>\n",
       "      <td>HS-grad</td>\n",
       "      <td>Widowed</td>\n",
       "      <td>?</td>\n",
       "      <td>Not-in-family</td>\n",
       "      <td>White</td>\n",
       "      <td>Female</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Private</td>\n",
       "      <td>HS-grad</td>\n",
       "      <td>Widowed</td>\n",
       "      <td>Exec-managerial</td>\n",
       "      <td>Not-in-family</td>\n",
       "      <td>White</td>\n",
       "      <td>Female</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>?</td>\n",
       "      <td>Some-college</td>\n",
       "      <td>Widowed</td>\n",
       "      <td>?</td>\n",
       "      <td>Unmarried</td>\n",
       "      <td>Black</td>\n",
       "      <td>Female</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Private</td>\n",
       "      <td>7th-8th</td>\n",
       "      <td>Divorced</td>\n",
       "      <td>Machine-op-inspct</td>\n",
       "      <td>Unmarried</td>\n",
       "      <td>White</td>\n",
       "      <td>Female</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Private</td>\n",
       "      <td>Some-college</td>\n",
       "      <td>Separated</td>\n",
       "      <td>Prof-specialty</td>\n",
       "      <td>Own-child</td>\n",
       "      <td>White</td>\n",
       "      <td>Female</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  workclass     education marital.status         occupation   relationship  \\\n",
       "0         ?       HS-grad        Widowed                  ?  Not-in-family   \n",
       "1   Private       HS-grad        Widowed    Exec-managerial  Not-in-family   \n",
       "2         ?  Some-college        Widowed                  ?      Unmarried   \n",
       "3   Private       7th-8th       Divorced  Machine-op-inspct      Unmarried   \n",
       "4   Private  Some-college      Separated     Prof-specialty      Own-child   \n",
       "\n",
       "    race     sex  \n",
       "0  White  Female  \n",
       "1  White  Female  \n",
       "2  Black  Female  \n",
       "3  White  Female  \n",
       "4  White  Female  "
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.select_dtypes(exclude=[np.number]).head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Replace all '?'s with NaNs.\n",
    "data = data.applymap(lambda x: np.nan if x=='?' else x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "data.isnull().sum(axis=0) # How many issing values are there in the dataset?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "data.shape[0] - data.dropna(axis=0).shape[0]   # how many rows will be removed if I remove all the NaN's?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = data.dropna(axis=0) ## Drop all the NaNs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#data.education.value_counts()  # I will label-encode the education column since it is an ordinal categorical variable"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "## This computes the fraction of people by country who earn >50k per annum\n",
    "#mean_income_bycountry_df = data[['native.country','income']].groupby(['native.country']).mean().reset_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#edu_encode_dict = {'Preschool':0,'1st-4th':1, '5th-6th':2, '7th-8th':3, '9th':4, '10th':5,\n",
    "#                  '11th':6, '12th':7, 'HS-grad':8, 'Some-college':9, 'Bachelors':10, 'Masters':11, 'Assoc-voc':12, \n",
    "#                   'Assoc-acdm':13, 'Doctorate':14, 'Prof-school':15}\n",
    "\n",
    "#data['education'] = data['education'].apply(lambda ed_level: edu_encode_dict[ed_level])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### One-hot encoding of the categorical columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = pd.get_dummies(data,columns=['workclass','sex', 'marital.status',\n",
    "                                    'race','relationship','occupation'],\n",
    "               prefix=['workclass', 'is', 'is', 'race_is', 'relation', 'is'], drop_first=True)\n",
    "### native country is ignored because that feature will be dropped later"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x1a25433c18>"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x864 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(20,12))\n",
    "sns.heatmap(data.corr())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(32561, 1)"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.select_dtypes(exclude=[np.number]).shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "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>age</th>\n",
       "      <th>fnlwgt</th>\n",
       "      <th>education.num</th>\n",
       "      <th>capital.gain</th>\n",
       "      <th>capital.loss</th>\n",
       "      <th>hours.per.week</th>\n",
       "      <th>native.country</th>\n",
       "      <th>workclass_Local-gov</th>\n",
       "      <th>workclass_Never-worked</th>\n",
       "      <th>workclass_Private</th>\n",
       "      <th>...</th>\n",
       "      <th>is_Farming-fishing</th>\n",
       "      <th>is_Handlers-cleaners</th>\n",
       "      <th>is_Machine-op-inspct</th>\n",
       "      <th>is_Other-service</th>\n",
       "      <th>is_Priv-house-serv</th>\n",
       "      <th>is_Prof-specialty</th>\n",
       "      <th>is_Protective-serv</th>\n",
       "      <th>is_Sales</th>\n",
       "      <th>is_Tech-support</th>\n",
       "      <th>is_Transport-moving</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>income</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>36.783738</td>\n",
       "      <td>190340.86517</td>\n",
       "      <td>9.595065</td>\n",
       "      <td>148.752468</td>\n",
       "      <td>53.142921</td>\n",
       "      <td>38.840210</td>\n",
       "      <td>0.889927</td>\n",
       "      <td>0.059709</td>\n",
       "      <td>0.000283</td>\n",
       "      <td>0.717354</td>\n",
       "      <td>...</td>\n",
       "      <td>0.035558</td>\n",
       "      <td>0.051942</td>\n",
       "      <td>0.070874</td>\n",
       "      <td>0.127751</td>\n",
       "      <td>0.005987</td>\n",
       "      <td>0.092273</td>\n",
       "      <td>0.017718</td>\n",
       "      <td>0.107888</td>\n",
       "      <td>0.026092</td>\n",
       "      <td>0.051659</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>44.249841</td>\n",
       "      <td>188005.00000</td>\n",
       "      <td>11.611657</td>\n",
       "      <td>4006.142456</td>\n",
       "      <td>195.001530</td>\n",
       "      <td>45.473026</td>\n",
       "      <td>0.914552</td>\n",
       "      <td>0.078689</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.632955</td>\n",
       "      <td>...</td>\n",
       "      <td>0.014666</td>\n",
       "      <td>0.010968</td>\n",
       "      <td>0.031884</td>\n",
       "      <td>0.017472</td>\n",
       "      <td>0.000128</td>\n",
       "      <td>0.237087</td>\n",
       "      <td>0.026910</td>\n",
       "      <td>0.125367</td>\n",
       "      <td>0.036092</td>\n",
       "      <td>0.040811</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2 rows × 43 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "              age        fnlwgt  education.num  capital.gain  capital.loss  \\\n",
       "income                                                                       \n",
       "0       36.783738  190340.86517       9.595065    148.752468     53.142921   \n",
       "1       44.249841  188005.00000      11.611657   4006.142456    195.001530   \n",
       "\n",
       "        hours.per.week  native.country  workclass_Local-gov  \\\n",
       "income                                                        \n",
       "0            38.840210        0.889927             0.059709   \n",
       "1            45.473026        0.914552             0.078689   \n",
       "\n",
       "        workclass_Never-worked  workclass_Private  ...  is_Farming-fishing  \\\n",
       "income                                             ...                       \n",
       "0                     0.000283           0.717354  ...            0.035558   \n",
       "1                     0.000000           0.632955  ...            0.014666   \n",
       "\n",
       "        is_Handlers-cleaners  is_Machine-op-inspct  is_Other-service  \\\n",
       "income                                                                 \n",
       "0                   0.051942              0.070874          0.127751   \n",
       "1                   0.010968              0.031884          0.017472   \n",
       "\n",
       "        is_Priv-house-serv  is_Prof-specialty  is_Protective-serv  is_Sales  \\\n",
       "income                                                                        \n",
       "0                 0.005987           0.092273            0.017718  0.107888   \n",
       "1                 0.000128           0.237087            0.026910  0.125367   \n",
       "\n",
       "        is_Tech-support  is_Transport-moving  \n",
       "income                                        \n",
       "0              0.026092             0.051659  \n",
       "1              0.036092             0.040811  \n",
       "\n",
       "[2 rows x 43 columns]"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.groupby('income').mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(32561, 45)"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "y = data.income\n",
    "X = data.drop(['income', 'education', 'native.country', 'fnlwgt'],axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- `income` is dropped from X because it is the target variable.\n",
    "- `Education` is dropped because it is already label-encoded in `education.num`. One can notice the high correlation between `education` and `education.num` in the heatmap.\n",
    "- `native country` is dropped because it showed very little feature importance in random forest classifer.\n",
    "- `fnlwgt` is dropped because it has no correlation with `income`."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Modelling\n",
    "This section explores different classification algorithms to maximise the accuracy for predicting income of a person (> 50k/yr or < 50k/yr)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Split the dataset into training and testing set\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.ensemble import ExtraTreesClassifier\n",
    "from xgboost import XGBClassifier as xgb\n",
    "from sklearn import metrics"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Baseline model\n",
    "In the baseline model, we predict the minority class for all our train and test (or validation) examples. The resulting accuracy will serve as a benchmark for the ML models. In other words, the sophisticated ML models should have an accuracy which should at least better the baseline one."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy on train data: 0.758643%\n",
      "Accuracy on test data: 0.760467%\n"
     ]
    }
   ],
   "source": [
    "baseline_train = np.zeros(y_train.shape[0])\n",
    "baseline_test = np.zeros(y_test.shape[0])\n",
    "print('Accuracy on train data: %f%%' % (metrics.accuracy_score(y_train, baseline_train)))\n",
    "print('Accuracy on test data: %f%%' %  (metrics.accuracy_score(y_test, baseline_test)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Random Forest classifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RandomForestClassifier(bootstrap=True, class_weight=None, criterion='gini',\n",
      "                       max_depth=10, max_features='auto', max_leaf_nodes=None,\n",
      "                       min_impurity_decrease=0.0, min_impurity_split=None,\n",
      "                       min_samples_leaf=1, min_samples_split=5,\n",
      "                       min_weight_fraction_leaf=0.0, n_estimators=300,\n",
      "                       n_jobs=None, oob_score=True, random_state=10, verbose=0,\n",
      "                       warm_start=False)\n"
     ]
    }
   ],
   "source": [
    "rfmodel = RandomForestClassifier(n_estimators=300,oob_score=True,min_samples_split=5, max_depth=10,random_state=10)\n",
    "rfmodel.fit(X_train,y_train)\n",
    "print(rfmodel)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "def show_classifier_metrics(clf, y_train=y_train,y_test=y_test, print_classification_report=True, print_confusion_matrix=True):\n",
    "    print(clf)\n",
    "    if print_confusion_matrix:\n",
    "        print('confusion matrix of training data')\n",
    "        print(metrics.confusion_matrix(y_train, clf.predict(X_train)))\n",
    "        print('confusion matrix of test data')\n",
    "        print(metrics.confusion_matrix(y_test, clf.predict(X_test)))\n",
    "    if print_classification_report:\n",
    "        print('classification report of test data')\n",
    "        print(metrics.classification_report(y_test, clf.predict(X_test)))\n",
    "    print('Accuracy on test data: %f%%' % (metrics.accuracy_score(y_test, clf.predict(X_test))*100))\n",
    "    print('Accuracy on training data: %f%%' % (metrics.accuracy_score(y_train, clf.predict(X_train))*100))\n",
    "    print('Area under the ROC curve : %f' % (metrics.roc_auc_score(y_test, clf.predict(X_test))))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RandomForestClassifier(bootstrap=True, class_weight=None, criterion='gini',\n",
      "                       max_depth=10, max_features='auto', max_leaf_nodes=None,\n",
      "                       min_impurity_decrease=0.0, min_impurity_split=None,\n",
      "                       min_samples_leaf=1, min_samples_split=5,\n",
      "                       min_weight_fraction_leaf=0.0, n_estimators=300,\n",
      "                       n_jobs=None, oob_score=True, random_state=10, verbose=0,\n",
      "                       warm_start=False)\n",
      "confusion matrix of training data\n",
      "[[16610   681]\n",
      " [ 2333  3168]]\n",
      "confusion matrix of test data\n",
      "[[7125  304]\n",
      " [1072 1268]]\n",
      "classification report of test data\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.87      0.96      0.91      7429\n",
      "           1       0.81      0.54      0.65      2340\n",
      "\n",
      "    accuracy                           0.86      9769\n",
      "   macro avg       0.84      0.75      0.78      9769\n",
      "weighted avg       0.85      0.86      0.85      9769\n",
      "\n",
      "Accuracy on test data: 85.914628%\n",
      "Accuracy on training data: 86.776062%\n",
      "Area under the ROC curve : 0.750480\n",
      "oob score = 0.859907\n"
     ]
    }
   ],
   "source": [
    "show_classifier_metrics(rfmodel,y_train)\n",
    "print('oob score = %f'% rfmodel.oob_score_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "importance_list = rfmodel.feature_importances_\n",
    "name_list = X_train.columns\n",
    "importance_list, name_list = zip(*sorted(zip(importance_list, name_list)))\n",
    "plt.figure(figsize=(20,10))\n",
    "plt.barh(range(len(name_list)),importance_list,align='center')\n",
    "plt.yticks(range(len(name_list)),name_list)\n",
    "plt.xlabel('Relative Importance in the Random Forest')\n",
    "plt.ylabel('Features')\n",
    "plt.title('Relative importance of Each Feature')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Random forest: Grid Search and cross-validation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import cross_val_score, GridSearchCV"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "def grid_search(clf, parameters, X, y, n_jobs= -1, n_folds=4, score_func=None):\n",
    "    if score_func:\n",
    "        gs = GridSearchCV(clf, param_grid=parameters, cv=n_folds, n_jobs=n_jobs, scoring=score_func,verbose =2)\n",
    "    else:\n",
    "        print('Doing grid search')\n",
    "        gs = GridSearchCV(clf, param_grid=parameters, n_jobs=n_jobs, cv=n_folds, verbose =2)\n",
    "    gs.fit(X, y)\n",
    "    print(\"mean test score (weighted by split size) of CV rounds: \",gs.cv_results_['mean_test_score'] )\n",
    "    print (\"Best parameter set\", gs.best_params_, \"Corresponding mean CV score\",gs.best_score_)\n",
    "    best = gs.best_estimator_\n",
    "    return best"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Doing grid search\n",
      "Fitting 4 folds for each of 12 candidates, totalling 48 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 4 concurrent workers.\n",
      "[Parallel(n_jobs=-1)]: Done  33 tasks      | elapsed:   45.1s\n",
      "[Parallel(n_jobs=-1)]: Done  48 out of  48 | elapsed:  1.1min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "mean test score (weighted by split size) of CV rounds:  [0.85990698 0.85995086 0.85968761 0.86113549 0.86157424 0.86104774\n",
      " 0.86161811 0.86205686 0.86267111 0.86038961 0.86188136 0.86166199]\n",
      "Best parameter set {'max_depth': 15, 'n_estimators': 300} Corresponding mean CV score 0.8626711126711126\n"
     ]
    }
   ],
   "source": [
    "rfmodel2 = RandomForestClassifier(min_samples_split=5,oob_score=True, n_jobs=-1,random_state=10)\n",
    "parameters = {'n_estimators': [100,200,300], 'max_depth': [10,13,15,20]}\n",
    "rfmodelCV = grid_search(rfmodel2, parameters,X_train,y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RandomForestClassifier(bootstrap=True, class_weight=None, criterion='gini',\n",
      "                       max_depth=15, max_features='auto', max_leaf_nodes=None,\n",
      "                       min_impurity_decrease=0.0, min_impurity_split=None,\n",
      "                       min_samples_leaf=1, min_samples_split=5,\n",
      "                       min_weight_fraction_leaf=0.0, n_estimators=300,\n",
      "                       n_jobs=-1, oob_score=True, random_state=10, verbose=0,\n",
      "                       warm_start=False)\n",
      "confusion matrix of training data\n",
      "[[16686   605]\n",
      " [ 1995  3506]]\n",
      "confusion matrix of test data\n",
      "[[7103  326]\n",
      " [1017 1323]]\n",
      "classification report of test data\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.87      0.96      0.91      7429\n",
      "           1       0.80      0.57      0.66      2340\n",
      "\n",
      "    accuracy                           0.86      9769\n",
      "   macro avg       0.84      0.76      0.79      9769\n",
      "weighted avg       0.86      0.86      0.85      9769\n",
      "\n",
      "Accuracy on test data: 86.252431%\n",
      "Accuracy on training data: 88.592489%\n",
      "Area under the ROC curve : 0.760751\n",
      "oob score = 0.863285\n"
     ]
    }
   ],
   "source": [
    "rfmodelCV.fit(X_train,y_train)\n",
    "show_classifier_metrics(rfmodelCV,y_train)\n",
    "print('oob score = %f'% rfmodelCV.oob_score_)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## XGBoost "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [],
   "source": [
    "from xgboost.sklearn import XGBClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0]\tvalidation_0-error:0.175325\tvalidation_1-error:0.178831\n",
      "Multiple eval metrics have been passed: 'validation_1-error' will be used for early stopping.\n",
      "\n",
      "Will train until validation_1-error hasn't improved in 30 rounds.\n",
      "[1]\tvalidation_0-error:0.144963\tvalidation_1-error:0.145665\n",
      "[2]\tvalidation_0-error:0.150623\tvalidation_1-error:0.152523\n",
      "[3]\tvalidation_0-error:0.144568\tvalidation_1-error:0.146074\n",
      "[4]\tvalidation_0-error:0.143384\tvalidation_1-error:0.143618\n",
      "[5]\tvalidation_0-error:0.143384\tvalidation_1-error:0.144334\n",
      "[6]\tvalidation_0-error:0.14334\tvalidation_1-error:0.143822\n",
      "[7]\tvalidation_0-error:0.143077\tvalidation_1-error:0.144334\n",
      "[8]\tvalidation_0-error:0.143164\tvalidation_1-error:0.144436\n",
      "[9]\tvalidation_0-error:0.14334\tvalidation_1-error:0.144641\n",
      "[10]\tvalidation_0-error:0.143252\tvalidation_1-error:0.144436\n",
      "[11]\tvalidation_0-error:0.14391\tvalidation_1-error:0.144948\n",
      "[12]\tvalidation_0-error:0.142857\tvalidation_1-error:0.144232\n",
      "[13]\tvalidation_0-error:0.142682\tvalidation_1-error:0.144334\n",
      "[14]\tvalidation_0-error:0.140444\tvalidation_1-error:0.143106\n",
      "[15]\tvalidation_0-error:0.140049\tvalidation_1-error:0.142901\n",
      "[16]\tvalidation_0-error:0.139391\tvalidation_1-error:0.143003\n",
      "[17]\tvalidation_0-error:0.139259\tvalidation_1-error:0.142799\n",
      "[18]\tvalidation_0-error:0.139303\tvalidation_1-error:0.142594\n",
      "[19]\tvalidation_0-error:0.139259\tvalidation_1-error:0.142799\n",
      "[20]\tvalidation_0-error:0.138996\tvalidation_1-error:0.142696\n",
      "[21]\tvalidation_0-error:0.139347\tvalidation_1-error:0.142696\n",
      "[22]\tvalidation_0-error:0.138996\tvalidation_1-error:0.142492\n",
      "[23]\tvalidation_0-error:0.138996\tvalidation_1-error:0.142696\n",
      "[24]\tvalidation_0-error:0.138294\tvalidation_1-error:0.142184\n",
      "[25]\tvalidation_0-error:0.137461\tvalidation_1-error:0.141673\n",
      "[26]\tvalidation_0-error:0.137724\tvalidation_1-error:0.14198\n",
      "[27]\tvalidation_0-error:0.137724\tvalidation_1-error:0.141877\n",
      "[28]\tvalidation_0-error:0.136802\tvalidation_1-error:0.141468\n",
      "[29]\tvalidation_0-error:0.136978\tvalidation_1-error:0.141263\n",
      "[30]\tvalidation_0-error:0.136539\tvalidation_1-error:0.141468\n",
      "[31]\tvalidation_0-error:0.136188\tvalidation_1-error:0.141161\n",
      "[32]\tvalidation_0-error:0.136057\tvalidation_1-error:0.141263\n",
      "[33]\tvalidation_0-error:0.136232\tvalidation_1-error:0.141161\n",
      "[34]\tvalidation_0-error:0.135662\tvalidation_1-error:0.141058\n",
      "[35]\tvalidation_0-error:0.135486\tvalidation_1-error:0.140751\n",
      "[36]\tvalidation_0-error:0.135267\tvalidation_1-error:0.141058\n",
      "[37]\tvalidation_0-error:0.135311\tvalidation_1-error:0.140956\n",
      "[38]\tvalidation_0-error:0.135179\tvalidation_1-error:0.140649\n",
      "[39]\tvalidation_0-error:0.135398\tvalidation_1-error:0.140137\n",
      "[40]\tvalidation_0-error:0.134345\tvalidation_1-error:0.139625\n",
      "[41]\tvalidation_0-error:0.134038\tvalidation_1-error:0.139932\n",
      "[42]\tvalidation_0-error:0.133819\tvalidation_1-error:0.139523\n",
      "[43]\tvalidation_0-error:0.134038\tvalidation_1-error:0.139421\n",
      "[44]\tvalidation_0-error:0.133994\tvalidation_1-error:0.139523\n",
      "[45]\tvalidation_0-error:0.133643\tvalidation_1-error:0.139318\n",
      "[46]\tvalidation_0-error:0.133468\tvalidation_1-error:0.139318\n",
      "[47]\tvalidation_0-error:0.133292\tvalidation_1-error:0.139114\n",
      "[48]\tvalidation_0-error:0.132985\tvalidation_1-error:0.138704\n",
      "[49]\tvalidation_0-error:0.132854\tvalidation_1-error:0.138602\n",
      "[50]\tvalidation_0-error:0.13259\tvalidation_1-error:0.138499\n",
      "[51]\tvalidation_0-error:0.131976\tvalidation_1-error:0.137988\n",
      "[52]\tvalidation_0-error:0.131581\tvalidation_1-error:0.137783\n",
      "[53]\tvalidation_0-error:0.131713\tvalidation_1-error:0.137578\n",
      "[54]\tvalidation_0-error:0.131274\tvalidation_1-error:0.136862\n",
      "[55]\tvalidation_0-error:0.130923\tvalidation_1-error:0.136554\n",
      "[56]\tvalidation_0-error:0.130704\tvalidation_1-error:0.136657\n",
      "[57]\tvalidation_0-error:0.130309\tvalidation_1-error:0.137169\n",
      "[58]\tvalidation_0-error:0.130265\tvalidation_1-error:0.137271\n",
      "[59]\tvalidation_0-error:0.130397\tvalidation_1-error:0.136554\n",
      "[60]\tvalidation_0-error:0.130265\tvalidation_1-error:0.135633\n",
      "[61]\tvalidation_0-error:0.129739\tvalidation_1-error:0.135531\n",
      "[62]\tvalidation_0-error:0.1293\tvalidation_1-error:0.135633\n",
      "[63]\tvalidation_0-error:0.128861\tvalidation_1-error:0.135428\n",
      "[64]\tvalidation_0-error:0.128993\tvalidation_1-error:0.135326\n",
      "[65]\tvalidation_0-error:0.128905\tvalidation_1-error:0.135326\n",
      "[66]\tvalidation_0-error:0.128642\tvalidation_1-error:0.135531\n",
      "[67]\tvalidation_0-error:0.128598\tvalidation_1-error:0.135224\n",
      "[68]\tvalidation_0-error:0.127764\tvalidation_1-error:0.134814\n",
      "[69]\tvalidation_0-error:0.12772\tvalidation_1-error:0.134405\n",
      "[70]\tvalidation_0-error:0.127632\tvalidation_1-error:0.134302\n",
      "[71]\tvalidation_0-error:0.127457\tvalidation_1-error:0.133176\n",
      "[72]\tvalidation_0-error:0.12715\tvalidation_1-error:0.133074\n",
      "[73]\tvalidation_0-error:0.126623\tvalidation_1-error:0.132767\n",
      "[74]\tvalidation_0-error:0.126536\tvalidation_1-error:0.132153\n",
      "[75]\tvalidation_0-error:0.126141\tvalidation_1-error:0.131641\n",
      "[76]\tvalidation_0-error:0.125877\tvalidation_1-error:0.131641\n",
      "[77]\tvalidation_0-error:0.125921\tvalidation_1-error:0.130208\n",
      "[78]\tvalidation_0-error:0.125877\tvalidation_1-error:0.131231\n",
      "[79]\tvalidation_0-error:0.125483\tvalidation_1-error:0.13072\n",
      "[80]\tvalidation_0-error:0.125483\tvalidation_1-error:0.130515\n",
      "[81]\tvalidation_0-error:0.125395\tvalidation_1-error:0.130515\n",
      "[82]\tvalidation_0-error:0.124956\tvalidation_1-error:0.129798\n",
      "[83]\tvalidation_0-error:0.125\tvalidation_1-error:0.129287\n",
      "[84]\tvalidation_0-error:0.125088\tvalidation_1-error:0.129594\n",
      "[85]\tvalidation_0-error:0.124868\tvalidation_1-error:0.129594\n",
      "[86]\tvalidation_0-error:0.124825\tvalidation_1-error:0.129184\n",
      "[87]\tvalidation_0-error:0.124605\tvalidation_1-error:0.129184\n",
      "[88]\tvalidation_0-error:0.124693\tvalidation_1-error:0.128775\n",
      "[89]\tvalidation_0-error:0.124386\tvalidation_1-error:0.128775\n",
      "[90]\tvalidation_0-error:0.124386\tvalidation_1-error:0.128775\n",
      "[91]\tvalidation_0-error:0.123903\tvalidation_1-error:0.12857\n",
      "[92]\tvalidation_0-error:0.123815\tvalidation_1-error:0.128468\n",
      "[93]\tvalidation_0-error:0.123771\tvalidation_1-error:0.128365\n",
      "[94]\tvalidation_0-error:0.123508\tvalidation_1-error:0.128365\n",
      "[95]\tvalidation_0-error:0.123377\tvalidation_1-error:0.128263\n",
      "[96]\tvalidation_0-error:0.123289\tvalidation_1-error:0.128263\n",
      "[97]\tvalidation_0-error:0.123157\tvalidation_1-error:0.128161\n",
      "[98]\tvalidation_0-error:0.122938\tvalidation_1-error:0.127444\n",
      "[99]\tvalidation_0-error:0.122762\tvalidation_1-error:0.127649\n",
      "[100]\tvalidation_0-error:0.122894\tvalidation_1-error:0.127342\n",
      "[101]\tvalidation_0-error:0.122631\tvalidation_1-error:0.127035\n",
      "[102]\tvalidation_0-error:0.122631\tvalidation_1-error:0.127239\n",
      "[103]\tvalidation_0-error:0.122455\tvalidation_1-error:0.127239\n",
      "[104]\tvalidation_0-error:0.122499\tvalidation_1-error:0.127137\n",
      "[105]\tvalidation_0-error:0.122587\tvalidation_1-error:0.127035\n",
      "[106]\tvalidation_0-error:0.122367\tvalidation_1-error:0.127239\n",
      "[107]\tvalidation_0-error:0.122367\tvalidation_1-error:0.127342\n",
      "[108]\tvalidation_0-error:0.122192\tvalidation_1-error:0.127035\n",
      "[109]\tvalidation_0-error:0.122104\tvalidation_1-error:0.127137\n",
      "[110]\tvalidation_0-error:0.122016\tvalidation_1-error:0.127035\n",
      "[111]\tvalidation_0-error:0.122192\tvalidation_1-error:0.127035\n",
      "[112]\tvalidation_0-error:0.12206\tvalidation_1-error:0.126932\n",
      "[113]\tvalidation_0-error:0.121973\tvalidation_1-error:0.126727\n",
      "[114]\tvalidation_0-error:0.121885\tvalidation_1-error:0.126625\n",
      "[115]\tvalidation_0-error:0.121885\tvalidation_1-error:0.126625\n",
      "[116]\tvalidation_0-error:0.121841\tvalidation_1-error:0.12642\n",
      "[117]\tvalidation_0-error:0.121578\tvalidation_1-error:0.126318\n",
      "[118]\tvalidation_0-error:0.121622\tvalidation_1-error:0.12642\n",
      "[119]\tvalidation_0-error:0.121666\tvalidation_1-error:0.125908\n",
      "[120]\tvalidation_0-error:0.121753\tvalidation_1-error:0.125908\n",
      "[121]\tvalidation_0-error:0.121578\tvalidation_1-error:0.125704\n",
      "[122]\tvalidation_0-error:0.121622\tvalidation_1-error:0.125601\n",
      "[123]\tvalidation_0-error:0.12092\tvalidation_1-error:0.126216\n",
      "[124]\tvalidation_0-error:0.120963\tvalidation_1-error:0.126216\n",
      "[125]\tvalidation_0-error:0.120876\tvalidation_1-error:0.12642\n",
      "[126]\tvalidation_0-error:0.1207\tvalidation_1-error:0.126523\n",
      "[127]\tvalidation_0-error:0.120569\tvalidation_1-error:0.12642\n",
      "[128]\tvalidation_0-error:0.120525\tvalidation_1-error:0.126216\n",
      "[129]\tvalidation_0-error:0.120481\tvalidation_1-error:0.126216\n",
      "[130]\tvalidation_0-error:0.120525\tvalidation_1-error:0.126113\n",
      "[131]\tvalidation_0-error:0.120437\tvalidation_1-error:0.126113\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[132]\tvalidation_0-error:0.120349\tvalidation_1-error:0.126113\n",
      "[133]\tvalidation_0-error:0.120305\tvalidation_1-error:0.126011\n",
      "[134]\tvalidation_0-error:0.120218\tvalidation_1-error:0.126113\n",
      "[135]\tvalidation_0-error:0.120174\tvalidation_1-error:0.126216\n",
      "[136]\tvalidation_0-error:0.120042\tvalidation_1-error:0.126523\n",
      "[137]\tvalidation_0-error:0.119779\tvalidation_1-error:0.12642\n",
      "[138]\tvalidation_0-error:0.119998\tvalidation_1-error:0.126318\n",
      "[139]\tvalidation_0-error:0.120086\tvalidation_1-error:0.12642\n",
      "[140]\tvalidation_0-error:0.119998\tvalidation_1-error:0.126932\n",
      "[141]\tvalidation_0-error:0.119691\tvalidation_1-error:0.12683\n",
      "[142]\tvalidation_0-error:0.119867\tvalidation_1-error:0.126523\n",
      "[143]\tvalidation_0-error:0.119691\tvalidation_1-error:0.126216\n",
      "[144]\tvalidation_0-error:0.119735\tvalidation_1-error:0.126318\n",
      "[145]\tvalidation_0-error:0.119472\tvalidation_1-error:0.126318\n",
      "[146]\tvalidation_0-error:0.119472\tvalidation_1-error:0.126318\n",
      "[147]\tvalidation_0-error:0.119428\tvalidation_1-error:0.126625\n",
      "[148]\tvalidation_0-error:0.119077\tvalidation_1-error:0.126625\n",
      "[149]\tvalidation_0-error:0.118814\tvalidation_1-error:0.127342\n",
      "[150]\tvalidation_0-error:0.118594\tvalidation_1-error:0.127342\n",
      "[151]\tvalidation_0-error:0.11855\tvalidation_1-error:0.127035\n",
      "[152]\tvalidation_0-error:0.11855\tvalidation_1-error:0.126932\n",
      "Stopping. Best iteration:\n",
      "[122]\tvalidation_0-error:0.121622\tvalidation_1-error:0.125601\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=0.9,\n",
       "              colsample_bytree=0.9, gamma=0, learning_rate=0.1,\n",
       "              max_delta_step=0, max_depth=5, min_child_weight=1, missing=None,\n",
       "              n_estimators=400, n_jobs=1, nthread=None,\n",
       "              objective='binary:logistic', random_state=0, reg_alpha=0,\n",
       "              reg_lambda=1.5, scale_pos_weight=1, seed=10, silent=True,\n",
       "              subsample=0.9, verbosity=1)"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "param = {}\n",
    "param['learning_rate'] = 0.1\n",
    "param['verbosity'] = 1\n",
    "param['colsample_bylevel'] = 0.9\n",
    "param['colsample_bytree'] = 0.9\n",
    "param['subsample'] = 0.9\n",
    "param['reg_lambda']= 1.5\n",
    "param['max_depth'] = 5\n",
    "param['n_estimators'] = 400\n",
    "param['seed']=10\n",
    "xgb= XGBClassifier(**param)\n",
    "xgb.fit(X_train, y_train, eval_metric=['error'], eval_set=[(X_train, y_train),(X_test, y_test)],early_stopping_rounds=30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=0.9,\n",
      "              colsample_bytree=0.9, gamma=0, learning_rate=0.1,\n",
      "              max_delta_step=0, max_depth=5, min_child_weight=1, missing=None,\n",
      "              n_estimators=400, n_jobs=1, nthread=None,\n",
      "              objective='binary:logistic', random_state=0, reg_alpha=0,\n",
      "              reg_lambda=1.5, scale_pos_weight=1, seed=10, silent=True,\n",
      "              subsample=0.9, verbosity=1)\n",
      "confusion matrix of training data\n",
      "[[16408   883]\n",
      " [ 1889  3612]]\n",
      "confusion matrix of test data\n",
      "[[7054  375]\n",
      " [ 852 1488]]\n",
      "classification report of test data\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.89      0.95      0.92      7429\n",
      "           1       0.80      0.64      0.71      2340\n",
      "\n",
      "    accuracy                           0.87      9769\n",
      "   macro avg       0.85      0.79      0.81      9769\n",
      "weighted avg       0.87      0.87      0.87      9769\n",
      "\n",
      "Accuracy on test data: 87.439861%\n",
      "Accuracy on training data: 87.837838%\n",
      "Area under the ROC curve : 0.792710\n"
     ]
    }
   ],
   "source": [
    "show_classifier_metrics(xgb,y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "importance_list = xgb.feature_importances_\n",
    "name_list = X_train.columns\n",
    "importance_list, name_list = zip(*sorted(zip(importance_list, name_list)))\n",
    "plt.figure(figsize=(20,10))\n",
    "plt.barh(range(len(name_list)),importance_list,align='center')\n",
    "plt.yticks(range(len(name_list)),name_list)\n",
    "plt.xlabel('Relative Importance in XGBoost')\n",
    "plt.ylabel('Features')\n",
    "plt.title('Relative importance of Each Feature')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Grid search with cross validation: XGBoost model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Doing grid search\n",
      "Fitting 4 folds for each of 9 candidates, totalling 36 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 4 concurrent workers.\n",
      "[Parallel(n_jobs=-1)]: Done  36 out of  36 | elapsed:  2.4min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "mean test score (weighted by split size) of CV rounds:  [0.86982274 0.86938399 0.86876974 0.86938399 0.87017375 0.870788\n",
      " 0.86565462 0.86587399 0.86565462]\n",
      "Best parameter set {'colsample_bylevel': 0.9, 'colsample_bytree': 0.9, 'learning_rate': 0.1, 'max_depth': 5, 'n_estimators': 300, 'reg_lambda': 2.5, 'seed': 10, 'subsample': 0.9} Corresponding mean CV score 0.8707879957879958\n"
     ]
    }
   ],
   "source": [
    "xgbmodel2 = XGBClassifier(seed=42)\n",
    "param = {\n",
    "'learning_rate': [0.1],#[0.1,0.2],\n",
    "#'verbosity': [1],\n",
    "'colsample_bylevel': [0.9],\n",
    "'colsample_bytree': [0.9],\n",
    "'subsample' : [0.9],\n",
    "'n_estimators': [300],\n",
    "'reg_lambda': [1.5,2,2.5],\n",
    "'max_depth': [3,5,7],\n",
    " 'seed': [10]   \n",
    "}\n",
    "xgbCV = grid_search(xgbmodel2, param,X_train,y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0]\tvalidation_0-error:0.174316\tvalidation_1-error:0.178831\n",
      "Multiple eval metrics have been passed: 'validation_1-error' will be used for early stopping.\n",
      "\n",
      "Will train until validation_1-error hasn't improved in 30 rounds.\n",
      "[1]\tvalidation_0-error:0.144086\tvalidation_1-error:0.145358\n",
      "[2]\tvalidation_0-error:0.151501\tvalidation_1-error:0.153854\n",
      "[3]\tvalidation_0-error:0.14413\tvalidation_1-error:0.145665\n",
      "[4]\tvalidation_0-error:0.144612\tvalidation_1-error:0.145767\n",
      "[5]\tvalidation_0-error:0.144437\tvalidation_1-error:0.145255\n",
      "[6]\tvalidation_0-error:0.143954\tvalidation_1-error:0.144641\n",
      "[7]\tvalidation_0-error:0.143735\tvalidation_1-error:0.144744\n",
      "[8]\tvalidation_0-error:0.143559\tvalidation_1-error:0.144744\n",
      "[9]\tvalidation_0-error:0.14119\tvalidation_1-error:0.143618\n",
      "[10]\tvalidation_0-error:0.143252\tvalidation_1-error:0.144641\n",
      "[11]\tvalidation_0-error:0.14255\tvalidation_1-error:0.144129\n",
      "[12]\tvalidation_0-error:0.142287\tvalidation_1-error:0.143618\n",
      "[13]\tvalidation_0-error:0.140883\tvalidation_1-error:0.142901\n",
      "[14]\tvalidation_0-error:0.14062\tvalidation_1-error:0.143515\n",
      "[15]\tvalidation_0-error:0.139874\tvalidation_1-error:0.142799\n",
      "[16]\tvalidation_0-error:0.13961\tvalidation_1-error:0.142901\n",
      "[17]\tvalidation_0-error:0.139567\tvalidation_1-error:0.142799\n",
      "[18]\tvalidation_0-error:0.139479\tvalidation_1-error:0.142696\n",
      "[19]\tvalidation_0-error:0.139391\tvalidation_1-error:0.143106\n",
      "[20]\tvalidation_0-error:0.139523\tvalidation_1-error:0.142594\n",
      "[21]\tvalidation_0-error:0.139786\tvalidation_1-error:0.142492\n",
      "[22]\tvalidation_0-error:0.139479\tvalidation_1-error:0.14198\n",
      "[23]\tvalidation_0-error:0.139259\tvalidation_1-error:0.142082\n",
      "[24]\tvalidation_0-error:0.138865\tvalidation_1-error:0.142082\n",
      "[25]\tvalidation_0-error:0.137548\tvalidation_1-error:0.141366\n",
      "[26]\tvalidation_0-error:0.138163\tvalidation_1-error:0.141366\n",
      "[27]\tvalidation_0-error:0.13768\tvalidation_1-error:0.140956\n",
      "[28]\tvalidation_0-error:0.137373\tvalidation_1-error:0.140854\n",
      "[29]\tvalidation_0-error:0.13711\tvalidation_1-error:0.140751\n",
      "[30]\tvalidation_0-error:0.136715\tvalidation_1-error:0.140547\n",
      "[31]\tvalidation_0-error:0.136276\tvalidation_1-error:0.140444\n",
      "[32]\tvalidation_0-error:0.136144\tvalidation_1-error:0.140547\n",
      "[33]\tvalidation_0-error:0.135749\tvalidation_1-error:0.140137\n",
      "[34]\tvalidation_0-error:0.135311\tvalidation_1-error:0.139932\n",
      "[35]\tvalidation_0-error:0.135267\tvalidation_1-error:0.13983\n",
      "[36]\tvalidation_0-error:0.135135\tvalidation_1-error:0.139625\n",
      "[37]\tvalidation_0-error:0.135267\tvalidation_1-error:0.139625\n",
      "[38]\tvalidation_0-error:0.134828\tvalidation_1-error:0.139625\n",
      "[39]\tvalidation_0-error:0.134653\tvalidation_1-error:0.139728\n",
      "[40]\tvalidation_0-error:0.134258\tvalidation_1-error:0.138602\n",
      "[41]\tvalidation_0-error:0.133994\tvalidation_1-error:0.139114\n",
      "[42]\tvalidation_0-error:0.133336\tvalidation_1-error:0.138602\n",
      "[43]\tvalidation_0-error:0.13338\tvalidation_1-error:0.139114\n",
      "[44]\tvalidation_0-error:0.133073\tvalidation_1-error:0.139011\n",
      "[45]\tvalidation_0-error:0.133073\tvalidation_1-error:0.138909\n",
      "[46]\tvalidation_0-error:0.132371\tvalidation_1-error:0.139011\n",
      "[47]\tvalidation_0-error:0.132327\tvalidation_1-error:0.139216\n",
      "[48]\tvalidation_0-error:0.132941\tvalidation_1-error:0.138602\n",
      "[49]\tvalidation_0-error:0.131976\tvalidation_1-error:0.138295\n",
      "[50]\tvalidation_0-error:0.131845\tvalidation_1-error:0.137988\n",
      "[51]\tvalidation_0-error:0.131494\tvalidation_1-error:0.137373\n",
      "[52]\tvalidation_0-error:0.130879\tvalidation_1-error:0.136964\n",
      "[53]\tvalidation_0-error:0.130835\tvalidation_1-error:0.137169\n",
      "[54]\tvalidation_0-error:0.130528\tvalidation_1-error:0.137066\n",
      "[55]\tvalidation_0-error:0.130528\tvalidation_1-error:0.137271\n",
      "[56]\tvalidation_0-error:0.13066\tvalidation_1-error:0.136862\n",
      "[57]\tvalidation_0-error:0.130353\tvalidation_1-error:0.136759\n",
      "[58]\tvalidation_0-error:0.130177\tvalidation_1-error:0.136657\n",
      "[59]\tvalidation_0-error:0.129563\tvalidation_1-error:0.136247\n",
      "[60]\tvalidation_0-error:0.129387\tvalidation_1-error:0.13594\n",
      "[61]\tvalidation_0-error:0.129212\tvalidation_1-error:0.135428\n",
      "[62]\tvalidation_0-error:0.129124\tvalidation_1-error:0.135224\n",
      "[63]\tvalidation_0-error:0.129212\tvalidation_1-error:0.135224\n",
      "[64]\tvalidation_0-error:0.128817\tvalidation_1-error:0.134712\n",
      "[65]\tvalidation_0-error:0.127983\tvalidation_1-error:0.1342\n",
      "[66]\tvalidation_0-error:0.128071\tvalidation_1-error:0.134507\n",
      "[67]\tvalidation_0-error:0.128159\tvalidation_1-error:0.134405\n",
      "[68]\tvalidation_0-error:0.12794\tvalidation_1-error:0.134098\n",
      "[69]\tvalidation_0-error:0.127764\tvalidation_1-error:0.134098\n",
      "[70]\tvalidation_0-error:0.127676\tvalidation_1-error:0.134098\n",
      "[71]\tvalidation_0-error:0.127983\tvalidation_1-error:0.133176\n",
      "[72]\tvalidation_0-error:0.127238\tvalidation_1-error:0.132255\n",
      "[73]\tvalidation_0-error:0.127194\tvalidation_1-error:0.13205\n",
      "[74]\tvalidation_0-error:0.126887\tvalidation_1-error:0.132153\n",
      "[75]\tvalidation_0-error:0.126931\tvalidation_1-error:0.132255\n",
      "[76]\tvalidation_0-error:0.126887\tvalidation_1-error:0.131948\n",
      "[77]\tvalidation_0-error:0.12658\tvalidation_1-error:0.131641\n",
      "[78]\tvalidation_0-error:0.126536\tvalidation_1-error:0.131641\n",
      "[79]\tvalidation_0-error:0.126228\tvalidation_1-error:0.131231\n",
      "[80]\tvalidation_0-error:0.126272\tvalidation_1-error:0.131436\n",
      "[81]\tvalidation_0-error:0.125921\tvalidation_1-error:0.130924\n",
      "[82]\tvalidation_0-error:0.125527\tvalidation_1-error:0.130617\n",
      "[83]\tvalidation_0-error:0.125351\tvalidation_1-error:0.13072\n",
      "[84]\tvalidation_0-error:0.125132\tvalidation_1-error:0.130822\n",
      "[85]\tvalidation_0-error:0.125044\tvalidation_1-error:0.130617\n",
      "[86]\tvalidation_0-error:0.125176\tvalidation_1-error:0.130515\n",
      "[87]\tvalidation_0-error:0.125219\tvalidation_1-error:0.13031\n",
      "[88]\tvalidation_0-error:0.125132\tvalidation_1-error:0.13031\n",
      "[89]\tvalidation_0-error:0.124956\tvalidation_1-error:0.130208\n",
      "[90]\tvalidation_0-error:0.124386\tvalidation_1-error:0.129696\n",
      "[91]\tvalidation_0-error:0.124386\tvalidation_1-error:0.129287\n",
      "[92]\tvalidation_0-error:0.124254\tvalidation_1-error:0.128672\n",
      "[93]\tvalidation_0-error:0.124166\tvalidation_1-error:0.128672\n",
      "[94]\tvalidation_0-error:0.124079\tvalidation_1-error:0.12857\n",
      "[95]\tvalidation_0-error:0.123859\tvalidation_1-error:0.128877\n",
      "[96]\tvalidation_0-error:0.123464\tvalidation_1-error:0.128672\n",
      "[97]\tvalidation_0-error:0.123377\tvalidation_1-error:0.128775\n",
      "[98]\tvalidation_0-error:0.123377\tvalidation_1-error:0.128979\n",
      "[99]\tvalidation_0-error:0.123069\tvalidation_1-error:0.128877\n",
      "[100]\tvalidation_0-error:0.123245\tvalidation_1-error:0.12857\n",
      "[101]\tvalidation_0-error:0.123201\tvalidation_1-error:0.128877\n",
      "[102]\tvalidation_0-error:0.122982\tvalidation_1-error:0.128877\n",
      "[103]\tvalidation_0-error:0.122982\tvalidation_1-error:0.128877\n",
      "[104]\tvalidation_0-error:0.122762\tvalidation_1-error:0.128468\n",
      "[105]\tvalidation_0-error:0.123069\tvalidation_1-error:0.128263\n",
      "[106]\tvalidation_0-error:0.122938\tvalidation_1-error:0.128468\n",
      "[107]\tvalidation_0-error:0.12285\tvalidation_1-error:0.128161\n",
      "[108]\tvalidation_0-error:0.122587\tvalidation_1-error:0.127956\n",
      "[109]\tvalidation_0-error:0.122587\tvalidation_1-error:0.127546\n",
      "[110]\tvalidation_0-error:0.122543\tvalidation_1-error:0.127546\n",
      "[111]\tvalidation_0-error:0.122236\tvalidation_1-error:0.127853\n",
      "[112]\tvalidation_0-error:0.121973\tvalidation_1-error:0.127649\n",
      "[113]\tvalidation_0-error:0.122104\tvalidation_1-error:0.127546\n",
      "[114]\tvalidation_0-error:0.122148\tvalidation_1-error:0.127649\n",
      "[115]\tvalidation_0-error:0.122016\tvalidation_1-error:0.127649\n",
      "[116]\tvalidation_0-error:0.121929\tvalidation_1-error:0.12683\n",
      "[117]\tvalidation_0-error:0.121973\tvalidation_1-error:0.126932\n",
      "[118]\tvalidation_0-error:0.121841\tvalidation_1-error:0.127342\n",
      "[119]\tvalidation_0-error:0.121753\tvalidation_1-error:0.127035\n",
      "[120]\tvalidation_0-error:0.121753\tvalidation_1-error:0.12683\n",
      "[121]\tvalidation_0-error:0.121666\tvalidation_1-error:0.12683\n",
      "[122]\tvalidation_0-error:0.121666\tvalidation_1-error:0.126727\n",
      "[123]\tvalidation_0-error:0.121622\tvalidation_1-error:0.126625\n",
      "[124]\tvalidation_0-error:0.12149\tvalidation_1-error:0.126523\n",
      "[125]\tvalidation_0-error:0.121446\tvalidation_1-error:0.12642\n",
      "[126]\tvalidation_0-error:0.121358\tvalidation_1-error:0.12642\n",
      "[127]\tvalidation_0-error:0.121314\tvalidation_1-error:0.126318\n",
      "[128]\tvalidation_0-error:0.121271\tvalidation_1-error:0.12642\n",
      "[129]\tvalidation_0-error:0.121314\tvalidation_1-error:0.126523\n",
      "[130]\tvalidation_0-error:0.121183\tvalidation_1-error:0.12642\n",
      "[131]\tvalidation_0-error:0.120963\tvalidation_1-error:0.12642\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[132]\tvalidation_0-error:0.12092\tvalidation_1-error:0.126113\n",
      "[133]\tvalidation_0-error:0.120876\tvalidation_1-error:0.126318\n",
      "[134]\tvalidation_0-error:0.120876\tvalidation_1-error:0.126318\n",
      "[135]\tvalidation_0-error:0.120876\tvalidation_1-error:0.126318\n",
      "[136]\tvalidation_0-error:0.120612\tvalidation_1-error:0.126011\n",
      "[137]\tvalidation_0-error:0.120656\tvalidation_1-error:0.126216\n",
      "[138]\tvalidation_0-error:0.120788\tvalidation_1-error:0.126011\n",
      "[139]\tvalidation_0-error:0.120744\tvalidation_1-error:0.12642\n",
      "[140]\tvalidation_0-error:0.120656\tvalidation_1-error:0.12642\n",
      "[141]\tvalidation_0-error:0.120744\tvalidation_1-error:0.12642\n",
      "[142]\tvalidation_0-error:0.120612\tvalidation_1-error:0.126113\n",
      "[143]\tvalidation_0-error:0.120481\tvalidation_1-error:0.126011\n",
      "[144]\tvalidation_0-error:0.12013\tvalidation_1-error:0.126011\n",
      "[145]\tvalidation_0-error:0.119954\tvalidation_1-error:0.125908\n",
      "[146]\tvalidation_0-error:0.119998\tvalidation_1-error:0.125806\n",
      "[147]\tvalidation_0-error:0.120086\tvalidation_1-error:0.126216\n",
      "[148]\tvalidation_0-error:0.120086\tvalidation_1-error:0.126216\n",
      "[149]\tvalidation_0-error:0.120042\tvalidation_1-error:0.126523\n",
      "[150]\tvalidation_0-error:0.120042\tvalidation_1-error:0.12642\n",
      "[151]\tvalidation_0-error:0.119998\tvalidation_1-error:0.126216\n",
      "[152]\tvalidation_0-error:0.119954\tvalidation_1-error:0.126216\n",
      "[153]\tvalidation_0-error:0.119691\tvalidation_1-error:0.126113\n",
      "[154]\tvalidation_0-error:0.119779\tvalidation_1-error:0.126113\n",
      "[155]\tvalidation_0-error:0.119735\tvalidation_1-error:0.126216\n",
      "[156]\tvalidation_0-error:0.119603\tvalidation_1-error:0.126113\n",
      "[157]\tvalidation_0-error:0.119472\tvalidation_1-error:0.126011\n",
      "[158]\tvalidation_0-error:0.119559\tvalidation_1-error:0.126216\n",
      "[159]\tvalidation_0-error:0.119472\tvalidation_1-error:0.126216\n",
      "[160]\tvalidation_0-error:0.119516\tvalidation_1-error:0.126113\n",
      "[161]\tvalidation_0-error:0.119559\tvalidation_1-error:0.126216\n",
      "[162]\tvalidation_0-error:0.119516\tvalidation_1-error:0.126216\n",
      "[163]\tvalidation_0-error:0.11934\tvalidation_1-error:0.126523\n",
      "[164]\tvalidation_0-error:0.119384\tvalidation_1-error:0.126216\n",
      "[165]\tvalidation_0-error:0.119165\tvalidation_1-error:0.126318\n",
      "[166]\tvalidation_0-error:0.119252\tvalidation_1-error:0.12642\n",
      "[167]\tvalidation_0-error:0.11934\tvalidation_1-error:0.126318\n",
      "[168]\tvalidation_0-error:0.119252\tvalidation_1-error:0.126011\n",
      "[169]\tvalidation_0-error:0.119208\tvalidation_1-error:0.126113\n",
      "[170]\tvalidation_0-error:0.119033\tvalidation_1-error:0.126318\n",
      "[171]\tvalidation_0-error:0.118989\tvalidation_1-error:0.126216\n",
      "[172]\tvalidation_0-error:0.11877\tvalidation_1-error:0.126011\n",
      "[173]\tvalidation_0-error:0.118726\tvalidation_1-error:0.126011\n",
      "[174]\tvalidation_0-error:0.118594\tvalidation_1-error:0.126318\n",
      "[175]\tvalidation_0-error:0.118814\tvalidation_1-error:0.126216\n",
      "[176]\tvalidation_0-error:0.118857\tvalidation_1-error:0.125806\n",
      "Stopping. Best iteration:\n",
      "[146]\tvalidation_0-error:0.119998\tvalidation_1-error:0.125806\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=0.9,\n",
       "              colsample_bytree=0.9, gamma=0, learning_rate=0.1,\n",
       "              max_delta_step=0, max_depth=5, min_child_weight=1, missing=None,\n",
       "              n_estimators=300, n_jobs=1, nthread=None,\n",
       "              objective='binary:logistic', random_state=0, reg_alpha=0,\n",
       "              reg_lambda=2.5, scale_pos_weight=1, seed=10, silent=True,\n",
       "              subsample=0.9)"
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "xgbCV.fit(X_train, y_train, eval_metric=['error'], eval_set=[(X_train, y_train),(X_test, y_test)],early_stopping_rounds=30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=0.9,\n",
      "              colsample_bytree=0.9, gamma=0, learning_rate=0.1,\n",
      "              max_delta_step=0, max_depth=5, min_child_weight=1, missing=None,\n",
      "              n_estimators=300, n_jobs=1, nthread=None,\n",
      "              objective='binary:logistic', random_state=0, reg_alpha=0,\n",
      "              reg_lambda=2.5, scale_pos_weight=1, seed=10, silent=True,\n",
      "              subsample=0.9)\n",
      "confusion matrix of training data\n",
      "[[16413   878]\n",
      " [ 1857  3644]]\n",
      "confusion matrix of test data\n",
      "[[7046  383]\n",
      " [ 846 1494]]\n",
      "classification report of test data\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.89      0.95      0.92      7429\n",
      "           1       0.80      0.64      0.71      2340\n",
      "\n",
      "    accuracy                           0.87      9769\n",
      "   macro avg       0.84      0.79      0.81      9769\n",
      "weighted avg       0.87      0.87      0.87      9769\n",
      "\n",
      "Accuracy on test data: 87.419388%\n",
      "Accuracy on training data: 88.000176%\n",
      "Area under the ROC curve : 0.793453\n"
     ]
    }
   ],
   "source": [
    "show_classifier_metrics(xgbCV,y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#X_test.iloc[np.where(y_test != xgbCV.predict(X_test))]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Logistic regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.linear_model import LogisticRegression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Doing grid search\n",
      "Fitting 3 folds for each of 12 candidates, totalling 36 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 4 concurrent workers.\n",
      "[Parallel(n_jobs=-1)]: Done  36 out of  36 | elapsed:    3.7s finished\n",
      "/Users/debashis/anaconda3/lib/python3.6/site-packages/sklearn/linear_model/logistic.py:432: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[LibLinear]mean test score (weighted by split size) of CV rounds:  [0.8498596 0.8498596 0.8498596 0.8498596 0.8502106 0.8502106 0.8502106\n",
      " 0.8502106 0.8495086 0.8495086 0.8495086 0.8495086]\n",
      "Best parameter set {'C': 5, 'max_iter': 100, 'verbose': 1} Corresponding mean CV score 0.8502106002106002\n"
     ]
    }
   ],
   "source": [
    "param = {\n",
    "'C': [3,5,10], \n",
    "'verbose': [1],\n",
    "    'max_iter': [100,200,500,700]\n",
    "}   \n",
    "logreg = LogisticRegression(random_state=10)\n",
    "logreg_grid = grid_search(logreg, param, X_train,y_train, n_folds=3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[LibLinear]"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/debashis/anaconda3/lib/python3.6/site-packages/sklearn/linear_model/logistic.py:432: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "LogisticRegression(C=5, class_weight=None, dual=False, fit_intercept=True,\n",
       "                   intercept_scaling=1, l1_ratio=None, max_iter=100,\n",
       "                   multi_class='warn', n_jobs=None, penalty='l2',\n",
       "                   random_state=10, solver='warn', tol=0.0001, verbose=1,\n",
       "                   warm_start=False)"
      ]
     },
     "execution_count": 100,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "logreg_grid.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LogisticRegression(C=5, class_weight=None, dual=False, fit_intercept=True,\n",
      "                   intercept_scaling=1, l1_ratio=None, max_iter=100,\n",
      "                   multi_class='warn', n_jobs=None, penalty='l2',\n",
      "                   random_state=10, solver='warn', tol=0.0001, verbose=1,\n",
      "                   warm_start=False)\n",
      "confusion matrix of training data\n",
      "[[16110  1181]\n",
      " [ 2213  3288]]\n",
      "confusion matrix of test data\n",
      "[[6929  500]\n",
      " [ 990 1350]]\n",
      "classification report of test data\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.87      0.93      0.90      7429\n",
      "           1       0.73      0.58      0.64      2340\n",
      "\n",
      "    accuracy                           0.85      9769\n",
      "   macro avg       0.80      0.75      0.77      9769\n",
      "weighted avg       0.84      0.85      0.84      9769\n",
      "\n",
      "Accuracy on test data: 84.747671%\n",
      "Accuracy on training data: 85.108810%\n",
      "Area under the ROC curve : 0.754810\n"
     ]
    }
   ],
   "source": [
    "show_classifier_metrics(logreg_grid)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Naive Bayes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.naive_bayes import GaussianNB"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {},
   "outputs": [],
   "source": [
    "NBmodel = GaussianNB()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GaussianNB(priors=None, var_smoothing=1e-09)"
      ]
     },
     "execution_count": 104,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "NBmodel.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 0, 1, ..., 0, 0, 0])"
      ]
     },
     "execution_count": 105,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "NBmodel.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "GaussianNB(priors=None, var_smoothing=1e-09)\n",
      "confusion matrix of training data\n",
      "[[14764  2527]\n",
      " [ 1443  4058]]\n",
      "confusion matrix of test data\n",
      "[[6394 1035]\n",
      " [ 640 1700]]\n",
      "classification report of test data\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.91      0.86      0.88      7429\n",
      "           1       0.62      0.73      0.67      2340\n",
      "\n",
      "    accuracy                           0.83      9769\n",
      "   macro avg       0.77      0.79      0.78      9769\n",
      "weighted avg       0.84      0.83      0.83      9769\n",
      "\n",
      "Accuracy on test data: 82.853926%\n",
      "Accuracy on training data: 82.581608%\n",
      "Area under the ROC curve : 0.793588\n"
     ]
    }
   ],
   "source": [
    "show_classifier_metrics(NBmodel,y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Stacked model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "metadata": {},
   "outputs": [],
   "source": [
    "def create_stacked_dataset(clfs,modelnames, X_train=X_train,X_test=X_test):\n",
    "    X_train_stack, X_test_stack = X_train, X_test\n",
    "    for clf,modelname in zip(clfs,modelnames):\n",
    "        temptrain = pd.DataFrame(clf.predict(X_train),index = X_train.index,columns=[modelname+'_prediction'])\n",
    "        temptest  = pd.DataFrame(clf.predict(X_test),index = X_test.index,columns=[modelname+'_prediction'])\n",
    "        X_train_stack = pd.concat([X_train_stack, temptrain], axis=1)\n",
    "        X_test_stack = pd.concat([X_test_stack, temptest], axis=1)\n",
    "    return (X_train_stack,X_test_stack)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_stack,X_test_stack = create_stacked_dataset([NBmodel,rfmodelCV,logreg_grid,xgb],modelnames=['NBmodel','rfmodelCV','logreg', 'xgb'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "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>age</th>\n",
       "      <th>education.num</th>\n",
       "      <th>capital.gain</th>\n",
       "      <th>capital.loss</th>\n",
       "      <th>hours.per.week</th>\n",
       "      <th>workclass_Local-gov</th>\n",
       "      <th>workclass_Never-worked</th>\n",
       "      <th>workclass_Private</th>\n",
       "      <th>workclass_Self-emp-inc</th>\n",
       "      <th>workclass_Self-emp-not-inc</th>\n",
       "      <th>...</th>\n",
       "      <th>is_Priv-house-serv</th>\n",
       "      <th>is_Prof-specialty</th>\n",
       "      <th>is_Protective-serv</th>\n",
       "      <th>is_Sales</th>\n",
       "      <th>is_Tech-support</th>\n",
       "      <th>is_Transport-moving</th>\n",
       "      <th>NBmodel_prediction</th>\n",
       "      <th>rfmodelCV_prediction</th>\n",
       "      <th>logreg_prediction</th>\n",
       "      <th>xgb_prediction</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>19749</th>\n",
       "      <td>58</td>\n",
       "      <td>9</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>40</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1216</th>\n",
       "      <td>62</td>\n",
       "      <td>10</td>\n",
       "      <td>0</td>\n",
       "      <td>1617</td>\n",
       "      <td>33</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2 rows × 45 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       age  education.num  capital.gain  capital.loss  hours.per.week  \\\n",
       "19749   58              9             0             0              40   \n",
       "1216    62             10             0          1617              33   \n",
       "\n",
       "       workclass_Local-gov  workclass_Never-worked  workclass_Private  \\\n",
       "19749                    0                       0                  1   \n",
       "1216                     0                       0                  1   \n",
       "\n",
       "       workclass_Self-emp-inc  workclass_Self-emp-not-inc  ...  \\\n",
       "19749                       0                           0  ...   \n",
       "1216                        0                           0  ...   \n",
       "\n",
       "       is_Priv-house-serv  is_Prof-specialty  is_Protective-serv  is_Sales  \\\n",
       "19749                   0                  0                   0         0   \n",
       "1216                    0                  0                   0         1   \n",
       "\n",
       "       is_Tech-support  is_Transport-moving  NBmodel_prediction  \\\n",
       "19749                0                    0                   0   \n",
       "1216                 0                    0                   0   \n",
       "\n",
       "       rfmodelCV_prediction  logreg_prediction  xgb_prediction  \n",
       "19749                     0                  0               0  \n",
       "1216                      0                  0               0  \n",
       "\n",
       "[2 rows x 45 columns]"
      ]
     },
     "execution_count": 114,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train_stack.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0]\tvalidation_0-error:0.114075\tvalidation_1-error:0.137476\n",
      "Multiple eval metrics have been passed: 'validation_1-error' will be used for early stopping.\n",
      "\n",
      "Will train until validation_1-error hasn't improved in 30 rounds.\n",
      "[1]\tvalidation_0-error:0.112759\tvalidation_1-error:0.13594\n",
      "[2]\tvalidation_0-error:0.111794\tvalidation_1-error:0.132665\n",
      "[3]\tvalidation_0-error:0.111969\tvalidation_1-error:0.132767\n",
      "[4]\tvalidation_0-error:0.111837\tvalidation_1-error:0.132665\n",
      "[5]\tvalidation_0-error:0.111881\tvalidation_1-error:0.133176\n",
      "[6]\tvalidation_0-error:0.112145\tvalidation_1-error:0.133586\n",
      "[7]\tvalidation_0-error:0.112364\tvalidation_1-error:0.133893\n",
      "[8]\tvalidation_0-error:0.112188\tvalidation_1-error:0.133688\n",
      "[9]\tvalidation_0-error:0.112232\tvalidation_1-error:0.133688\n",
      "[10]\tvalidation_0-error:0.111837\tvalidation_1-error:0.133279\n",
      "[11]\tvalidation_0-error:0.112013\tvalidation_1-error:0.133483\n",
      "[12]\tvalidation_0-error:0.112057\tvalidation_1-error:0.133381\n",
      "[13]\tvalidation_0-error:0.11175\tvalidation_1-error:0.133074\n",
      "[14]\tvalidation_0-error:0.111969\tvalidation_1-error:0.133176\n",
      "[15]\tvalidation_0-error:0.112057\tvalidation_1-error:0.133381\n",
      "[16]\tvalidation_0-error:0.112013\tvalidation_1-error:0.133381\n",
      "[17]\tvalidation_0-error:0.111837\tvalidation_1-error:0.132972\n",
      "[18]\tvalidation_0-error:0.111969\tvalidation_1-error:0.133279\n",
      "[19]\tvalidation_0-error:0.111969\tvalidation_1-error:0.133381\n",
      "[20]\tvalidation_0-error:0.111837\tvalidation_1-error:0.132972\n",
      "[21]\tvalidation_0-error:0.111881\tvalidation_1-error:0.132972\n",
      "[22]\tvalidation_0-error:0.111794\tvalidation_1-error:0.133381\n",
      "[23]\tvalidation_0-error:0.11175\tvalidation_1-error:0.133074\n",
      "[24]\tvalidation_0-error:0.111706\tvalidation_1-error:0.133279\n",
      "[25]\tvalidation_0-error:0.111706\tvalidation_1-error:0.133074\n",
      "[26]\tvalidation_0-error:0.111443\tvalidation_1-error:0.132153\n",
      "[27]\tvalidation_0-error:0.111443\tvalidation_1-error:0.132153\n",
      "[28]\tvalidation_0-error:0.111399\tvalidation_1-error:0.132255\n",
      "[29]\tvalidation_0-error:0.111399\tvalidation_1-error:0.132562\n",
      "[30]\tvalidation_0-error:0.111267\tvalidation_1-error:0.132357\n",
      "[31]\tvalidation_0-error:0.111135\tvalidation_1-error:0.132255\n",
      "[32]\tvalidation_0-error:0.111179\tvalidation_1-error:0.132357\n",
      "[33]\tvalidation_0-error:0.111092\tvalidation_1-error:0.132255\n",
      "[34]\tvalidation_0-error:0.111004\tvalidation_1-error:0.132255\n",
      "[35]\tvalidation_0-error:0.111004\tvalidation_1-error:0.132255\n",
      "[36]\tvalidation_0-error:0.111004\tvalidation_1-error:0.132255\n",
      "[37]\tvalidation_0-error:0.111004\tvalidation_1-error:0.132153\n",
      "[38]\tvalidation_0-error:0.111004\tvalidation_1-error:0.132153\n",
      "[39]\tvalidation_0-error:0.11096\tvalidation_1-error:0.132153\n",
      "[40]\tvalidation_0-error:0.111004\tvalidation_1-error:0.132153\n",
      "[41]\tvalidation_0-error:0.11096\tvalidation_1-error:0.131948\n",
      "[42]\tvalidation_0-error:0.111004\tvalidation_1-error:0.131948\n",
      "[43]\tvalidation_0-error:0.111004\tvalidation_1-error:0.131948\n",
      "[44]\tvalidation_0-error:0.111004\tvalidation_1-error:0.131948\n",
      "[45]\tvalidation_0-error:0.111004\tvalidation_1-error:0.131948\n",
      "[46]\tvalidation_0-error:0.110741\tvalidation_1-error:0.131846\n",
      "[47]\tvalidation_0-error:0.110916\tvalidation_1-error:0.130822\n",
      "[48]\tvalidation_0-error:0.111092\tvalidation_1-error:0.131027\n",
      "[49]\tvalidation_0-error:0.111135\tvalidation_1-error:0.131129\n",
      "[50]\tvalidation_0-error:0.111004\tvalidation_1-error:0.131027\n",
      "[51]\tvalidation_0-error:0.111048\tvalidation_1-error:0.130924\n",
      "[52]\tvalidation_0-error:0.111092\tvalidation_1-error:0.130822\n",
      "[53]\tvalidation_0-error:0.11096\tvalidation_1-error:0.130822\n",
      "[54]\tvalidation_0-error:0.11096\tvalidation_1-error:0.130413\n",
      "[55]\tvalidation_0-error:0.110828\tvalidation_1-error:0.130003\n",
      "[56]\tvalidation_0-error:0.110872\tvalidation_1-error:0.130003\n",
      "[57]\tvalidation_0-error:0.111048\tvalidation_1-error:0.13031\n",
      "[58]\tvalidation_0-error:0.110916\tvalidation_1-error:0.129901\n",
      "[59]\tvalidation_0-error:0.110741\tvalidation_1-error:0.129491\n",
      "[60]\tvalidation_0-error:0.110477\tvalidation_1-error:0.129389\n",
      "[61]\tvalidation_0-error:0.110609\tvalidation_1-error:0.129594\n",
      "[62]\tvalidation_0-error:0.110653\tvalidation_1-error:0.129594\n",
      "[63]\tvalidation_0-error:0.110653\tvalidation_1-error:0.129491\n",
      "[64]\tvalidation_0-error:0.110433\tvalidation_1-error:0.129389\n",
      "[65]\tvalidation_0-error:0.110346\tvalidation_1-error:0.129389\n",
      "[66]\tvalidation_0-error:0.110258\tvalidation_1-error:0.128877\n",
      "[67]\tvalidation_0-error:0.110258\tvalidation_1-error:0.128775\n",
      "[68]\tvalidation_0-error:0.110039\tvalidation_1-error:0.128877\n",
      "[69]\tvalidation_0-error:0.109951\tvalidation_1-error:0.128877\n",
      "[70]\tvalidation_0-error:0.109951\tvalidation_1-error:0.128877\n",
      "[71]\tvalidation_0-error:0.110039\tvalidation_1-error:0.128775\n",
      "[72]\tvalidation_0-error:0.110082\tvalidation_1-error:0.128672\n",
      "[73]\tvalidation_0-error:0.110039\tvalidation_1-error:0.128672\n",
      "[74]\tvalidation_0-error:0.110082\tvalidation_1-error:0.128672\n",
      "[75]\tvalidation_0-error:0.109951\tvalidation_1-error:0.128775\n",
      "[76]\tvalidation_0-error:0.109907\tvalidation_1-error:0.128775\n",
      "[77]\tvalidation_0-error:0.109951\tvalidation_1-error:0.128775\n",
      "[78]\tvalidation_0-error:0.109951\tvalidation_1-error:0.128775\n",
      "[79]\tvalidation_0-error:0.109863\tvalidation_1-error:0.128775\n",
      "[80]\tvalidation_0-error:0.109775\tvalidation_1-error:0.12857\n",
      "[81]\tvalidation_0-error:0.109731\tvalidation_1-error:0.128775\n",
      "[82]\tvalidation_0-error:0.109731\tvalidation_1-error:0.128775\n",
      "[83]\tvalidation_0-error:0.109775\tvalidation_1-error:0.12857\n",
      "[84]\tvalidation_0-error:0.109688\tvalidation_1-error:0.12857\n",
      "[85]\tvalidation_0-error:0.109819\tvalidation_1-error:0.128468\n",
      "[86]\tvalidation_0-error:0.109644\tvalidation_1-error:0.12857\n",
      "[87]\tvalidation_0-error:0.109644\tvalidation_1-error:0.128672\n",
      "[88]\tvalidation_0-error:0.109644\tvalidation_1-error:0.128672\n",
      "[89]\tvalidation_0-error:0.109644\tvalidation_1-error:0.128672\n",
      "[90]\tvalidation_0-error:0.109644\tvalidation_1-error:0.128468\n",
      "[91]\tvalidation_0-error:0.109644\tvalidation_1-error:0.12857\n",
      "[92]\tvalidation_0-error:0.109644\tvalidation_1-error:0.128672\n",
      "[93]\tvalidation_0-error:0.109644\tvalidation_1-error:0.12857\n",
      "[94]\tvalidation_0-error:0.109644\tvalidation_1-error:0.12857\n",
      "[95]\tvalidation_0-error:0.109556\tvalidation_1-error:0.12857\n",
      "[96]\tvalidation_0-error:0.109512\tvalidation_1-error:0.12857\n",
      "[97]\tvalidation_0-error:0.109512\tvalidation_1-error:0.128672\n",
      "[98]\tvalidation_0-error:0.109556\tvalidation_1-error:0.12857\n",
      "[99]\tvalidation_0-error:0.1096\tvalidation_1-error:0.12857\n",
      "[100]\tvalidation_0-error:0.1096\tvalidation_1-error:0.128468\n",
      "[101]\tvalidation_0-error:0.109424\tvalidation_1-error:0.128468\n",
      "[102]\tvalidation_0-error:0.109468\tvalidation_1-error:0.128468\n",
      "[103]\tvalidation_0-error:0.109424\tvalidation_1-error:0.128468\n",
      "[104]\tvalidation_0-error:0.109468\tvalidation_1-error:0.128468\n",
      "[105]\tvalidation_0-error:0.109468\tvalidation_1-error:0.128468\n",
      "[106]\tvalidation_0-error:0.109468\tvalidation_1-error:0.128365\n",
      "[107]\tvalidation_0-error:0.109468\tvalidation_1-error:0.128365\n",
      "[108]\tvalidation_0-error:0.109424\tvalidation_1-error:0.128365\n",
      "[109]\tvalidation_0-error:0.109468\tvalidation_1-error:0.128263\n",
      "[110]\tvalidation_0-error:0.109468\tvalidation_1-error:0.128263\n",
      "[111]\tvalidation_0-error:0.109468\tvalidation_1-error:0.128263\n",
      "[112]\tvalidation_0-error:0.109424\tvalidation_1-error:0.128263\n",
      "[113]\tvalidation_0-error:0.109512\tvalidation_1-error:0.128365\n",
      "[114]\tvalidation_0-error:0.109117\tvalidation_1-error:0.128468\n",
      "[115]\tvalidation_0-error:0.109117\tvalidation_1-error:0.128263\n",
      "[116]\tvalidation_0-error:0.109073\tvalidation_1-error:0.128263\n",
      "[117]\tvalidation_0-error:0.109073\tvalidation_1-error:0.128058\n",
      "[118]\tvalidation_0-error:0.109161\tvalidation_1-error:0.128058\n",
      "[119]\tvalidation_0-error:0.109161\tvalidation_1-error:0.128058\n",
      "[120]\tvalidation_0-error:0.108986\tvalidation_1-error:0.128263\n",
      "[121]\tvalidation_0-error:0.108722\tvalidation_1-error:0.128263\n",
      "[122]\tvalidation_0-error:0.108722\tvalidation_1-error:0.128161\n",
      "[123]\tvalidation_0-error:0.108678\tvalidation_1-error:0.128263\n",
      "[124]\tvalidation_0-error:0.108678\tvalidation_1-error:0.128058\n",
      "[125]\tvalidation_0-error:0.108678\tvalidation_1-error:0.128058\n",
      "[126]\tvalidation_0-error:0.108722\tvalidation_1-error:0.128058\n",
      "[127]\tvalidation_0-error:0.108722\tvalidation_1-error:0.128161\n",
      "[128]\tvalidation_0-error:0.108722\tvalidation_1-error:0.128161\n",
      "[129]\tvalidation_0-error:0.108722\tvalidation_1-error:0.128161\n",
      "[130]\tvalidation_0-error:0.108766\tvalidation_1-error:0.128161\n",
      "[131]\tvalidation_0-error:0.108766\tvalidation_1-error:0.128161\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[132]\tvalidation_0-error:0.108942\tvalidation_1-error:0.128161\n",
      "[133]\tvalidation_0-error:0.108942\tvalidation_1-error:0.128161\n",
      "[134]\tvalidation_0-error:0.108898\tvalidation_1-error:0.128263\n",
      "[135]\tvalidation_0-error:0.108898\tvalidation_1-error:0.128263\n",
      "[136]\tvalidation_0-error:0.108898\tvalidation_1-error:0.128263\n",
      "[137]\tvalidation_0-error:0.108898\tvalidation_1-error:0.128161\n",
      "[138]\tvalidation_0-error:0.108898\tvalidation_1-error:0.127956\n",
      "[139]\tvalidation_0-error:0.108854\tvalidation_1-error:0.127956\n",
      "[140]\tvalidation_0-error:0.10881\tvalidation_1-error:0.128058\n",
      "[141]\tvalidation_0-error:0.108854\tvalidation_1-error:0.128058\n",
      "[142]\tvalidation_0-error:0.108854\tvalidation_1-error:0.128058\n",
      "[143]\tvalidation_0-error:0.108854\tvalidation_1-error:0.128058\n",
      "[144]\tvalidation_0-error:0.108766\tvalidation_1-error:0.128161\n",
      "[145]\tvalidation_0-error:0.108678\tvalidation_1-error:0.128058\n",
      "[146]\tvalidation_0-error:0.108635\tvalidation_1-error:0.128058\n",
      "[147]\tvalidation_0-error:0.108635\tvalidation_1-error:0.128263\n",
      "[148]\tvalidation_0-error:0.108591\tvalidation_1-error:0.128263\n",
      "[149]\tvalidation_0-error:0.108591\tvalidation_1-error:0.128263\n",
      "[150]\tvalidation_0-error:0.108591\tvalidation_1-error:0.128263\n",
      "[151]\tvalidation_0-error:0.108327\tvalidation_1-error:0.12857\n",
      "[152]\tvalidation_0-error:0.108196\tvalidation_1-error:0.12857\n",
      "[153]\tvalidation_0-error:0.10824\tvalidation_1-error:0.128672\n",
      "[154]\tvalidation_0-error:0.108196\tvalidation_1-error:0.128672\n",
      "[155]\tvalidation_0-error:0.108284\tvalidation_1-error:0.128672\n",
      "[156]\tvalidation_0-error:0.108196\tvalidation_1-error:0.12857\n",
      "[157]\tvalidation_0-error:0.108196\tvalidation_1-error:0.128365\n",
      "[158]\tvalidation_0-error:0.108064\tvalidation_1-error:0.128263\n",
      "[159]\tvalidation_0-error:0.10802\tvalidation_1-error:0.128263\n",
      "[160]\tvalidation_0-error:0.107976\tvalidation_1-error:0.128672\n",
      "[161]\tvalidation_0-error:0.107757\tvalidation_1-error:0.128877\n",
      "[162]\tvalidation_0-error:0.107801\tvalidation_1-error:0.128775\n",
      "[163]\tvalidation_0-error:0.107713\tvalidation_1-error:0.128672\n",
      "[164]\tvalidation_0-error:0.107669\tvalidation_1-error:0.128979\n",
      "[165]\tvalidation_0-error:0.107713\tvalidation_1-error:0.128979\n",
      "[166]\tvalidation_0-error:0.107713\tvalidation_1-error:0.128877\n",
      "[167]\tvalidation_0-error:0.107582\tvalidation_1-error:0.128877\n",
      "[168]\tvalidation_0-error:0.107582\tvalidation_1-error:0.128877\n",
      "Stopping. Best iteration:\n",
      "[138]\tvalidation_0-error:0.108898\tvalidation_1-error:0.127956\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=0.9,\n",
       "              colsample_bytree=0.9, gamma=0, learning_rate=0.1,\n",
       "              max_delta_step=0, max_depth=5, min_child_weight=1, missing=None,\n",
       "              n_estimators=400, n_jobs=1, nthread=None,\n",
       "              objective='binary:logistic', random_state=0, reg_alpha=0,\n",
       "              reg_lambda=1.5, scale_pos_weight=1, seed=10, silent=True,\n",
       "              subsample=0.9, verbosity=1)"
      ]
     },
     "execution_count": 115,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "param = {}\n",
    "param['learning_rate'] = 0.1\n",
    "param['verbosity'] = 1\n",
    "param['colsample_bylevel'] = 0.9\n",
    "param['colsample_bytree'] = 0.9\n",
    "param['subsample'] = 0.9\n",
    "param['reg_lambda']= 1.5\n",
    "param['max_depth'] = 5#10\n",
    "param['n_estimators'] = 400\n",
    "param['seed']=10\n",
    "xgbstack= XGBClassifier(**param)\n",
    "xgbstack.fit(X_train_stack, y_train, eval_metric=['error'], eval_set=[(X_train_stack, y_train),(X_test_stack, y_test)],early_stopping_rounds=30)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.89      0.95      0.92      7429\n",
      "           1       0.81      0.61      0.70      2340\n",
      "\n",
      "    accuracy                           0.87      9769\n",
      "   macro avg       0.85      0.78      0.81      9769\n",
      "weighted avg       0.87      0.87      0.87      9769\n",
      "\n",
      "Accuracy on test data: 87.204422%\n",
      "Accuracy on training data: 89.110214%\n"
     ]
    }
   ],
   "source": [
    "print(metrics.classification_report(y_test, xgbstack.predict(X_test_stack)))\n",
    "print('Accuracy on test data: %f%%' % (metrics.accuracy_score(y_test, xgbstack.predict(X_test_stack))*100))\n",
    "print('Accuracy on training data: %f%%' % (metrics.accuracy_score(y_train, xgbstack.predict(X_train_stack))*100))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Stacked model Grid Search"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Doing grid search\n",
      "Fitting 4 folds for each of 4 candidates, totalling 16 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 4 concurrent workers.\n",
      "[Parallel(n_jobs=-1)]: Done  16 out of  16 | elapsed:  1.5min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "mean test score (weighted by split size) of CV rounds:  [0.88842576 0.88566164 0.87802738 0.87285012]\n",
      "Best parameter set {'colsample_bylevel': 0.9, 'colsample_bytree': 0.9, 'learning_rate': 0.1, 'max_depth': 3, 'n_estimators': 300, 'reg_lambda': 1.5, 'seed': 10, 'subsample': 0.9} Corresponding mean CV score 0.8884257634257634\n"
     ]
    }
   ],
   "source": [
    "xgbstackCV = XGBClassifier(seed=10)\n",
    "param_grid = {}\n",
    "param_grid['learning_rate'] = [0.1]\n",
    "param_grid['colsample_bylevel'] = [0.9]\n",
    "param_grid['colsample_bytree'] = [0.9]\n",
    "param_grid['subsample'] = [0.9]\n",
    "param_grid['n_estimators'] = [300]\n",
    "param_grid['reg_lambda']= [1.5]\n",
    "param_grid['seed'] =[10]\n",
    "param_grid['max_depth'] = [3,5,8,10]\n",
    "xgbstackCV_grid = grid_search(xgbstackCV, param_grid,X_train_stack,y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0]\tvalidation_0-error:0.114075\tvalidation_1-error:0.137476\n",
      "Multiple eval metrics have been passed: 'validation_1-error' will be used for early stopping.\n",
      "\n",
      "Will train until validation_1-error hasn't improved in 30 rounds.\n",
      "[1]\tvalidation_0-error:0.114075\tvalidation_1-error:0.137476\n",
      "[2]\tvalidation_0-error:0.114075\tvalidation_1-error:0.137476\n",
      "[3]\tvalidation_0-error:0.114075\tvalidation_1-error:0.137476\n",
      "[4]\tvalidation_0-error:0.114075\tvalidation_1-error:0.137476\n",
      "[5]\tvalidation_0-error:0.114075\tvalidation_1-error:0.137476\n",
      "[6]\tvalidation_0-error:0.114075\tvalidation_1-error:0.137476\n",
      "[7]\tvalidation_0-error:0.114075\tvalidation_1-error:0.137476\n",
      "[8]\tvalidation_0-error:0.114075\tvalidation_1-error:0.137476\n",
      "[9]\tvalidation_0-error:0.114075\tvalidation_1-error:0.137476\n",
      "[10]\tvalidation_0-error:0.114075\tvalidation_1-error:0.137476\n",
      "[11]\tvalidation_0-error:0.114075\tvalidation_1-error:0.137476\n",
      "[12]\tvalidation_0-error:0.114075\tvalidation_1-error:0.137476\n",
      "[13]\tvalidation_0-error:0.114075\tvalidation_1-error:0.137476\n",
      "[14]\tvalidation_0-error:0.114075\tvalidation_1-error:0.137476\n",
      "[15]\tvalidation_0-error:0.112803\tvalidation_1-error:0.135019\n",
      "[16]\tvalidation_0-error:0.112803\tvalidation_1-error:0.135019\n",
      "[17]\tvalidation_0-error:0.112847\tvalidation_1-error:0.135121\n",
      "[18]\tvalidation_0-error:0.112847\tvalidation_1-error:0.135121\n",
      "[19]\tvalidation_0-error:0.112671\tvalidation_1-error:0.134405\n",
      "[20]\tvalidation_0-error:0.112759\tvalidation_1-error:0.134609\n",
      "[21]\tvalidation_0-error:0.112715\tvalidation_1-error:0.134814\n",
      "[22]\tvalidation_0-error:0.112715\tvalidation_1-error:0.134814\n",
      "[23]\tvalidation_0-error:0.112759\tvalidation_1-error:0.134814\n",
      "[24]\tvalidation_0-error:0.112759\tvalidation_1-error:0.134917\n",
      "[25]\tvalidation_0-error:0.11289\tvalidation_1-error:0.134917\n",
      "[26]\tvalidation_0-error:0.11289\tvalidation_1-error:0.134917\n",
      "[27]\tvalidation_0-error:0.11289\tvalidation_1-error:0.134917\n",
      "[28]\tvalidation_0-error:0.112803\tvalidation_1-error:0.134917\n",
      "[29]\tvalidation_0-error:0.112803\tvalidation_1-error:0.134917\n",
      "[30]\tvalidation_0-error:0.112803\tvalidation_1-error:0.134917\n",
      "[31]\tvalidation_0-error:0.112803\tvalidation_1-error:0.134917\n",
      "[32]\tvalidation_0-error:0.112759\tvalidation_1-error:0.134917\n",
      "[33]\tvalidation_0-error:0.112759\tvalidation_1-error:0.134917\n",
      "[34]\tvalidation_0-error:0.112847\tvalidation_1-error:0.135224\n",
      "[35]\tvalidation_0-error:0.11289\tvalidation_1-error:0.135224\n",
      "[36]\tvalidation_0-error:0.11289\tvalidation_1-error:0.135224\n",
      "[37]\tvalidation_0-error:0.11289\tvalidation_1-error:0.135224\n",
      "[38]\tvalidation_0-error:0.112847\tvalidation_1-error:0.135224\n",
      "[39]\tvalidation_0-error:0.112847\tvalidation_1-error:0.135224\n",
      "[40]\tvalidation_0-error:0.112847\tvalidation_1-error:0.135224\n",
      "[41]\tvalidation_0-error:0.112847\tvalidation_1-error:0.135121\n",
      "[42]\tvalidation_0-error:0.112847\tvalidation_1-error:0.135121\n",
      "[43]\tvalidation_0-error:0.11289\tvalidation_1-error:0.135224\n",
      "[44]\tvalidation_0-error:0.11289\tvalidation_1-error:0.135224\n",
      "[45]\tvalidation_0-error:0.11289\tvalidation_1-error:0.135224\n",
      "[46]\tvalidation_0-error:0.112847\tvalidation_1-error:0.135224\n",
      "[47]\tvalidation_0-error:0.112847\tvalidation_1-error:0.135224\n",
      "[48]\tvalidation_0-error:0.112847\tvalidation_1-error:0.135121\n",
      "[49]\tvalidation_0-error:0.112847\tvalidation_1-error:0.134712\n",
      "Stopping. Best iteration:\n",
      "[19]\tvalidation_0-error:0.112671\tvalidation_1-error:0.134405\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=0.9,\n",
       "              colsample_bytree=0.9, gamma=0, learning_rate=0.1,\n",
       "              max_delta_step=0, max_depth=3, min_child_weight=1, missing=None,\n",
       "              n_estimators=300, n_jobs=1, nthread=None,\n",
       "              objective='binary:logistic', random_state=0, reg_alpha=0,\n",
       "              reg_lambda=1.5, scale_pos_weight=1, seed=10, silent=True,\n",
       "              subsample=0.9)"
      ]
     },
     "execution_count": 118,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "xgbstackCV_grid.fit(X_train_stack, y_train, eval_metric=['error'], eval_set=[(X_train_stack, y_train),(X_test_stack, y_test)],early_stopping_rounds=30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.89      0.95      0.92      7429\n",
      "           1       0.81      0.61      0.70      2340\n",
      "\n",
      "    accuracy                           0.87      9769\n",
      "   macro avg       0.85      0.78      0.81      9769\n",
      "weighted avg       0.87      0.87      0.87      9769\n",
      "\n",
      "Accuracy on test data: 87.204422%\n",
      "Accuracy on training data: 89.110214%\n"
     ]
    }
   ],
   "source": [
    "print(metrics.classification_report(y_test, xgbstack.predict(X_test_stack)))\n",
    "print('Accuracy on test data: %f%%' % (metrics.accuracy_score(y_test, xgbstack.predict(X_test_stack))*100))\n",
    "print('Accuracy on training data: %f%%' % (metrics.accuracy_score(y_train, xgbstack.predict(X_train_stack))*100))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "metadata": {},
   "outputs": [],
   "source": [
    "from catboost import CatBoostClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "metadata": {},
   "outputs": [],
   "source": [
    "catb = CatBoostClassifier(learning_rate=0.3,iterations=400,verbose=1,random_seed=10,eval_metric='Accuracy',rsm=0.9)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0:\tlearn: 0.8477536\ttest: 0.8477536\ttest1: 0.8453271\tbest: 0.8453271 (0)\ttotal: 112ms\tremaining: 44.8s\n",
      "1:\tlearn: 0.8512197\ttest: 0.8512197\ttest1: 0.8458389\tbest: 0.8458389 (1)\ttotal: 162ms\tremaining: 32.2s\n",
      "2:\tlearn: 0.8495525\ttest: 0.8495525\ttest1: 0.8480909\tbest: 0.8480909 (2)\ttotal: 217ms\tremaining: 28.7s\n",
      "3:\tlearn: 0.8508687\ttest: 0.8508687\ttest1: 0.8507524\tbest: 0.8507524 (3)\ttotal: 384ms\tremaining: 38.1s\n",
      "4:\tlearn: 0.8545981\ttest: 0.8545981\ttest1: 0.8536186\tbest: 0.8536186 (4)\ttotal: 488ms\tremaining: 38.5s\n",
      "5:\tlearn: 0.8566602\ttest: 0.8566602\ttest1: 0.8543351\tbest: 0.8543351 (5)\ttotal: 550ms\tremaining: 36.1s\n",
      "6:\tlearn: 0.8561776\ttest: 0.8561776\ttest1: 0.8539257\tbest: 0.8543351 (5)\ttotal: 613ms\tremaining: 34.4s\n",
      "7:\tlearn: 0.8590734\ttest: 0.8590734\ttest1: 0.8540280\tbest: 0.8543351 (5)\ttotal: 687ms\tremaining: 33.7s\n",
      "8:\tlearn: 0.8589856\ttest: 0.8589856\ttest1: 0.8542328\tbest: 0.8543351 (5)\ttotal: 750ms\tremaining: 32.6s\n",
      "9:\tlearn: 0.8588540\ttest: 0.8588540\ttest1: 0.8539257\tbest: 0.8543351 (5)\ttotal: 864ms\tremaining: 33.7s\n",
      "10:\tlearn: 0.8600386\ttest: 0.8600386\ttest1: 0.8547446\tbest: 0.8547446 (10)\ttotal: 1.06s\tremaining: 37.5s\n",
      "11:\tlearn: 0.8622762\ttest: 0.8622762\ttest1: 0.8563824\tbest: 0.8563824 (11)\ttotal: 1.13s\tremaining: 36.6s\n",
      "12:\tlearn: 0.8636802\ttest: 0.8636802\ttest1: 0.8569966\tbest: 0.8569966 (12)\ttotal: 1.28s\tremaining: 38.1s\n",
      "13:\tlearn: 0.8643384\ttest: 0.8643384\ttest1: 0.8581226\tbest: 0.8581226 (13)\ttotal: 1.36s\tremaining: 37.4s\n",
      "14:\tlearn: 0.8633292\ttest: 0.8633292\ttest1: 0.8581226\tbest: 0.8581226 (13)\ttotal: 1.44s\tremaining: 36.9s\n",
      "15:\tlearn: 0.8650842\ttest: 0.8650842\ttest1: 0.8598628\tbest: 0.8598628 (15)\ttotal: 1.57s\tremaining: 37.7s\n",
      "16:\tlearn: 0.8651281\ttest: 0.8651281\ttest1: 0.8594534\tbest: 0.8598628 (15)\ttotal: 1.64s\tremaining: 37s\n",
      "17:\tlearn: 0.8654352\ttest: 0.8654352\ttest1: 0.8602723\tbest: 0.8602723 (17)\ttotal: 1.69s\tremaining: 36s\n",
      "18:\tlearn: 0.8658301\ttest: 0.8658301\ttest1: 0.8612959\tbest: 0.8612959 (18)\ttotal: 1.75s\tremaining: 35.1s\n",
      "19:\tlearn: 0.8664005\ttest: 0.8664005\ttest1: 0.8621149\tbest: 0.8621149 (19)\ttotal: 1.83s\tremaining: 34.9s\n",
      "20:\tlearn: 0.8664882\ttest: 0.8664882\ttest1: 0.8618078\tbest: 0.8621149 (19)\ttotal: 1.91s\tremaining: 34.5s\n",
      "21:\tlearn: 0.8668392\ttest: 0.8668392\ttest1: 0.8627290\tbest: 0.8627290 (21)\ttotal: 1.97s\tremaining: 33.9s\n",
      "22:\tlearn: 0.8667076\ttest: 0.8667076\ttest1: 0.8624219\tbest: 0.8627290 (21)\ttotal: 2.02s\tremaining: 33.2s\n",
      "23:\tlearn: 0.8671464\ttest: 0.8671464\ttest1: 0.8625243\tbest: 0.8627290 (21)\ttotal: 2.1s\tremaining: 32.9s\n",
      "24:\tlearn: 0.8674096\ttest: 0.8674096\ttest1: 0.8635480\tbest: 0.8635480 (24)\ttotal: 2.17s\tremaining: 32.5s\n",
      "25:\tlearn: 0.8674096\ttest: 0.8674096\ttest1: 0.8639574\tbest: 0.8639574 (25)\ttotal: 2.22s\tremaining: 31.9s\n",
      "26:\tlearn: 0.8674974\ttest: 0.8674974\ttest1: 0.8635480\tbest: 0.8639574 (25)\ttotal: 2.26s\tremaining: 31.3s\n",
      "27:\tlearn: 0.8677167\ttest: 0.8677167\ttest1: 0.8641621\tbest: 0.8641621 (27)\ttotal: 2.31s\tremaining: 30.7s\n",
      "28:\tlearn: 0.8681555\ttest: 0.8681555\ttest1: 0.8645716\tbest: 0.8645716 (28)\ttotal: 2.36s\tremaining: 30.3s\n",
      "29:\tlearn: 0.8680677\ttest: 0.8680677\ttest1: 0.8647763\tbest: 0.8647763 (29)\ttotal: 2.42s\tremaining: 29.9s\n",
      "30:\tlearn: 0.8682871\ttest: 0.8682871\ttest1: 0.8649811\tbest: 0.8649811 (30)\ttotal: 2.49s\tremaining: 29.6s\n",
      "31:\tlearn: 0.8682871\ttest: 0.8682871\ttest1: 0.8648787\tbest: 0.8649811 (30)\ttotal: 2.56s\tremaining: 29.5s\n",
      "32:\tlearn: 0.8684187\ttest: 0.8684187\ttest1: 0.8645716\tbest: 0.8649811 (30)\ttotal: 2.62s\tremaining: 29.2s\n",
      "33:\tlearn: 0.8697789\ttest: 0.8697789\ttest1: 0.8659023\tbest: 0.8659023 (33)\ttotal: 2.68s\tremaining: 28.8s\n",
      "34:\tlearn: 0.8700421\ttest: 0.8700421\ttest1: 0.8661071\tbest: 0.8661071 (34)\ttotal: 2.73s\tremaining: 28.4s\n",
      "35:\tlearn: 0.8707441\ttest: 0.8707441\ttest1: 0.8665165\tbest: 0.8665165 (35)\ttotal: 2.78s\tremaining: 28.1s\n",
      "36:\tlearn: 0.8711390\ttest: 0.8711390\ttest1: 0.8668236\tbest: 0.8668236 (36)\ttotal: 2.83s\tremaining: 27.7s\n",
      "37:\tlearn: 0.8717094\ttest: 0.8717094\ttest1: 0.8679496\tbest: 0.8679496 (37)\ttotal: 2.94s\tremaining: 28s\n",
      "38:\tlearn: 0.8718849\ttest: 0.8718849\ttest1: 0.8679496\tbest: 0.8679496 (37)\ttotal: 3.11s\tremaining: 28.8s\n",
      "39:\tlearn: 0.8719726\ttest: 0.8719726\ttest1: 0.8679496\tbest: 0.8679496 (37)\ttotal: 3.24s\tremaining: 29.2s\n",
      "40:\tlearn: 0.8721920\ttest: 0.8721920\ttest1: 0.8681544\tbest: 0.8681544 (40)\ttotal: 3.3s\tremaining: 28.9s\n",
      "41:\tlearn: 0.8722359\ttest: 0.8722359\ttest1: 0.8682567\tbest: 0.8682567 (41)\ttotal: 3.37s\tremaining: 28.7s\n",
      "42:\tlearn: 0.8723236\ttest: 0.8723236\ttest1: 0.8684615\tbest: 0.8684615 (42)\ttotal: 3.42s\tremaining: 28.4s\n",
      "43:\tlearn: 0.8724114\ttest: 0.8724114\ttest1: 0.8686662\tbest: 0.8686662 (43)\ttotal: 3.47s\tremaining: 28s\n",
      "44:\tlearn: 0.8728062\ttest: 0.8728062\ttest1: 0.8689733\tbest: 0.8689733 (44)\ttotal: 3.51s\tremaining: 27.7s\n",
      "45:\tlearn: 0.8730695\ttest: 0.8730695\ttest1: 0.8693827\tbest: 0.8693827 (45)\ttotal: 3.57s\tremaining: 27.5s\n",
      "46:\tlearn: 0.8728940\ttest: 0.8728940\ttest1: 0.8698946\tbest: 0.8698946 (46)\ttotal: 3.71s\tremaining: 27.9s\n",
      "47:\tlearn: 0.8736399\ttest: 0.8736399\ttest1: 0.8695875\tbest: 0.8698946 (46)\ttotal: 3.9s\tremaining: 28.6s\n",
      "48:\tlearn: 0.8742541\ttest: 0.8742541\ttest1: 0.8710206\tbest: 0.8710206 (48)\ttotal: 4.01s\tremaining: 28.7s\n",
      "49:\tlearn: 0.8742980\ttest: 0.8742980\ttest1: 0.8708158\tbest: 0.8710206 (48)\ttotal: 4.12s\tremaining: 28.9s\n",
      "50:\tlearn: 0.8743857\ttest: 0.8743857\ttest1: 0.8708158\tbest: 0.8710206 (48)\ttotal: 4.29s\tremaining: 29.3s\n",
      "51:\tlearn: 0.8743857\ttest: 0.8743857\ttest1: 0.8710206\tbest: 0.8710206 (48)\ttotal: 4.4s\tremaining: 29.5s\n",
      "52:\tlearn: 0.8744296\ttest: 0.8744296\ttest1: 0.8708158\tbest: 0.8710206 (48)\ttotal: 4.49s\tremaining: 29.4s\n",
      "53:\tlearn: 0.8746490\ttest: 0.8746490\ttest1: 0.8712253\tbest: 0.8712253 (53)\ttotal: 4.59s\tremaining: 29.4s\n",
      "54:\tlearn: 0.8746051\ttest: 0.8746051\ttest1: 0.8718395\tbest: 0.8718395 (54)\ttotal: 4.7s\tremaining: 29.5s\n",
      "55:\tlearn: 0.8750000\ttest: 0.8750000\ttest1: 0.8721466\tbest: 0.8721466 (55)\ttotal: 4.85s\tremaining: 29.8s\n",
      "56:\tlearn: 0.8753949\ttest: 0.8753949\ttest1: 0.8720442\tbest: 0.8721466 (55)\ttotal: 4.97s\tremaining: 29.9s\n",
      "57:\tlearn: 0.8754826\ttest: 0.8754826\ttest1: 0.8724537\tbest: 0.8724537 (57)\ttotal: 5.02s\tremaining: 29.6s\n",
      "58:\tlearn: 0.8754826\ttest: 0.8754826\ttest1: 0.8726584\tbest: 0.8726584 (58)\ttotal: 5.08s\tremaining: 29.3s\n",
      "59:\tlearn: 0.8759214\ttest: 0.8759214\ttest1: 0.8717371\tbest: 0.8726584 (58)\ttotal: 5.16s\tremaining: 29.3s\n",
      "60:\tlearn: 0.8758775\ttest: 0.8758775\ttest1: 0.8717371\tbest: 0.8726584 (58)\ttotal: 5.22s\tremaining: 29s\n",
      "61:\tlearn: 0.8760091\ttest: 0.8760091\ttest1: 0.8716348\tbest: 0.8726584 (58)\ttotal: 5.3s\tremaining: 28.9s\n",
      "62:\tlearn: 0.8755704\ttest: 0.8755704\ttest1: 0.8716348\tbest: 0.8726584 (58)\ttotal: 5.35s\tremaining: 28.6s\n",
      "63:\tlearn: 0.8758336\ttest: 0.8758336\ttest1: 0.8716348\tbest: 0.8726584 (58)\ttotal: 5.4s\tremaining: 28.4s\n",
      "64:\tlearn: 0.8759214\ttest: 0.8759214\ttest1: 0.8719419\tbest: 0.8726584 (58)\ttotal: 5.47s\tremaining: 28.2s\n",
      "65:\tlearn: 0.8759653\ttest: 0.8759653\ttest1: 0.8719419\tbest: 0.8726584 (58)\ttotal: 5.52s\tremaining: 27.9s\n",
      "66:\tlearn: 0.8760969\ttest: 0.8760969\ttest1: 0.8720442\tbest: 0.8726584 (58)\ttotal: 5.57s\tremaining: 27.7s\n",
      "67:\tlearn: 0.8763601\ttest: 0.8763601\ttest1: 0.8717371\tbest: 0.8726584 (58)\ttotal: 5.61s\tremaining: 27.4s\n",
      "68:\tlearn: 0.8764918\ttest: 0.8764918\ttest1: 0.8718395\tbest: 0.8726584 (58)\ttotal: 5.66s\tremaining: 27.2s\n",
      "69:\tlearn: 0.8768866\ttest: 0.8768866\ttest1: 0.8718395\tbest: 0.8726584 (58)\ttotal: 5.72s\tremaining: 27s\n",
      "70:\tlearn: 0.8768428\ttest: 0.8768428\ttest1: 0.8720442\tbest: 0.8726584 (58)\ttotal: 5.77s\tremaining: 26.7s\n",
      "71:\tlearn: 0.8767989\ttest: 0.8767989\ttest1: 0.8721466\tbest: 0.8726584 (58)\ttotal: 5.82s\tremaining: 26.5s\n",
      "72:\tlearn: 0.8768428\ttest: 0.8768428\ttest1: 0.8719419\tbest: 0.8726584 (58)\ttotal: 5.87s\tremaining: 26.3s\n",
      "73:\tlearn: 0.8768866\ttest: 0.8768866\ttest1: 0.8722490\tbest: 0.8726584 (58)\ttotal: 5.92s\tremaining: 26.1s\n",
      "74:\tlearn: 0.8772815\ttest: 0.8772815\ttest1: 0.8724537\tbest: 0.8726584 (58)\ttotal: 5.98s\tremaining: 25.9s\n",
      "75:\tlearn: 0.8774131\ttest: 0.8774131\ttest1: 0.8725560\tbest: 0.8726584 (58)\ttotal: 6.05s\tremaining: 25.8s\n",
      "76:\tlearn: 0.8771499\ttest: 0.8771499\ttest1: 0.8727608\tbest: 0.8727608 (76)\ttotal: 6.18s\tremaining: 25.9s\n",
      "77:\tlearn: 0.8771499\ttest: 0.8771499\ttest1: 0.8725560\tbest: 0.8727608 (76)\ttotal: 6.25s\tremaining: 25.8s\n",
      "78:\tlearn: 0.8772376\ttest: 0.8772376\ttest1: 0.8730679\tbest: 0.8730679 (78)\ttotal: 6.3s\tremaining: 25.6s\n",
      "79:\tlearn: 0.8773693\ttest: 0.8773693\ttest1: 0.8730679\tbest: 0.8730679 (78)\ttotal: 6.35s\tremaining: 25.4s\n",
      "80:\tlearn: 0.8776325\ttest: 0.8776325\ttest1: 0.8723513\tbest: 0.8730679 (78)\ttotal: 6.4s\tremaining: 25.2s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "81:\tlearn: 0.8777641\ttest: 0.8777641\ttest1: 0.8723513\tbest: 0.8730679 (78)\ttotal: 6.45s\tremaining: 25s\n",
      "82:\tlearn: 0.8777203\ttest: 0.8777203\ttest1: 0.8724537\tbest: 0.8730679 (78)\ttotal: 6.5s\tremaining: 24.8s\n",
      "83:\tlearn: 0.8777641\ttest: 0.8777641\ttest1: 0.8725560\tbest: 0.8730679 (78)\ttotal: 6.55s\tremaining: 24.6s\n",
      "84:\tlearn: 0.8777641\ttest: 0.8777641\ttest1: 0.8725560\tbest: 0.8730679 (78)\ttotal: 6.6s\tremaining: 24.4s\n",
      "85:\tlearn: 0.8778958\ttest: 0.8778958\ttest1: 0.8724537\tbest: 0.8730679 (78)\ttotal: 6.65s\tremaining: 24.3s\n",
      "86:\tlearn: 0.8780274\ttest: 0.8780274\ttest1: 0.8726584\tbest: 0.8730679 (78)\ttotal: 6.79s\tremaining: 24.4s\n",
      "87:\tlearn: 0.8782906\ttest: 0.8782906\ttest1: 0.8729655\tbest: 0.8730679 (78)\ttotal: 6.89s\tremaining: 24.4s\n",
      "88:\tlearn: 0.8782468\ttest: 0.8782468\ttest1: 0.8729655\tbest: 0.8730679 (78)\ttotal: 7.04s\tremaining: 24.6s\n",
      "89:\tlearn: 0.8786416\ttest: 0.8786416\ttest1: 0.8730679\tbest: 0.8730679 (78)\ttotal: 7.14s\tremaining: 24.6s\n",
      "90:\tlearn: 0.8786416\ttest: 0.8786416\ttest1: 0.8727608\tbest: 0.8730679 (78)\ttotal: 7.26s\tremaining: 24.7s\n",
      "91:\tlearn: 0.8786855\ttest: 0.8786855\ttest1: 0.8727608\tbest: 0.8730679 (78)\ttotal: 7.34s\tremaining: 24.6s\n",
      "92:\tlearn: 0.8785539\ttest: 0.8785539\ttest1: 0.8729655\tbest: 0.8730679 (78)\ttotal: 7.41s\tremaining: 24.5s\n",
      "93:\tlearn: 0.8786416\ttest: 0.8786416\ttest1: 0.8729655\tbest: 0.8730679 (78)\ttotal: 7.46s\tremaining: 24.3s\n",
      "94:\tlearn: 0.8785978\ttest: 0.8785978\ttest1: 0.8728631\tbest: 0.8730679 (78)\ttotal: 7.51s\tremaining: 24.1s\n",
      "95:\tlearn: 0.8788610\ttest: 0.8788610\ttest1: 0.8729655\tbest: 0.8730679 (78)\ttotal: 7.56s\tremaining: 23.9s\n",
      "96:\tlearn: 0.8788610\ttest: 0.8788610\ttest1: 0.8732726\tbest: 0.8732726 (96)\ttotal: 7.62s\tremaining: 23.8s\n",
      "97:\tlearn: 0.8789049\ttest: 0.8789049\ttest1: 0.8732726\tbest: 0.8732726 (96)\ttotal: 7.67s\tremaining: 23.6s\n",
      "98:\tlearn: 0.8789926\ttest: 0.8789926\ttest1: 0.8730679\tbest: 0.8732726 (96)\ttotal: 7.72s\tremaining: 23.5s\n",
      "99:\tlearn: 0.8789488\ttest: 0.8789488\ttest1: 0.8729655\tbest: 0.8732726 (96)\ttotal: 7.76s\tremaining: 23.3s\n",
      "100:\tlearn: 0.8790365\ttest: 0.8790365\ttest1: 0.8733750\tbest: 0.8733750 (100)\ttotal: 7.81s\tremaining: 23.1s\n",
      "101:\tlearn: 0.8792120\ttest: 0.8792120\ttest1: 0.8733750\tbest: 0.8733750 (100)\ttotal: 7.86s\tremaining: 22.9s\n",
      "102:\tlearn: 0.8792120\ttest: 0.8792120\ttest1: 0.8734773\tbest: 0.8734773 (102)\ttotal: 7.9s\tremaining: 22.8s\n",
      "103:\tlearn: 0.8790365\ttest: 0.8790365\ttest1: 0.8735797\tbest: 0.8735797 (103)\ttotal: 7.96s\tremaining: 22.6s\n",
      "104:\tlearn: 0.8790365\ttest: 0.8790365\ttest1: 0.8736821\tbest: 0.8736821 (104)\ttotal: 8s\tremaining: 22.5s\n",
      "105:\tlearn: 0.8790365\ttest: 0.8790365\ttest1: 0.8737844\tbest: 0.8737844 (105)\ttotal: 8.05s\tremaining: 22.3s\n",
      "106:\tlearn: 0.8791243\ttest: 0.8791243\ttest1: 0.8739891\tbest: 0.8739891 (106)\ttotal: 8.12s\tremaining: 22.2s\n",
      "107:\tlearn: 0.8792120\ttest: 0.8792120\ttest1: 0.8735797\tbest: 0.8739891 (106)\ttotal: 8.17s\tremaining: 22.1s\n",
      "108:\tlearn: 0.8792559\ttest: 0.8792559\ttest1: 0.8733750\tbest: 0.8739891 (106)\ttotal: 8.22s\tremaining: 22s\n",
      "109:\tlearn: 0.8792998\ttest: 0.8792998\ttest1: 0.8732726\tbest: 0.8739891 (106)\ttotal: 8.27s\tremaining: 21.8s\n",
      "110:\tlearn: 0.8792120\ttest: 0.8792120\ttest1: 0.8733750\tbest: 0.8739891 (106)\ttotal: 8.31s\tremaining: 21.6s\n",
      "111:\tlearn: 0.8792120\ttest: 0.8792120\ttest1: 0.8732726\tbest: 0.8739891 (106)\ttotal: 8.36s\tremaining: 21.5s\n",
      "112:\tlearn: 0.8791243\ttest: 0.8791243\ttest1: 0.8730679\tbest: 0.8739891 (106)\ttotal: 8.41s\tremaining: 21.4s\n",
      "113:\tlearn: 0.8791243\ttest: 0.8791243\ttest1: 0.8730679\tbest: 0.8739891 (106)\ttotal: 8.46s\tremaining: 21.2s\n",
      "114:\tlearn: 0.8792998\ttest: 0.8792998\ttest1: 0.8733750\tbest: 0.8739891 (106)\ttotal: 8.53s\tremaining: 21.1s\n",
      "115:\tlearn: 0.8792998\ttest: 0.8792998\ttest1: 0.8734773\tbest: 0.8739891 (106)\ttotal: 8.59s\tremaining: 21s\n",
      "116:\tlearn: 0.8792120\ttest: 0.8792120\ttest1: 0.8733750\tbest: 0.8739891 (106)\ttotal: 8.68s\tremaining: 21s\n",
      "117:\tlearn: 0.8794314\ttest: 0.8794314\ttest1: 0.8733750\tbest: 0.8739891 (106)\ttotal: 8.82s\tremaining: 21.1s\n",
      "118:\tlearn: 0.8792120\ttest: 0.8792120\ttest1: 0.8738868\tbest: 0.8739891 (106)\ttotal: 9s\tremaining: 21.2s\n",
      "119:\tlearn: 0.8792998\ttest: 0.8792998\ttest1: 0.8736821\tbest: 0.8739891 (106)\ttotal: 9.13s\tremaining: 21.3s\n",
      "120:\tlearn: 0.8793436\ttest: 0.8793436\ttest1: 0.8737844\tbest: 0.8739891 (106)\ttotal: 9.23s\tremaining: 21.3s\n",
      "121:\tlearn: 0.8792120\ttest: 0.8792120\ttest1: 0.8735797\tbest: 0.8739891 (106)\ttotal: 9.28s\tremaining: 21.1s\n",
      "122:\tlearn: 0.8796508\ttest: 0.8796508\ttest1: 0.8734773\tbest: 0.8739891 (106)\ttotal: 9.33s\tremaining: 21s\n",
      "123:\tlearn: 0.8796069\ttest: 0.8796069\ttest1: 0.8734773\tbest: 0.8739891 (106)\ttotal: 9.38s\tremaining: 20.9s\n",
      "124:\tlearn: 0.8795630\ttest: 0.8795630\ttest1: 0.8736821\tbest: 0.8739891 (106)\ttotal: 9.43s\tremaining: 20.7s\n",
      "125:\tlearn: 0.8800456\ttest: 0.8800456\ttest1: 0.8731702\tbest: 0.8739891 (106)\ttotal: 9.49s\tremaining: 20.6s\n",
      "126:\tlearn: 0.8800456\ttest: 0.8800456\ttest1: 0.8732726\tbest: 0.8739891 (106)\ttotal: 9.54s\tremaining: 20.5s\n",
      "127:\tlearn: 0.8800456\ttest: 0.8800456\ttest1: 0.8733750\tbest: 0.8739891 (106)\ttotal: 9.6s\tremaining: 20.4s\n",
      "128:\tlearn: 0.8800456\ttest: 0.8800456\ttest1: 0.8730679\tbest: 0.8739891 (106)\ttotal: 9.65s\tremaining: 20.3s\n",
      "129:\tlearn: 0.8798263\ttest: 0.8798263\ttest1: 0.8732726\tbest: 0.8739891 (106)\ttotal: 9.7s\tremaining: 20.1s\n",
      "130:\tlearn: 0.8799140\ttest: 0.8799140\ttest1: 0.8733750\tbest: 0.8739891 (106)\ttotal: 9.75s\tremaining: 20s\n",
      "131:\tlearn: 0.8799579\ttest: 0.8799579\ttest1: 0.8731702\tbest: 0.8739891 (106)\ttotal: 9.8s\tremaining: 19.9s\n",
      "132:\tlearn: 0.8799579\ttest: 0.8799579\ttest1: 0.8731702\tbest: 0.8739891 (106)\ttotal: 9.88s\tremaining: 19.8s\n",
      "133:\tlearn: 0.8804844\ttest: 0.8804844\ttest1: 0.8735797\tbest: 0.8739891 (106)\ttotal: 10s\tremaining: 19.9s\n",
      "134:\tlearn: 0.8805283\ttest: 0.8805283\ttest1: 0.8735797\tbest: 0.8739891 (106)\ttotal: 10.1s\tremaining: 19.8s\n",
      "135:\tlearn: 0.8806160\ttest: 0.8806160\ttest1: 0.8737844\tbest: 0.8739891 (106)\ttotal: 10.2s\tremaining: 19.7s\n",
      "136:\tlearn: 0.8805721\ttest: 0.8805721\ttest1: 0.8737844\tbest: 0.8739891 (106)\ttotal: 10.2s\tremaining: 19.6s\n",
      "137:\tlearn: 0.8804844\ttest: 0.8804844\ttest1: 0.8737844\tbest: 0.8739891 (106)\ttotal: 10.3s\tremaining: 19.5s\n",
      "138:\tlearn: 0.8806160\ttest: 0.8806160\ttest1: 0.8738868\tbest: 0.8739891 (106)\ttotal: 10.3s\tremaining: 19.3s\n",
      "139:\tlearn: 0.8807915\ttest: 0.8807915\ttest1: 0.8742962\tbest: 0.8742962 (139)\ttotal: 10.4s\tremaining: 19.2s\n",
      "140:\tlearn: 0.8808354\ttest: 0.8808354\ttest1: 0.8741939\tbest: 0.8742962 (139)\ttotal: 10.4s\tremaining: 19.1s\n",
      "141:\tlearn: 0.8808793\ttest: 0.8808793\ttest1: 0.8739891\tbest: 0.8742962 (139)\ttotal: 10.5s\tremaining: 19s\n",
      "142:\tlearn: 0.8809670\ttest: 0.8809670\ttest1: 0.8739891\tbest: 0.8742962 (139)\ttotal: 10.5s\tremaining: 18.9s\n",
      "143:\tlearn: 0.8810548\ttest: 0.8810548\ttest1: 0.8742962\tbest: 0.8742962 (139)\ttotal: 10.6s\tremaining: 18.8s\n",
      "144:\tlearn: 0.8808793\ttest: 0.8808793\ttest1: 0.8741939\tbest: 0.8742962 (139)\ttotal: 10.6s\tremaining: 18.7s\n",
      "145:\tlearn: 0.8810109\ttest: 0.8810109\ttest1: 0.8741939\tbest: 0.8742962 (139)\ttotal: 10.7s\tremaining: 18.5s\n",
      "146:\tlearn: 0.8810109\ttest: 0.8810109\ttest1: 0.8745010\tbest: 0.8745010 (146)\ttotal: 10.7s\tremaining: 18.4s\n",
      "147:\tlearn: 0.8810109\ttest: 0.8810109\ttest1: 0.8743986\tbest: 0.8745010 (146)\ttotal: 10.8s\tremaining: 18.3s\n",
      "148:\tlearn: 0.8810548\ttest: 0.8810548\ttest1: 0.8745010\tbest: 0.8745010 (146)\ttotal: 10.8s\tremaining: 18.2s\n",
      "149:\tlearn: 0.8810548\ttest: 0.8810548\ttest1: 0.8746033\tbest: 0.8746033 (149)\ttotal: 10.9s\tremaining: 18.1s\n",
      "150:\tlearn: 0.8810986\ttest: 0.8810986\ttest1: 0.8746033\tbest: 0.8746033 (149)\ttotal: 10.9s\tremaining: 18s\n",
      "151:\tlearn: 0.8811425\ttest: 0.8811425\ttest1: 0.8743986\tbest: 0.8746033 (149)\ttotal: 11s\tremaining: 17.9s\n",
      "152:\tlearn: 0.8809670\ttest: 0.8809670\ttest1: 0.8749104\tbest: 0.8749104 (152)\ttotal: 11.1s\tremaining: 17.9s\n",
      "153:\tlearn: 0.8809231\ttest: 0.8809231\ttest1: 0.8748081\tbest: 0.8749104 (152)\ttotal: 11.1s\tremaining: 17.8s\n",
      "154:\tlearn: 0.8808793\ttest: 0.8808793\ttest1: 0.8747057\tbest: 0.8749104 (152)\ttotal: 11.2s\tremaining: 17.7s\n",
      "155:\tlearn: 0.8809231\ttest: 0.8809231\ttest1: 0.8749104\tbest: 0.8749104 (152)\ttotal: 11.2s\tremaining: 17.6s\n",
      "156:\tlearn: 0.8809231\ttest: 0.8809231\ttest1: 0.8748081\tbest: 0.8749104 (152)\ttotal: 11.3s\tremaining: 17.4s\n",
      "157:\tlearn: 0.8808354\ttest: 0.8808354\ttest1: 0.8748081\tbest: 0.8749104 (152)\ttotal: 11.3s\tremaining: 17.3s\n",
      "158:\tlearn: 0.8808793\ttest: 0.8808793\ttest1: 0.8749104\tbest: 0.8749104 (152)\ttotal: 11.4s\tremaining: 17.2s\n",
      "159:\tlearn: 0.8808354\ttest: 0.8808354\ttest1: 0.8741939\tbest: 0.8749104 (152)\ttotal: 11.4s\tremaining: 17.1s\n",
      "160:\tlearn: 0.8809670\ttest: 0.8809670\ttest1: 0.8741939\tbest: 0.8749104 (152)\ttotal: 11.5s\tremaining: 17s\n",
      "161:\tlearn: 0.8815813\ttest: 0.8815813\ttest1: 0.8740915\tbest: 0.8749104 (152)\ttotal: 11.5s\tremaining: 16.9s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "162:\tlearn: 0.8815813\ttest: 0.8815813\ttest1: 0.8740915\tbest: 0.8749104 (152)\ttotal: 11.6s\tremaining: 16.8s\n",
      "163:\tlearn: 0.8813619\ttest: 0.8813619\ttest1: 0.8734773\tbest: 0.8749104 (152)\ttotal: 11.6s\tremaining: 16.7s\n",
      "164:\tlearn: 0.8814058\ttest: 0.8814058\ttest1: 0.8734773\tbest: 0.8749104 (152)\ttotal: 11.7s\tremaining: 16.6s\n",
      "165:\tlearn: 0.8817568\ttest: 0.8817568\ttest1: 0.8738868\tbest: 0.8749104 (152)\ttotal: 11.7s\tremaining: 16.5s\n",
      "166:\tlearn: 0.8816251\ttest: 0.8816251\ttest1: 0.8741939\tbest: 0.8749104 (152)\ttotal: 11.8s\tremaining: 16.4s\n",
      "167:\tlearn: 0.8815374\ttest: 0.8815374\ttest1: 0.8741939\tbest: 0.8749104 (152)\ttotal: 11.8s\tremaining: 16.3s\n",
      "168:\tlearn: 0.8814935\ttest: 0.8814935\ttest1: 0.8742962\tbest: 0.8749104 (152)\ttotal: 11.9s\tremaining: 16.2s\n",
      "169:\tlearn: 0.8814496\ttest: 0.8814496\ttest1: 0.8743986\tbest: 0.8749104 (152)\ttotal: 11.9s\tremaining: 16.1s\n",
      "170:\tlearn: 0.8814496\ttest: 0.8814496\ttest1: 0.8743986\tbest: 0.8749104 (152)\ttotal: 12s\tremaining: 16s\n",
      "171:\tlearn: 0.8814058\ttest: 0.8814058\ttest1: 0.8743986\tbest: 0.8749104 (152)\ttotal: 12s\tremaining: 16s\n",
      "172:\tlearn: 0.8814496\ttest: 0.8814496\ttest1: 0.8742962\tbest: 0.8749104 (152)\ttotal: 12.1s\tremaining: 15.8s\n",
      "173:\tlearn: 0.8814058\ttest: 0.8814058\ttest1: 0.8743986\tbest: 0.8749104 (152)\ttotal: 12.1s\tremaining: 15.8s\n",
      "174:\tlearn: 0.8815374\ttest: 0.8815374\ttest1: 0.8743986\tbest: 0.8749104 (152)\ttotal: 12.2s\tremaining: 15.7s\n",
      "175:\tlearn: 0.8815813\ttest: 0.8815813\ttest1: 0.8743986\tbest: 0.8749104 (152)\ttotal: 12.2s\tremaining: 15.6s\n",
      "176:\tlearn: 0.8813180\ttest: 0.8813180\ttest1: 0.8742962\tbest: 0.8749104 (152)\ttotal: 12.3s\tremaining: 15.5s\n",
      "177:\tlearn: 0.8813180\ttest: 0.8813180\ttest1: 0.8743986\tbest: 0.8749104 (152)\ttotal: 12.3s\tremaining: 15.4s\n",
      "178:\tlearn: 0.8813180\ttest: 0.8813180\ttest1: 0.8743986\tbest: 0.8749104 (152)\ttotal: 12.4s\tremaining: 15.3s\n",
      "179:\tlearn: 0.8813180\ttest: 0.8813180\ttest1: 0.8743986\tbest: 0.8749104 (152)\ttotal: 12.4s\tremaining: 15.2s\n",
      "180:\tlearn: 0.8813619\ttest: 0.8813619\ttest1: 0.8742962\tbest: 0.8749104 (152)\ttotal: 12.4s\tremaining: 15s\n",
      "181:\tlearn: 0.8815374\ttest: 0.8815374\ttest1: 0.8743986\tbest: 0.8749104 (152)\ttotal: 12.5s\tremaining: 14.9s\n",
      "182:\tlearn: 0.8816690\ttest: 0.8816690\ttest1: 0.8741939\tbest: 0.8749104 (152)\ttotal: 12.5s\tremaining: 14.9s\n",
      "183:\tlearn: 0.8815813\ttest: 0.8815813\ttest1: 0.8742962\tbest: 0.8749104 (152)\ttotal: 12.6s\tremaining: 14.8s\n",
      "184:\tlearn: 0.8816251\ttest: 0.8816251\ttest1: 0.8742962\tbest: 0.8749104 (152)\ttotal: 12.6s\tremaining: 14.7s\n",
      "185:\tlearn: 0.8816251\ttest: 0.8816251\ttest1: 0.8742962\tbest: 0.8749104 (152)\ttotal: 12.7s\tremaining: 14.6s\n",
      "186:\tlearn: 0.8815813\ttest: 0.8815813\ttest1: 0.8743986\tbest: 0.8749104 (152)\ttotal: 12.7s\tremaining: 14.5s\n",
      "187:\tlearn: 0.8815813\ttest: 0.8815813\ttest1: 0.8743986\tbest: 0.8749104 (152)\ttotal: 12.7s\tremaining: 14.4s\n",
      "188:\tlearn: 0.8815813\ttest: 0.8815813\ttest1: 0.8743986\tbest: 0.8749104 (152)\ttotal: 12.8s\tremaining: 14.3s\n",
      "189:\tlearn: 0.8815813\ttest: 0.8815813\ttest1: 0.8743986\tbest: 0.8749104 (152)\ttotal: 12.9s\tremaining: 14.3s\n",
      "190:\tlearn: 0.8815813\ttest: 0.8815813\ttest1: 0.8743986\tbest: 0.8749104 (152)\ttotal: 13s\tremaining: 14.3s\n",
      "191:\tlearn: 0.8816251\ttest: 0.8816251\ttest1: 0.8743986\tbest: 0.8749104 (152)\ttotal: 13.2s\tremaining: 14.3s\n",
      "192:\tlearn: 0.8816690\ttest: 0.8816690\ttest1: 0.8743986\tbest: 0.8749104 (152)\ttotal: 13.2s\tremaining: 14.2s\n",
      "193:\tlearn: 0.8817129\ttest: 0.8817129\ttest1: 0.8742962\tbest: 0.8749104 (152)\ttotal: 13.3s\tremaining: 14.1s\n",
      "194:\tlearn: 0.8816690\ttest: 0.8816690\ttest1: 0.8743986\tbest: 0.8749104 (152)\ttotal: 13.4s\tremaining: 14.1s\n",
      "195:\tlearn: 0.8815813\ttest: 0.8815813\ttest1: 0.8743986\tbest: 0.8749104 (152)\ttotal: 13.5s\tremaining: 14.1s\n",
      "196:\tlearn: 0.8816251\ttest: 0.8816251\ttest1: 0.8746033\tbest: 0.8749104 (152)\ttotal: 13.7s\tremaining: 14.1s\n",
      "197:\tlearn: 0.8816251\ttest: 0.8816251\ttest1: 0.8745010\tbest: 0.8749104 (152)\ttotal: 13.7s\tremaining: 14s\n",
      "198:\tlearn: 0.8816251\ttest: 0.8816251\ttest1: 0.8746033\tbest: 0.8749104 (152)\ttotal: 13.8s\tremaining: 13.9s\n",
      "199:\tlearn: 0.8818884\ttest: 0.8818884\ttest1: 0.8741939\tbest: 0.8749104 (152)\ttotal: 14s\tremaining: 14s\n",
      "200:\tlearn: 0.8817129\ttest: 0.8817129\ttest1: 0.8741939\tbest: 0.8749104 (152)\ttotal: 14.2s\tremaining: 14s\n",
      "201:\tlearn: 0.8818445\ttest: 0.8818445\ttest1: 0.8740915\tbest: 0.8749104 (152)\ttotal: 14.2s\tremaining: 14s\n",
      "202:\tlearn: 0.8819323\ttest: 0.8819323\ttest1: 0.8740915\tbest: 0.8749104 (152)\ttotal: 14.3s\tremaining: 13.9s\n",
      "Stopped by overfitting detector  (50 iterations wait)\n",
      "\n",
      "bestTest = 0.874910431\n",
      "bestIteration = 152\n",
      "\n",
      "Shrink model to first 153 iterations.\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<catboost.core.CatBoostClassifier at 0x1a267732b0>"
      ]
     },
     "execution_count": 122,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "catb.fit(X_train,y_train,eval_set=[(X_train,y_train), (X_test,y_test)],early_stopping_rounds=50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 123,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<catboost.core.CatBoostClassifier object at 0x1a267732b0>\n",
      "confusion matrix of training data\n",
      "[[16420   871]\n",
      " [ 1842  3659]]\n",
      "confusion matrix of test data\n",
      "[[7058  371]\n",
      " [ 851 1489]]\n",
      "classification report of test data\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.89      0.95      0.92      7429\n",
      "           1       0.80      0.64      0.71      2340\n",
      "\n",
      "    accuracy                           0.87      9769\n",
      "   macro avg       0.85      0.79      0.81      9769\n",
      "weighted avg       0.87      0.87      0.87      9769\n",
      "\n",
      "Accuracy on test data: 87.491043%\n",
      "Accuracy on training data: 88.096701%\n",
      "Area under the ROC curve : 0.793193\n"
     ]
    }
   ],
   "source": [
    "show_classifier_metrics(catb)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 167,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Doing grid search\n",
      "Fitting 4 folds for each of 4 candidates, totalling 16 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 4 concurrent workers.\n",
      "[Parallel(n_jobs=-1)]: Done  16 out of  16 | elapsed: 11.4min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0:\tlearn: 0.8489382\ttotal: 302ms\tremaining: 1m 30s\n",
      "1:\tlearn: 0.8508687\ttotal: 555ms\tremaining: 1m 22s\n",
      "2:\tlearn: 0.8531064\ttotal: 816ms\tremaining: 1m 20s\n",
      "3:\tlearn: 0.8520534\ttotal: 1.06s\tremaining: 1m 18s\n",
      "4:\tlearn: 0.8542471\ttotal: 1.31s\tremaining: 1m 17s\n",
      "5:\tlearn: 0.8554756\ttotal: 1.4s\tremaining: 1m 8s\n",
      "6:\tlearn: 0.8550369\ttotal: 1.65s\tremaining: 1m 8s\n",
      "7:\tlearn: 0.8545981\ttotal: 1.74s\tremaining: 1m 3s\n",
      "8:\tlearn: 0.8567041\ttotal: 1.98s\tremaining: 1m 3s\n",
      "9:\tlearn: 0.8580204\ttotal: 2.23s\tremaining: 1m 4s\n",
      "10:\tlearn: 0.8585030\ttotal: 2.48s\tremaining: 1m 5s\n",
      "11:\tlearn: 0.8596437\ttotal: 2.73s\tremaining: 1m 5s\n",
      "12:\tlearn: 0.8604335\ttotal: 2.96s\tremaining: 1m 5s\n",
      "13:\tlearn: 0.8607845\ttotal: 3.2s\tremaining: 1m 5s\n",
      "14:\tlearn: 0.8604774\ttotal: 3.44s\tremaining: 1m 5s\n",
      "15:\tlearn: 0.8617497\ttotal: 3.67s\tremaining: 1m 5s\n",
      "16:\tlearn: 0.8622762\ttotal: 3.77s\tremaining: 1m 2s\n",
      "17:\tlearn: 0.8623640\ttotal: 4.01s\tremaining: 1m 2s\n",
      "18:\tlearn: 0.8632415\ttotal: 4.23s\tremaining: 1m 2s\n",
      "19:\tlearn: 0.8636802\ttotal: 4.46s\tremaining: 1m 2s\n",
      "20:\tlearn: 0.8637241\ttotal: 4.71s\tremaining: 1m 2s\n",
      "21:\tlearn: 0.8637680\ttotal: 4.95s\tremaining: 1m 2s\n",
      "22:\tlearn: 0.8648649\ttotal: 5.17s\tremaining: 1m 2s\n",
      "23:\tlearn: 0.8652159\ttotal: 5.38s\tremaining: 1m 1s\n",
      "24:\tlearn: 0.8651720\ttotal: 5.45s\tremaining: 59.9s\n",
      "25:\tlearn: 0.8652159\ttotal: 5.67s\tremaining: 59.8s\n",
      "26:\tlearn: 0.8657862\ttotal: 5.89s\tremaining: 59.5s\n",
      "27:\tlearn: 0.8662689\ttotal: 6.11s\tremaining: 59.4s\n",
      "28:\tlearn: 0.8667076\ttotal: 6.33s\tremaining: 59.2s\n",
      "29:\tlearn: 0.8670147\ttotal: 6.5s\tremaining: 58.5s\n",
      "30:\tlearn: 0.8674535\ttotal: 6.74s\tremaining: 58.4s\n",
      "31:\tlearn: 0.8677167\ttotal: 6.97s\tremaining: 58.4s\n",
      "32:\tlearn: 0.8675851\ttotal: 7.03s\tremaining: 56.9s\n",
      "33:\tlearn: 0.8674974\ttotal: 7.26s\tremaining: 56.8s\n",
      "34:\tlearn: 0.8682871\ttotal: 7.43s\tremaining: 56.3s\n",
      "35:\tlearn: 0.8681994\ttotal: 7.54s\tremaining: 55.3s\n",
      "36:\tlearn: 0.8682432\ttotal: 7.78s\tremaining: 55.3s\n",
      "37:\tlearn: 0.8684626\ttotal: 7.94s\tremaining: 54.8s\n",
      "38:\tlearn: 0.8689014\ttotal: 8.01s\tremaining: 53.6s\n",
      "39:\tlearn: 0.8689014\ttotal: 8.24s\tremaining: 53.6s\n",
      "40:\tlearn: 0.8689014\ttotal: 8.46s\tremaining: 53.5s\n",
      "41:\tlearn: 0.8686381\ttotal: 8.7s\tremaining: 53.5s\n",
      "42:\tlearn: 0.8692524\ttotal: 8.93s\tremaining: 53.4s\n",
      "43:\tlearn: 0.8692524\ttotal: 8.97s\tremaining: 52.2s\n",
      "44:\tlearn: 0.8695595\ttotal: 9.22s\tremaining: 52.2s\n",
      "45:\tlearn: 0.8696034\ttotal: 9.28s\tremaining: 51.2s\n",
      "46:\tlearn: 0.8699982\ttotal: 9.51s\tremaining: 51.2s\n",
      "47:\tlearn: 0.8707002\ttotal: 9.74s\tremaining: 51.2s\n",
      "48:\tlearn: 0.8707441\ttotal: 9.96s\tremaining: 51s\n",
      "49:\tlearn: 0.8707880\ttotal: 10.1s\tremaining: 50.3s\n",
      "50:\tlearn: 0.8709196\ttotal: 10.2s\tremaining: 50s\n",
      "51:\tlearn: 0.8710074\ttotal: 10.5s\tremaining: 50s\n",
      "52:\tlearn: 0.8712706\ttotal: 10.7s\tremaining: 49.7s\n",
      "53:\tlearn: 0.8714022\ttotal: 10.8s\tremaining: 49s\n",
      "54:\tlearn: 0.8715777\ttotal: 10.8s\tremaining: 48.2s\n",
      "55:\tlearn: 0.8715777\ttotal: 10.9s\tremaining: 47.4s\n",
      "56:\tlearn: 0.8716655\ttotal: 10.9s\tremaining: 46.5s\n",
      "57:\tlearn: 0.8718410\ttotal: 11s\tremaining: 46s\n",
      "58:\tlearn: 0.8718849\ttotal: 11.1s\tremaining: 45.2s\n",
      "59:\tlearn: 0.8722359\ttotal: 11.3s\tremaining: 45.4s\n",
      "60:\tlearn: 0.8724552\ttotal: 11.6s\tremaining: 45.5s\n",
      "61:\tlearn: 0.8724552\ttotal: 11.7s\tremaining: 44.7s\n",
      "62:\tlearn: 0.8725430\ttotal: 11.9s\tremaining: 44.9s\n",
      "63:\tlearn: 0.8726746\ttotal: 12.1s\tremaining: 44.7s\n",
      "64:\tlearn: 0.8726746\ttotal: 12.2s\tremaining: 44.1s\n",
      "65:\tlearn: 0.8728062\ttotal: 12.4s\tremaining: 44.1s\n",
      "66:\tlearn: 0.8731134\ttotal: 12.7s\tremaining: 44.1s\n",
      "67:\tlearn: 0.8728062\ttotal: 12.9s\tremaining: 44.1s\n",
      "68:\tlearn: 0.8732889\ttotal: 13.2s\tremaining: 44s\n",
      "69:\tlearn: 0.8732450\ttotal: 13.2s\tremaining: 43.4s\n",
      "70:\tlearn: 0.8734205\ttotal: 13.4s\tremaining: 43.3s\n",
      "71:\tlearn: 0.8739470\ttotal: 13.5s\tremaining: 42.7s\n",
      "72:\tlearn: 0.8737276\ttotal: 13.7s\tremaining: 42.7s\n",
      "73:\tlearn: 0.8742980\ttotal: 14s\tremaining: 42.8s\n",
      "74:\tlearn: 0.8742980\ttotal: 14.2s\tremaining: 42.5s\n",
      "75:\tlearn: 0.8741225\ttotal: 14.4s\tremaining: 42.5s\n",
      "76:\tlearn: 0.8741225\ttotal: 14.7s\tremaining: 42.5s\n",
      "77:\tlearn: 0.8745174\ttotal: 14.8s\tremaining: 42s\n",
      "78:\tlearn: 0.8746929\ttotal: 15s\tremaining: 41.9s\n",
      "79:\tlearn: 0.8748684\ttotal: 15.2s\tremaining: 41.8s\n",
      "80:\tlearn: 0.8753071\ttotal: 15.5s\tremaining: 41.8s\n",
      "81:\tlearn: 0.8753071\ttotal: 15.5s\tremaining: 41.2s\n",
      "82:\tlearn: 0.8753071\ttotal: 15.7s\tremaining: 41s\n",
      "83:\tlearn: 0.8755265\ttotal: 16s\tremaining: 41s\n",
      "84:\tlearn: 0.8754388\ttotal: 16.3s\tremaining: 41.2s\n",
      "85:\tlearn: 0.8752633\ttotal: 16.6s\tremaining: 41.2s\n",
      "86:\tlearn: 0.8752633\ttotal: 16.6s\tremaining: 40.7s\n",
      "87:\tlearn: 0.8753510\ttotal: 16.9s\tremaining: 40.8s\n",
      "88:\tlearn: 0.8753949\ttotal: 17.2s\tremaining: 40.8s\n",
      "89:\tlearn: 0.8753510\ttotal: 17.4s\tremaining: 40.6s\n",
      "90:\tlearn: 0.8755704\ttotal: 17.6s\tremaining: 40.5s\n",
      "91:\tlearn: 0.8757020\ttotal: 17.7s\tremaining: 39.9s\n",
      "92:\tlearn: 0.8754388\ttotal: 17.7s\tremaining: 39.5s\n",
      "93:\tlearn: 0.8754388\ttotal: 17.8s\tremaining: 38.9s\n",
      "94:\tlearn: 0.8753949\ttotal: 18s\tremaining: 38.8s\n",
      "95:\tlearn: 0.8755704\ttotal: 18.2s\tremaining: 38.7s\n",
      "96:\tlearn: 0.8760530\ttotal: 18.5s\tremaining: 38.6s\n",
      "97:\tlearn: 0.8759214\ttotal: 18.6s\tremaining: 38.2s\n",
      "98:\tlearn: 0.8756581\ttotal: 18.8s\tremaining: 38.2s\n",
      "99:\tlearn: 0.8756581\ttotal: 18.8s\tremaining: 37.7s\n",
      "100:\tlearn: 0.8759653\ttotal: 18.9s\tremaining: 37.2s\n",
      "101:\tlearn: 0.8760530\ttotal: 19s\tremaining: 36.8s\n",
      "102:\tlearn: 0.8758775\ttotal: 19.2s\tremaining: 36.7s\n",
      "103:\tlearn: 0.8760969\ttotal: 19.3s\tremaining: 36.3s\n",
      "104:\tlearn: 0.8761408\ttotal: 19.4s\tremaining: 36.1s\n",
      "105:\tlearn: 0.8761408\ttotal: 19.7s\tremaining: 36.1s\n",
      "106:\tlearn: 0.8761408\ttotal: 20s\tremaining: 36s\n",
      "107:\tlearn: 0.8760969\ttotal: 20.1s\tremaining: 35.8s\n",
      "108:\tlearn: 0.8763601\ttotal: 20.5s\tremaining: 35.8s\n",
      "109:\tlearn: 0.8766673\ttotal: 20.8s\tremaining: 35.8s\n",
      "110:\tlearn: 0.8768866\ttotal: 21.1s\tremaining: 35.8s\n",
      "111:\tlearn: 0.8769305\ttotal: 21.1s\tremaining: 35.5s\n",
      "112:\tlearn: 0.8770621\ttotal: 21.4s\tremaining: 35.5s\n",
      "113:\tlearn: 0.8771938\ttotal: 21.7s\tremaining: 35.3s\n",
      "114:\tlearn: 0.8772376\ttotal: 21.9s\tremaining: 35.2s\n",
      "115:\tlearn: 0.8772815\ttotal: 22.1s\tremaining: 35.1s\n",
      "116:\tlearn: 0.8772376\ttotal: 22.4s\tremaining: 35s\n",
      "117:\tlearn: 0.8773693\ttotal: 22.6s\tremaining: 34.9s\n",
      "118:\tlearn: 0.8772376\ttotal: 22.9s\tremaining: 34.8s\n",
      "119:\tlearn: 0.8774570\ttotal: 23.1s\tremaining: 34.7s\n",
      "120:\tlearn: 0.8774570\ttotal: 23.1s\tremaining: 34.2s\n",
      "121:\tlearn: 0.8776325\ttotal: 23.4s\tremaining: 34.1s\n",
      "122:\tlearn: 0.8778958\ttotal: 23.6s\tremaining: 34s\n",
      "123:\tlearn: 0.8778958\ttotal: 23.9s\tremaining: 33.9s\n",
      "124:\tlearn: 0.8779396\ttotal: 24.1s\tremaining: 33.7s\n",
      "125:\tlearn: 0.8780274\ttotal: 24.3s\tremaining: 33.6s\n",
      "126:\tlearn: 0.8778958\ttotal: 24.6s\tremaining: 33.5s\n",
      "127:\tlearn: 0.8778958\ttotal: 24.9s\tremaining: 33.4s\n",
      "128:\tlearn: 0.8780713\ttotal: 25.1s\tremaining: 33.3s\n",
      "129:\tlearn: 0.8779835\ttotal: 25.3s\tremaining: 33.1s\n",
      "130:\tlearn: 0.8781151\ttotal: 25.6s\tremaining: 33s\n",
      "131:\tlearn: 0.8781590\ttotal: 25.8s\tremaining: 32.9s\n",
      "132:\tlearn: 0.8781590\ttotal: 26.1s\tremaining: 32.7s\n",
      "133:\tlearn: 0.8782029\ttotal: 26.3s\tremaining: 32.6s\n",
      "134:\tlearn: 0.8783345\ttotal: 26.5s\tremaining: 32.4s\n",
      "135:\tlearn: 0.8782468\ttotal: 26.7s\tremaining: 32.2s\n",
      "136:\tlearn: 0.8782468\ttotal: 27s\tremaining: 32.1s\n",
      "137:\tlearn: 0.8784223\ttotal: 27.2s\tremaining: 32s\n",
      "138:\tlearn: 0.8783784\ttotal: 27.5s\tremaining: 31.8s\n",
      "139:\tlearn: 0.8784223\ttotal: 27.7s\tremaining: 31.7s\n",
      "140:\tlearn: 0.8786416\ttotal: 27.9s\tremaining: 31.5s\n",
      "141:\tlearn: 0.8785978\ttotal: 28.2s\tremaining: 31.4s\n",
      "142:\tlearn: 0.8785539\ttotal: 28.4s\tremaining: 31.2s\n",
      "143:\tlearn: 0.8785539\ttotal: 28.6s\tremaining: 31s\n",
      "144:\tlearn: 0.8784223\ttotal: 28.9s\tremaining: 30.9s\n",
      "145:\tlearn: 0.8785100\ttotal: 29s\tremaining: 30.5s\n",
      "146:\tlearn: 0.8784661\ttotal: 29.2s\tremaining: 30.4s\n",
      "147:\tlearn: 0.8785539\ttotal: 29.4s\tremaining: 30.2s\n",
      "148:\tlearn: 0.8791243\ttotal: 29.7s\tremaining: 30.1s\n",
      "149:\tlearn: 0.8792998\ttotal: 29.9s\tremaining: 29.9s\n",
      "150:\tlearn: 0.8792998\ttotal: 30.2s\tremaining: 29.8s\n",
      "151:\tlearn: 0.8792998\ttotal: 30.4s\tremaining: 29.6s\n",
      "152:\tlearn: 0.8792998\ttotal: 30.6s\tremaining: 29.4s\n",
      "153:\tlearn: 0.8797824\ttotal: 30.9s\tremaining: 29.3s\n",
      "154:\tlearn: 0.8798701\ttotal: 31.1s\tremaining: 29.1s\n",
      "155:\tlearn: 0.8800018\ttotal: 31.3s\tremaining: 28.9s\n",
      "156:\tlearn: 0.8798701\ttotal: 31.6s\tremaining: 28.8s\n",
      "157:\tlearn: 0.8799140\ttotal: 31.8s\tremaining: 28.6s\n",
      "158:\tlearn: 0.8799140\ttotal: 32s\tremaining: 28.4s\n",
      "159:\tlearn: 0.8799140\ttotal: 32.2s\tremaining: 28.2s\n",
      "160:\tlearn: 0.8800018\ttotal: 32.5s\tremaining: 28.1s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "161:\tlearn: 0.8801334\ttotal: 32.7s\tremaining: 27.9s\n",
      "162:\tlearn: 0.8801334\ttotal: 33s\tremaining: 27.7s\n",
      "163:\tlearn: 0.8801334\ttotal: 33.2s\tremaining: 27.5s\n",
      "164:\tlearn: 0.8800456\ttotal: 33.4s\tremaining: 27.3s\n",
      "165:\tlearn: 0.8801773\ttotal: 33.7s\tremaining: 27.2s\n",
      "166:\tlearn: 0.8802211\ttotal: 33.9s\tremaining: 27s\n",
      "167:\tlearn: 0.8800456\ttotal: 34.1s\tremaining: 26.8s\n",
      "168:\tlearn: 0.8800018\ttotal: 34.4s\tremaining: 26.6s\n",
      "169:\tlearn: 0.8800018\ttotal: 34.6s\tremaining: 26.5s\n",
      "170:\tlearn: 0.8803089\ttotal: 34.8s\tremaining: 26.3s\n",
      "171:\tlearn: 0.8803089\ttotal: 34.9s\tremaining: 26s\n",
      "172:\tlearn: 0.8803089\ttotal: 35.1s\tremaining: 25.8s\n",
      "173:\tlearn: 0.8803089\ttotal: 35.2s\tremaining: 25.5s\n",
      "174:\tlearn: 0.8802650\ttotal: 35.4s\tremaining: 25.3s\n",
      "175:\tlearn: 0.8803089\ttotal: 35.6s\tremaining: 25.1s\n",
      "176:\tlearn: 0.8803089\ttotal: 35.9s\tremaining: 24.9s\n",
      "177:\tlearn: 0.8802211\ttotal: 36.1s\tremaining: 24.7s\n",
      "178:\tlearn: 0.8801773\ttotal: 36.3s\tremaining: 24.6s\n",
      "179:\tlearn: 0.8801773\ttotal: 36.4s\tremaining: 24.3s\n",
      "180:\tlearn: 0.8802211\ttotal: 36.6s\tremaining: 24.1s\n",
      "181:\tlearn: 0.8803089\ttotal: 36.9s\tremaining: 23.9s\n",
      "182:\tlearn: 0.8804405\ttotal: 37.1s\tremaining: 23.7s\n",
      "183:\tlearn: 0.8804844\ttotal: 37.3s\tremaining: 23.5s\n",
      "184:\tlearn: 0.8807476\ttotal: 37.5s\tremaining: 23.3s\n",
      "185:\tlearn: 0.8807476\ttotal: 37.7s\tremaining: 23.1s\n",
      "186:\tlearn: 0.8807476\ttotal: 37.8s\tremaining: 22.8s\n",
      "187:\tlearn: 0.8808354\ttotal: 38s\tremaining: 22.6s\n",
      "188:\tlearn: 0.8808354\ttotal: 38.2s\tremaining: 22.4s\n",
      "189:\tlearn: 0.8810986\ttotal: 38.4s\tremaining: 22.3s\n",
      "190:\tlearn: 0.8811425\ttotal: 38.7s\tremaining: 22.1s\n",
      "191:\tlearn: 0.8810109\ttotal: 38.9s\tremaining: 21.9s\n",
      "192:\tlearn: 0.8810109\ttotal: 39.1s\tremaining: 21.7s\n",
      "193:\tlearn: 0.8810986\ttotal: 39.3s\tremaining: 21.5s\n",
      "194:\tlearn: 0.8811425\ttotal: 39.6s\tremaining: 21.3s\n",
      "195:\tlearn: 0.8815813\ttotal: 39.8s\tremaining: 21.1s\n",
      "196:\tlearn: 0.8815813\ttotal: 40s\tremaining: 20.9s\n",
      "197:\tlearn: 0.8815813\ttotal: 40.2s\tremaining: 20.7s\n",
      "198:\tlearn: 0.8817129\ttotal: 40.4s\tremaining: 20.5s\n",
      "199:\tlearn: 0.8821516\ttotal: 40.7s\tremaining: 20.3s\n",
      "200:\tlearn: 0.8821516\ttotal: 40.9s\tremaining: 20.2s\n",
      "201:\tlearn: 0.8823710\ttotal: 41.1s\tremaining: 20s\n",
      "202:\tlearn: 0.8823271\ttotal: 41.4s\tremaining: 19.8s\n",
      "203:\tlearn: 0.8821955\ttotal: 41.6s\tremaining: 19.6s\n",
      "204:\tlearn: 0.8822833\ttotal: 41.9s\tremaining: 19.4s\n",
      "205:\tlearn: 0.8822833\ttotal: 42.1s\tremaining: 19.2s\n",
      "206:\tlearn: 0.8825026\ttotal: 42.3s\tremaining: 19s\n",
      "207:\tlearn: 0.8826781\ttotal: 42.5s\tremaining: 18.8s\n",
      "208:\tlearn: 0.8825026\ttotal: 42.8s\tremaining: 18.6s\n",
      "209:\tlearn: 0.8824588\ttotal: 43s\tremaining: 18.4s\n",
      "210:\tlearn: 0.8825904\ttotal: 43.2s\tremaining: 18.2s\n",
      "211:\tlearn: 0.8827220\ttotal: 43.4s\tremaining: 18s\n",
      "212:\tlearn: 0.8828536\ttotal: 43.7s\tremaining: 17.8s\n",
      "213:\tlearn: 0.8828975\ttotal: 43.9s\tremaining: 17.6s\n",
      "214:\tlearn: 0.8830730\ttotal: 44.1s\tremaining: 17.4s\n",
      "215:\tlearn: 0.8829414\ttotal: 44.3s\tremaining: 17.2s\n",
      "216:\tlearn: 0.8828536\ttotal: 44.6s\tremaining: 17s\n",
      "217:\tlearn: 0.8828536\ttotal: 44.8s\tremaining: 16.9s\n",
      "218:\tlearn: 0.8828975\ttotal: 45s\tremaining: 16.7s\n",
      "219:\tlearn: 0.8828098\ttotal: 45.3s\tremaining: 16.5s\n",
      "220:\tlearn: 0.8828098\ttotal: 45.5s\tremaining: 16.3s\n",
      "221:\tlearn: 0.8828975\ttotal: 45.7s\tremaining: 16.1s\n",
      "222:\tlearn: 0.8828975\ttotal: 46s\tremaining: 15.9s\n",
      "223:\tlearn: 0.8830730\ttotal: 46.2s\tremaining: 15.7s\n",
      "224:\tlearn: 0.8830730\ttotal: 46.4s\tremaining: 15.5s\n",
      "225:\tlearn: 0.8830291\ttotal: 46.6s\tremaining: 15.3s\n",
      "226:\tlearn: 0.8830730\ttotal: 46.9s\tremaining: 15.1s\n",
      "227:\tlearn: 0.8831169\ttotal: 47.1s\tremaining: 14.9s\n",
      "228:\tlearn: 0.8831608\ttotal: 47.4s\tremaining: 14.7s\n",
      "229:\tlearn: 0.8831169\ttotal: 47.6s\tremaining: 14.5s\n",
      "230:\tlearn: 0.8831169\ttotal: 47.9s\tremaining: 14.3s\n",
      "231:\tlearn: 0.8830730\ttotal: 47.9s\tremaining: 14.1s\n",
      "232:\tlearn: 0.8830730\ttotal: 48.2s\tremaining: 13.9s\n",
      "233:\tlearn: 0.8830730\ttotal: 48.4s\tremaining: 13.6s\n",
      "234:\tlearn: 0.8831169\ttotal: 48.6s\tremaining: 13.4s\n",
      "235:\tlearn: 0.8831169\ttotal: 48.7s\tremaining: 13.2s\n",
      "236:\tlearn: 0.8830730\ttotal: 48.9s\tremaining: 13s\n",
      "237:\tlearn: 0.8831169\ttotal: 49.1s\tremaining: 12.8s\n",
      "238:\tlearn: 0.8832046\ttotal: 49.4s\tremaining: 12.6s\n",
      "239:\tlearn: 0.8831608\ttotal: 49.6s\tremaining: 12.4s\n",
      "240:\tlearn: 0.8832046\ttotal: 49.7s\tremaining: 12.2s\n",
      "241:\tlearn: 0.8831608\ttotal: 50s\tremaining: 12s\n",
      "242:\tlearn: 0.8833363\ttotal: 50.2s\tremaining: 11.8s\n",
      "243:\tlearn: 0.8833363\ttotal: 50.4s\tremaining: 11.6s\n",
      "244:\tlearn: 0.8833363\ttotal: 50.7s\tremaining: 11.4s\n",
      "245:\tlearn: 0.8834240\ttotal: 50.9s\tremaining: 11.2s\n",
      "246:\tlearn: 0.8834240\ttotal: 51.1s\tremaining: 11s\n",
      "247:\tlearn: 0.8833801\ttotal: 51.3s\tremaining: 10.8s\n",
      "248:\tlearn: 0.8833801\ttotal: 51.5s\tremaining: 10.6s\n",
      "249:\tlearn: 0.8833801\ttotal: 51.7s\tremaining: 10.3s\n",
      "250:\tlearn: 0.8833801\ttotal: 52s\tremaining: 10.1s\n",
      "251:\tlearn: 0.8833801\ttotal: 52.2s\tremaining: 9.94s\n",
      "252:\tlearn: 0.8833801\ttotal: 52.4s\tremaining: 9.74s\n",
      "253:\tlearn: 0.8833801\ttotal: 52.7s\tremaining: 9.54s\n",
      "254:\tlearn: 0.8833801\ttotal: 52.8s\tremaining: 9.32s\n",
      "255:\tlearn: 0.8833801\ttotal: 53s\tremaining: 9.12s\n",
      "256:\tlearn: 0.8834240\ttotal: 53.3s\tremaining: 8.91s\n",
      "257:\tlearn: 0.8833363\ttotal: 53.5s\tremaining: 8.71s\n",
      "258:\tlearn: 0.8833363\ttotal: 53.7s\tremaining: 8.51s\n",
      "259:\tlearn: 0.8834240\ttotal: 54s\tremaining: 8.3s\n",
      "260:\tlearn: 0.8833363\ttotal: 54.2s\tremaining: 8.1s\n",
      "261:\tlearn: 0.8833801\ttotal: 54.4s\tremaining: 7.89s\n",
      "262:\tlearn: 0.8833801\ttotal: 54.7s\tremaining: 7.69s\n",
      "263:\tlearn: 0.8834240\ttotal: 54.9s\tremaining: 7.49s\n",
      "264:\tlearn: 0.8833801\ttotal: 55.1s\tremaining: 7.28s\n",
      "265:\tlearn: 0.8834240\ttotal: 55.2s\tremaining: 7.06s\n",
      "266:\tlearn: 0.8834240\ttotal: 55.4s\tremaining: 6.85s\n",
      "267:\tlearn: 0.8834240\ttotal: 55.7s\tremaining: 6.64s\n",
      "268:\tlearn: 0.8832924\ttotal: 55.9s\tremaining: 6.44s\n",
      "269:\tlearn: 0.8832485\ttotal: 56.1s\tremaining: 6.23s\n",
      "270:\tlearn: 0.8832485\ttotal: 56.3s\tremaining: 6.03s\n",
      "271:\tlearn: 0.8832485\ttotal: 56.4s\tremaining: 5.8s\n",
      "272:\tlearn: 0.8832924\ttotal: 56.6s\tremaining: 5.6s\n",
      "273:\tlearn: 0.8832924\ttotal: 56.8s\tremaining: 5.39s\n",
      "274:\tlearn: 0.8833363\ttotal: 56.9s\tremaining: 5.17s\n",
      "275:\tlearn: 0.8833363\ttotal: 57.1s\tremaining: 4.96s\n",
      "276:\tlearn: 0.8833801\ttotal: 57.3s\tremaining: 4.76s\n",
      "277:\tlearn: 0.8833363\ttotal: 57.5s\tremaining: 4.55s\n",
      "278:\tlearn: 0.8833363\ttotal: 57.6s\tremaining: 4.33s\n",
      "279:\tlearn: 0.8833363\ttotal: 57.8s\tremaining: 4.13s\n",
      "280:\tlearn: 0.8833363\ttotal: 57.8s\tremaining: 3.91s\n",
      "281:\tlearn: 0.8833363\ttotal: 58s\tremaining: 3.7s\n",
      "282:\tlearn: 0.8834240\ttotal: 58.3s\tremaining: 3.5s\n",
      "283:\tlearn: 0.8834240\ttotal: 58.5s\tremaining: 3.29s\n",
      "284:\tlearn: 0.8834240\ttotal: 58.5s\tremaining: 3.08s\n",
      "285:\tlearn: 0.8835556\ttotal: 58.8s\tremaining: 2.88s\n",
      "286:\tlearn: 0.8835995\ttotal: 59s\tremaining: 2.67s\n",
      "287:\tlearn: 0.8835995\ttotal: 59.2s\tremaining: 2.47s\n",
      "288:\tlearn: 0.8835995\ttotal: 59.3s\tremaining: 2.26s\n",
      "289:\tlearn: 0.8835556\ttotal: 59.5s\tremaining: 2.05s\n",
      "290:\tlearn: 0.8835556\ttotal: 59.8s\tremaining: 1.85s\n",
      "291:\tlearn: 0.8835556\ttotal: 59.8s\tremaining: 1.64s\n",
      "292:\tlearn: 0.8835556\ttotal: 59.8s\tremaining: 1.43s\n",
      "293:\tlearn: 0.8835556\ttotal: 1m\tremaining: 1.23s\n",
      "294:\tlearn: 0.8835556\ttotal: 1m\tremaining: 1.02s\n",
      "295:\tlearn: 0.8836434\ttotal: 1m\tremaining: 818ms\n",
      "296:\tlearn: 0.8835995\ttotal: 1m\tremaining: 614ms\n",
      "297:\tlearn: 0.8835995\ttotal: 1m\tremaining: 409ms\n",
      "298:\tlearn: 0.8835995\ttotal: 1m 1s\tremaining: 205ms\n",
      "299:\tlearn: 0.8835556\ttotal: 1m 1s\tremaining: 0us\n",
      "mean test score (weighted by split size) of CV rounds:  [0.87056862 0.87030537 0.87034925 0.87091962]\n",
      "Best parameter set {'depth': 10, 'iterations': 300, 'learning_rate': 0.1, 'reg_lambda': 3, 'rsm': 0.9} Corresponding mean CV score 0.8709196209196209\n"
     ]
    }
   ],
   "source": [
    "### Catboost grid search\n",
    "\n",
    "catbCV = CatBoostClassifier(verbose=1,random_seed=10,eval_metric='Accuracy')\n",
    "param_grid = {}\n",
    "param_grid['learning_rate'] = [0.1]#, 0.3]\n",
    "param_grid['rsm'] = [0.9]\n",
    "#param_grid['subsample'] = [0.9]\n",
    "param_grid['iterations'] = [300]\n",
    "param_grid['reg_lambda']= [2,3]\n",
    "param_grid['depth'] = [8,10]#5\n",
    "catbCV_grid = grid_search(catbCV, param_grid,X_train,y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 168,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0:\tlearn: 0.8489382\ttest: 0.8489382\ttest1: 0.8447129\tbest: 0.8447129 (0)\ttotal: 226ms\tremaining: 1m 7s\n",
      "1:\tlearn: 0.8508687\ttest: 0.8508687\ttest1: 0.8468625\tbest: 0.8468625 (1)\ttotal: 453ms\tremaining: 1m 7s\n",
      "2:\tlearn: 0.8531064\ttest: 0.8531064\ttest1: 0.8488075\tbest: 0.8488075 (2)\ttotal: 704ms\tremaining: 1m 9s\n",
      "3:\tlearn: 0.8520534\ttest: 0.8520534\ttest1: 0.8472720\tbest: 0.8488075 (2)\ttotal: 940ms\tremaining: 1m 9s\n",
      "4:\tlearn: 0.8542471\ttest: 0.8542471\ttest1: 0.8503429\tbest: 0.8503429 (4)\ttotal: 1.16s\tremaining: 1m 8s\n",
      "5:\tlearn: 0.8554756\ttest: 0.8554756\ttest1: 0.8529020\tbest: 0.8529020 (5)\ttotal: 1.27s\tremaining: 1m 2s\n",
      "6:\tlearn: 0.8550369\ttest: 0.8550369\ttest1: 0.8523902\tbest: 0.8529020 (5)\ttotal: 1.52s\tremaining: 1m 3s\n",
      "7:\tlearn: 0.8545981\ttest: 0.8545981\ttest1: 0.8523902\tbest: 0.8529020 (5)\ttotal: 1.61s\tremaining: 59s\n",
      "8:\tlearn: 0.8567041\ttest: 0.8567041\ttest1: 0.8529020\tbest: 0.8529020 (5)\ttotal: 1.85s\tremaining: 59.9s\n",
      "9:\tlearn: 0.8580204\ttest: 0.8580204\ttest1: 0.8548470\tbest: 0.8548470 (9)\ttotal: 2.13s\tremaining: 1m 1s\n",
      "10:\tlearn: 0.8585030\ttest: 0.8585030\ttest1: 0.8551541\tbest: 0.8551541 (10)\ttotal: 2.37s\tremaining: 1m 2s\n",
      "11:\tlearn: 0.8596437\ttest: 0.8596437\ttest1: 0.8560753\tbest: 0.8560753 (11)\ttotal: 2.59s\tremaining: 1m 2s\n",
      "12:\tlearn: 0.8604335\ttest: 0.8604335\ttest1: 0.8552564\tbest: 0.8560753 (11)\ttotal: 2.85s\tremaining: 1m 2s\n",
      "13:\tlearn: 0.8607845\ttest: 0.8607845\ttest1: 0.8556659\tbest: 0.8560753 (11)\ttotal: 3.09s\tremaining: 1m 3s\n",
      "14:\tlearn: 0.8604774\ttest: 0.8604774\ttest1: 0.8557682\tbest: 0.8560753 (11)\ttotal: 3.31s\tremaining: 1m 2s\n",
      "15:\tlearn: 0.8617497\ttest: 0.8617497\ttest1: 0.8560753\tbest: 0.8560753 (11)\ttotal: 3.55s\tremaining: 1m 2s\n",
      "16:\tlearn: 0.8622762\ttest: 0.8622762\ttest1: 0.8568943\tbest: 0.8568943 (16)\ttotal: 3.65s\tremaining: 1m\n",
      "17:\tlearn: 0.8623640\ttest: 0.8623640\ttest1: 0.8567919\tbest: 0.8568943 (16)\ttotal: 3.88s\tremaining: 1m\n",
      "18:\tlearn: 0.8632415\ttest: 0.8632415\ttest1: 0.8573037\tbest: 0.8573037 (18)\ttotal: 4.1s\tremaining: 1m\n",
      "19:\tlearn: 0.8636802\ttest: 0.8636802\ttest1: 0.8579179\tbest: 0.8579179 (19)\ttotal: 4.33s\tremaining: 1m\n",
      "20:\tlearn: 0.8637241\ttest: 0.8637241\ttest1: 0.8583274\tbest: 0.8583274 (20)\ttotal: 4.56s\tremaining: 1m\n",
      "21:\tlearn: 0.8637680\ttest: 0.8637680\ttest1: 0.8586345\tbest: 0.8586345 (21)\ttotal: 4.79s\tremaining: 1m\n",
      "22:\tlearn: 0.8648649\ttest: 0.8648649\ttest1: 0.8591463\tbest: 0.8591463 (22)\ttotal: 5.01s\tremaining: 1m\n",
      "23:\tlearn: 0.8652159\ttest: 0.8652159\ttest1: 0.8586345\tbest: 0.8591463 (22)\ttotal: 5.24s\tremaining: 1m\n",
      "24:\tlearn: 0.8651720\ttest: 0.8651720\ttest1: 0.8588392\tbest: 0.8591463 (22)\ttotal: 5.3s\tremaining: 58.3s\n",
      "25:\tlearn: 0.8652159\ttest: 0.8652159\ttest1: 0.8592486\tbest: 0.8592486 (25)\ttotal: 5.52s\tremaining: 58.2s\n",
      "26:\tlearn: 0.8657862\ttest: 0.8657862\ttest1: 0.8598628\tbest: 0.8598628 (26)\ttotal: 5.76s\tremaining: 58.2s\n",
      "27:\tlearn: 0.8662689\ttest: 0.8662689\ttest1: 0.8603747\tbest: 0.8603747 (27)\ttotal: 5.96s\tremaining: 58s\n",
      "28:\tlearn: 0.8667076\ttest: 0.8667076\ttest1: 0.8607841\tbest: 0.8607841 (28)\ttotal: 6.19s\tremaining: 57.8s\n",
      "29:\tlearn: 0.8670147\ttest: 0.8670147\ttest1: 0.8610912\tbest: 0.8610912 (29)\ttotal: 6.37s\tremaining: 57.3s\n",
      "30:\tlearn: 0.8674535\ttest: 0.8674535\ttest1: 0.8613983\tbest: 0.8613983 (30)\ttotal: 6.59s\tremaining: 57.2s\n",
      "31:\tlearn: 0.8677167\ttest: 0.8677167\ttest1: 0.8617054\tbest: 0.8617054 (31)\ttotal: 6.81s\tremaining: 57.1s\n",
      "32:\tlearn: 0.8675851\ttest: 0.8675851\ttest1: 0.8618078\tbest: 0.8618078 (32)\ttotal: 6.88s\tremaining: 55.6s\n",
      "33:\tlearn: 0.8674974\ttest: 0.8674974\ttest1: 0.8616030\tbest: 0.8618078 (32)\ttotal: 7.11s\tremaining: 55.6s\n",
      "34:\tlearn: 0.8682871\ttest: 0.8682871\ttest1: 0.8618078\tbest: 0.8618078 (32)\ttotal: 7.28s\tremaining: 55.1s\n",
      "35:\tlearn: 0.8681994\ttest: 0.8681994\ttest1: 0.8618078\tbest: 0.8618078 (32)\ttotal: 7.4s\tremaining: 54.2s\n",
      "36:\tlearn: 0.8682432\ttest: 0.8682432\ttest1: 0.8625243\tbest: 0.8625243 (36)\ttotal: 7.62s\tremaining: 54.2s\n",
      "37:\tlearn: 0.8684626\ttest: 0.8684626\ttest1: 0.8630361\tbest: 0.8630361 (37)\ttotal: 7.8s\tremaining: 53.8s\n",
      "38:\tlearn: 0.8689014\ttest: 0.8689014\ttest1: 0.8627290\tbest: 0.8630361 (37)\ttotal: 7.87s\tremaining: 52.7s\n",
      "39:\tlearn: 0.8689014\ttest: 0.8689014\ttest1: 0.8628314\tbest: 0.8630361 (37)\ttotal: 8.1s\tremaining: 52.6s\n",
      "40:\tlearn: 0.8689014\ttest: 0.8689014\ttest1: 0.8627290\tbest: 0.8630361 (37)\ttotal: 8.34s\tremaining: 52.7s\n",
      "41:\tlearn: 0.8686381\ttest: 0.8686381\ttest1: 0.8624219\tbest: 0.8630361 (37)\ttotal: 8.57s\tremaining: 52.6s\n",
      "42:\tlearn: 0.8692524\ttest: 0.8692524\ttest1: 0.8628314\tbest: 0.8630361 (37)\ttotal: 8.8s\tremaining: 52.6s\n",
      "43:\tlearn: 0.8692524\ttest: 0.8692524\ttest1: 0.8628314\tbest: 0.8630361 (37)\ttotal: 8.84s\tremaining: 51.4s\n",
      "44:\tlearn: 0.8695595\ttest: 0.8695595\ttest1: 0.8633432\tbest: 0.8633432 (44)\ttotal: 9.08s\tremaining: 51.4s\n",
      "45:\tlearn: 0.8696034\ttest: 0.8696034\ttest1: 0.8635480\tbest: 0.8635480 (45)\ttotal: 9.14s\tremaining: 50.5s\n",
      "46:\tlearn: 0.8699982\ttest: 0.8699982\ttest1: 0.8636503\tbest: 0.8636503 (46)\ttotal: 9.38s\tremaining: 50.5s\n",
      "47:\tlearn: 0.8707002\ttest: 0.8707002\ttest1: 0.8639574\tbest: 0.8639574 (47)\ttotal: 9.61s\tremaining: 50.5s\n",
      "48:\tlearn: 0.8707441\ttest: 0.8707441\ttest1: 0.8639574\tbest: 0.8639574 (47)\ttotal: 9.83s\tremaining: 50.3s\n",
      "49:\tlearn: 0.8707880\ttest: 0.8707880\ttest1: 0.8642645\tbest: 0.8642645 (49)\ttotal: 9.93s\tremaining: 49.7s\n",
      "50:\tlearn: 0.8709196\ttest: 0.8709196\ttest1: 0.8646740\tbest: 0.8646740 (50)\ttotal: 10.1s\tremaining: 49.4s\n",
      "51:\tlearn: 0.8710074\ttest: 0.8710074\ttest1: 0.8649811\tbest: 0.8649811 (51)\ttotal: 10.3s\tremaining: 49.3s\n",
      "52:\tlearn: 0.8712706\ttest: 0.8712706\ttest1: 0.8650834\tbest: 0.8650834 (52)\ttotal: 10.5s\tremaining: 49s\n",
      "53:\tlearn: 0.8714022\ttest: 0.8714022\ttest1: 0.8655953\tbest: 0.8655953 (53)\ttotal: 10.6s\tremaining: 48.4s\n",
      "54:\tlearn: 0.8715777\ttest: 0.8715777\ttest1: 0.8655953\tbest: 0.8655953 (53)\ttotal: 10.7s\tremaining: 47.6s\n",
      "55:\tlearn: 0.8715777\ttest: 0.8715777\ttest1: 0.8655953\tbest: 0.8655953 (53)\ttotal: 10.7s\tremaining: 46.7s\n",
      "56:\tlearn: 0.8716655\ttest: 0.8716655\ttest1: 0.8656976\tbest: 0.8656976 (56)\ttotal: 10.8s\tremaining: 45.8s\n",
      "57:\tlearn: 0.8718410\ttest: 0.8718410\ttest1: 0.8653905\tbest: 0.8656976 (56)\ttotal: 10.8s\tremaining: 45.3s\n",
      "58:\tlearn: 0.8718849\ttest: 0.8718849\ttest1: 0.8653905\tbest: 0.8656976 (56)\ttotal: 10.9s\tremaining: 44.5s\n",
      "59:\tlearn: 0.8722359\ttest: 0.8722359\ttest1: 0.8664142\tbest: 0.8664142 (59)\ttotal: 11.1s\tremaining: 44.4s\n",
      "60:\tlearn: 0.8724552\ttest: 0.8724552\ttest1: 0.8665165\tbest: 0.8665165 (60)\ttotal: 11.3s\tremaining: 44.4s\n",
      "61:\tlearn: 0.8724552\ttest: 0.8724552\ttest1: 0.8664142\tbest: 0.8665165 (60)\ttotal: 11.4s\tremaining: 43.7s\n",
      "62:\tlearn: 0.8725430\ttest: 0.8725430\ttest1: 0.8669260\tbest: 0.8669260 (62)\ttotal: 11.6s\tremaining: 43.7s\n",
      "63:\tlearn: 0.8726746\ttest: 0.8726746\ttest1: 0.8667213\tbest: 0.8669260 (62)\ttotal: 11.8s\tremaining: 43.5s\n",
      "64:\tlearn: 0.8726746\ttest: 0.8726746\ttest1: 0.8667213\tbest: 0.8669260 (62)\ttotal: 11.9s\tremaining: 42.9s\n",
      "65:\tlearn: 0.8728062\ttest: 0.8728062\ttest1: 0.8665165\tbest: 0.8669260 (62)\ttotal: 12.1s\tremaining: 43s\n",
      "66:\tlearn: 0.8731134\ttest: 0.8731134\ttest1: 0.8665165\tbest: 0.8669260 (62)\ttotal: 12.4s\tremaining: 43s\n",
      "67:\tlearn: 0.8728062\ttest: 0.8728062\ttest1: 0.8666189\tbest: 0.8669260 (62)\ttotal: 12.6s\tremaining: 43s\n",
      "68:\tlearn: 0.8732889\ttest: 0.8732889\ttest1: 0.8671307\tbest: 0.8671307 (68)\ttotal: 12.8s\tremaining: 43s\n",
      "69:\tlearn: 0.8732450\ttest: 0.8732450\ttest1: 0.8671307\tbest: 0.8671307 (68)\ttotal: 12.9s\tremaining: 42.4s\n",
      "70:\tlearn: 0.8734205\ttest: 0.8734205\ttest1: 0.8672331\tbest: 0.8672331 (70)\ttotal: 13.1s\tremaining: 42.2s\n",
      "71:\tlearn: 0.8739470\ttest: 0.8739470\ttest1: 0.8671307\tbest: 0.8672331 (70)\ttotal: 13.2s\tremaining: 41.6s\n",
      "72:\tlearn: 0.8737276\ttest: 0.8737276\ttest1: 0.8673354\tbest: 0.8673354 (72)\ttotal: 13.4s\tremaining: 41.7s\n",
      "73:\tlearn: 0.8742980\ttest: 0.8742980\ttest1: 0.8675402\tbest: 0.8675402 (73)\ttotal: 13.6s\tremaining: 41.6s\n",
      "74:\tlearn: 0.8742980\ttest: 0.8742980\ttest1: 0.8676425\tbest: 0.8676425 (74)\ttotal: 13.8s\tremaining: 41.4s\n",
      "75:\tlearn: 0.8741225\ttest: 0.8741225\ttest1: 0.8674378\tbest: 0.8676425 (74)\ttotal: 14s\tremaining: 41.3s\n",
      "76:\tlearn: 0.8741225\ttest: 0.8741225\ttest1: 0.8677449\tbest: 0.8677449 (76)\ttotal: 14.3s\tremaining: 41.3s\n",
      "77:\tlearn: 0.8745174\ttest: 0.8745174\ttest1: 0.8683591\tbest: 0.8683591 (77)\ttotal: 14.4s\tremaining: 40.9s\n",
      "78:\tlearn: 0.8746929\ttest: 0.8746929\ttest1: 0.8680520\tbest: 0.8683591 (77)\ttotal: 14.5s\tremaining: 40.7s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "79:\tlearn: 0.8748684\ttest: 0.8748684\ttest1: 0.8679496\tbest: 0.8683591 (77)\ttotal: 14.8s\tremaining: 40.6s\n",
      "80:\tlearn: 0.8753071\ttest: 0.8753071\ttest1: 0.8685638\tbest: 0.8685638 (80)\ttotal: 15s\tremaining: 40.6s\n",
      "81:\tlearn: 0.8753071\ttest: 0.8753071\ttest1: 0.8688709\tbest: 0.8688709 (81)\ttotal: 15.1s\tremaining: 40s\n",
      "82:\tlearn: 0.8753071\ttest: 0.8753071\ttest1: 0.8687686\tbest: 0.8688709 (81)\ttotal: 15.2s\tremaining: 39.8s\n",
      "83:\tlearn: 0.8755265\ttest: 0.8755265\ttest1: 0.8684615\tbest: 0.8688709 (81)\ttotal: 15.5s\tremaining: 39.8s\n",
      "84:\tlearn: 0.8754388\ttest: 0.8754388\ttest1: 0.8686662\tbest: 0.8688709 (81)\ttotal: 15.7s\tremaining: 39.7s\n",
      "85:\tlearn: 0.8752633\ttest: 0.8752633\ttest1: 0.8686662\tbest: 0.8688709 (81)\ttotal: 15.9s\tremaining: 39.6s\n",
      "86:\tlearn: 0.8752633\ttest: 0.8752633\ttest1: 0.8686662\tbest: 0.8688709 (81)\ttotal: 16s\tremaining: 39.1s\n",
      "87:\tlearn: 0.8753510\ttest: 0.8753510\ttest1: 0.8685638\tbest: 0.8688709 (81)\ttotal: 16.2s\tremaining: 39s\n",
      "88:\tlearn: 0.8753949\ttest: 0.8753949\ttest1: 0.8687686\tbest: 0.8688709 (81)\ttotal: 16.4s\tremaining: 38.9s\n",
      "89:\tlearn: 0.8753510\ttest: 0.8753510\ttest1: 0.8689733\tbest: 0.8689733 (89)\ttotal: 16.6s\tremaining: 38.7s\n",
      "90:\tlearn: 0.8755704\ttest: 0.8755704\ttest1: 0.8691780\tbest: 0.8691780 (90)\ttotal: 16.8s\tremaining: 38.6s\n",
      "91:\tlearn: 0.8757020\ttest: 0.8757020\ttest1: 0.8691780\tbest: 0.8691780 (90)\ttotal: 16.8s\tremaining: 38.1s\n",
      "92:\tlearn: 0.8754388\ttest: 0.8754388\ttest1: 0.8694851\tbest: 0.8694851 (92)\ttotal: 16.9s\tremaining: 37.6s\n",
      "93:\tlearn: 0.8754388\ttest: 0.8754388\ttest1: 0.8693827\tbest: 0.8694851 (92)\ttotal: 17s\tremaining: 37.1s\n",
      "94:\tlearn: 0.8753949\ttest: 0.8753949\ttest1: 0.8694851\tbest: 0.8694851 (92)\ttotal: 17.2s\tremaining: 37.1s\n",
      "95:\tlearn: 0.8755704\ttest: 0.8755704\ttest1: 0.8696898\tbest: 0.8696898 (95)\ttotal: 17.4s\tremaining: 37s\n",
      "96:\tlearn: 0.8760530\ttest: 0.8760530\ttest1: 0.8694851\tbest: 0.8696898 (95)\ttotal: 17.6s\tremaining: 36.9s\n",
      "97:\tlearn: 0.8759214\ttest: 0.8759214\ttest1: 0.8694851\tbest: 0.8696898 (95)\ttotal: 17.7s\tremaining: 36.5s\n",
      "98:\tlearn: 0.8756581\ttest: 0.8756581\ttest1: 0.8693827\tbest: 0.8696898 (95)\ttotal: 18s\tremaining: 36.5s\n",
      "99:\tlearn: 0.8756581\ttest: 0.8756581\ttest1: 0.8693827\tbest: 0.8696898 (95)\ttotal: 18s\tremaining: 36s\n",
      "100:\tlearn: 0.8759653\ttest: 0.8759653\ttest1: 0.8697922\tbest: 0.8697922 (100)\ttotal: 18.1s\tremaining: 35.6s\n",
      "101:\tlearn: 0.8760530\ttest: 0.8760530\ttest1: 0.8703040\tbest: 0.8703040 (101)\ttotal: 18.1s\tremaining: 35.2s\n",
      "102:\tlearn: 0.8758775\ttest: 0.8758775\ttest1: 0.8704064\tbest: 0.8704064 (102)\ttotal: 18.3s\tremaining: 35.1s\n",
      "103:\tlearn: 0.8760969\ttest: 0.8760969\ttest1: 0.8704064\tbest: 0.8704064 (102)\ttotal: 18.4s\tremaining: 34.7s\n",
      "104:\tlearn: 0.8761408\ttest: 0.8761408\ttest1: 0.8705088\tbest: 0.8705088 (104)\ttotal: 18.6s\tremaining: 34.5s\n",
      "105:\tlearn: 0.8761408\ttest: 0.8761408\ttest1: 0.8705088\tbest: 0.8705088 (104)\ttotal: 18.8s\tremaining: 34.4s\n",
      "106:\tlearn: 0.8761408\ttest: 0.8761408\ttest1: 0.8706111\tbest: 0.8706111 (106)\ttotal: 19s\tremaining: 34.2s\n",
      "107:\tlearn: 0.8760969\ttest: 0.8760969\ttest1: 0.8706111\tbest: 0.8706111 (106)\ttotal: 19.2s\tremaining: 34s\n",
      "108:\tlearn: 0.8763601\ttest: 0.8763601\ttest1: 0.8699969\tbest: 0.8706111 (106)\ttotal: 19.4s\tremaining: 33.9s\n",
      "109:\tlearn: 0.8766673\ttest: 0.8766673\ttest1: 0.8702017\tbest: 0.8706111 (106)\ttotal: 19.6s\tremaining: 33.8s\n",
      "110:\tlearn: 0.8768866\ttest: 0.8768866\ttest1: 0.8703040\tbest: 0.8706111 (106)\ttotal: 19.8s\tremaining: 33.7s\n",
      "111:\tlearn: 0.8769305\ttest: 0.8769305\ttest1: 0.8704064\tbest: 0.8706111 (106)\ttotal: 19.9s\tremaining: 33.3s\n",
      "112:\tlearn: 0.8770621\ttest: 0.8770621\ttest1: 0.8707135\tbest: 0.8707135 (112)\ttotal: 20.1s\tremaining: 33.2s\n",
      "113:\tlearn: 0.8771938\ttest: 0.8771938\ttest1: 0.8710206\tbest: 0.8710206 (113)\ttotal: 20.3s\tremaining: 33.1s\n",
      "114:\tlearn: 0.8772376\ttest: 0.8772376\ttest1: 0.8714300\tbest: 0.8714300 (114)\ttotal: 20.5s\tremaining: 33s\n",
      "115:\tlearn: 0.8772815\ttest: 0.8772815\ttest1: 0.8717371\tbest: 0.8717371 (115)\ttotal: 20.8s\tremaining: 32.9s\n",
      "116:\tlearn: 0.8772376\ttest: 0.8772376\ttest1: 0.8717371\tbest: 0.8717371 (115)\ttotal: 21s\tremaining: 32.8s\n",
      "117:\tlearn: 0.8773693\ttest: 0.8773693\ttest1: 0.8716348\tbest: 0.8717371 (115)\ttotal: 21.2s\tremaining: 32.7s\n",
      "118:\tlearn: 0.8772376\ttest: 0.8772376\ttest1: 0.8712253\tbest: 0.8717371 (115)\ttotal: 21.5s\tremaining: 32.6s\n",
      "119:\tlearn: 0.8774570\ttest: 0.8774570\ttest1: 0.8714300\tbest: 0.8717371 (115)\ttotal: 21.7s\tremaining: 32.5s\n",
      "120:\tlearn: 0.8774570\ttest: 0.8774570\ttest1: 0.8715324\tbest: 0.8717371 (115)\ttotal: 21.7s\tremaining: 32.1s\n",
      "121:\tlearn: 0.8776325\ttest: 0.8776325\ttest1: 0.8714300\tbest: 0.8717371 (115)\ttotal: 21.9s\tremaining: 32s\n",
      "122:\tlearn: 0.8778958\ttest: 0.8778958\ttest1: 0.8717371\tbest: 0.8717371 (115)\ttotal: 22.1s\tremaining: 31.9s\n",
      "123:\tlearn: 0.8778958\ttest: 0.8778958\ttest1: 0.8718395\tbest: 0.8718395 (123)\ttotal: 22.4s\tremaining: 31.8s\n",
      "124:\tlearn: 0.8779396\ttest: 0.8779396\ttest1: 0.8717371\tbest: 0.8718395 (123)\ttotal: 22.6s\tremaining: 31.6s\n",
      "125:\tlearn: 0.8780274\ttest: 0.8780274\ttest1: 0.8719419\tbest: 0.8719419 (125)\ttotal: 22.8s\tremaining: 31.5s\n",
      "126:\tlearn: 0.8778958\ttest: 0.8778958\ttest1: 0.8720442\tbest: 0.8720442 (126)\ttotal: 23.1s\tremaining: 31.4s\n",
      "127:\tlearn: 0.8778958\ttest: 0.8778958\ttest1: 0.8720442\tbest: 0.8720442 (126)\ttotal: 23.3s\tremaining: 31.3s\n",
      "128:\tlearn: 0.8780713\ttest: 0.8780713\ttest1: 0.8720442\tbest: 0.8720442 (126)\ttotal: 23.5s\tremaining: 31.2s\n",
      "129:\tlearn: 0.8779835\ttest: 0.8779835\ttest1: 0.8721466\tbest: 0.8721466 (129)\ttotal: 23.8s\tremaining: 31.1s\n",
      "130:\tlearn: 0.8781151\ttest: 0.8781151\ttest1: 0.8721466\tbest: 0.8721466 (129)\ttotal: 24s\tremaining: 31s\n",
      "131:\tlearn: 0.8781590\ttest: 0.8781590\ttest1: 0.8722490\tbest: 0.8722490 (131)\ttotal: 24.2s\tremaining: 30.8s\n",
      "132:\tlearn: 0.8781590\ttest: 0.8781590\ttest1: 0.8723513\tbest: 0.8723513 (132)\ttotal: 24.5s\tremaining: 30.7s\n",
      "133:\tlearn: 0.8782029\ttest: 0.8782029\ttest1: 0.8723513\tbest: 0.8723513 (132)\ttotal: 24.7s\tremaining: 30.6s\n",
      "134:\tlearn: 0.8783345\ttest: 0.8783345\ttest1: 0.8728631\tbest: 0.8728631 (134)\ttotal: 24.9s\tremaining: 30.4s\n",
      "135:\tlearn: 0.8782468\ttest: 0.8782468\ttest1: 0.8731702\tbest: 0.8731702 (135)\ttotal: 25.1s\tremaining: 30.3s\n",
      "136:\tlearn: 0.8782468\ttest: 0.8782468\ttest1: 0.8732726\tbest: 0.8732726 (136)\ttotal: 25.4s\tremaining: 30.2s\n",
      "137:\tlearn: 0.8784223\ttest: 0.8784223\ttest1: 0.8733750\tbest: 0.8733750 (137)\ttotal: 25.6s\tremaining: 30.1s\n",
      "138:\tlearn: 0.8783784\ttest: 0.8783784\ttest1: 0.8733750\tbest: 0.8733750 (137)\ttotal: 25.9s\tremaining: 30s\n",
      "139:\tlearn: 0.8784223\ttest: 0.8784223\ttest1: 0.8734773\tbest: 0.8734773 (139)\ttotal: 26.1s\tremaining: 29.9s\n",
      "140:\tlearn: 0.8786416\ttest: 0.8786416\ttest1: 0.8733750\tbest: 0.8734773 (139)\ttotal: 26.4s\tremaining: 29.7s\n",
      "141:\tlearn: 0.8785978\ttest: 0.8785978\ttest1: 0.8733750\tbest: 0.8734773 (139)\ttotal: 26.7s\tremaining: 29.7s\n",
      "142:\tlearn: 0.8785539\ttest: 0.8785539\ttest1: 0.8734773\tbest: 0.8734773 (139)\ttotal: 26.9s\tremaining: 29.6s\n",
      "143:\tlearn: 0.8785539\ttest: 0.8785539\ttest1: 0.8734773\tbest: 0.8734773 (139)\ttotal: 27.2s\tremaining: 29.4s\n",
      "144:\tlearn: 0.8784223\ttest: 0.8784223\ttest1: 0.8735797\tbest: 0.8735797 (144)\ttotal: 27.5s\tremaining: 29.4s\n",
      "145:\tlearn: 0.8785100\ttest: 0.8785100\ttest1: 0.8735797\tbest: 0.8735797 (144)\ttotal: 27.6s\tremaining: 29.1s\n",
      "146:\tlearn: 0.8784661\ttest: 0.8784661\ttest1: 0.8735797\tbest: 0.8735797 (144)\ttotal: 27.9s\tremaining: 29.1s\n",
      "147:\tlearn: 0.8785539\ttest: 0.8785539\ttest1: 0.8735797\tbest: 0.8735797 (144)\ttotal: 28.2s\tremaining: 28.9s\n",
      "148:\tlearn: 0.8791243\ttest: 0.8791243\ttest1: 0.8739891\tbest: 0.8739891 (148)\ttotal: 28.4s\tremaining: 28.8s\n",
      "149:\tlearn: 0.8792998\ttest: 0.8792998\ttest1: 0.8740915\tbest: 0.8740915 (149)\ttotal: 28.7s\tremaining: 28.7s\n",
      "150:\tlearn: 0.8792998\ttest: 0.8792998\ttest1: 0.8740915\tbest: 0.8740915 (149)\ttotal: 28.9s\tremaining: 28.5s\n",
      "151:\tlearn: 0.8792998\ttest: 0.8792998\ttest1: 0.8741939\tbest: 0.8741939 (151)\ttotal: 29.1s\tremaining: 28.3s\n",
      "152:\tlearn: 0.8792998\ttest: 0.8792998\ttest1: 0.8742962\tbest: 0.8742962 (152)\ttotal: 29.3s\tremaining: 28.2s\n",
      "153:\tlearn: 0.8797824\ttest: 0.8797824\ttest1: 0.8745010\tbest: 0.8745010 (153)\ttotal: 29.6s\tremaining: 28s\n",
      "154:\tlearn: 0.8798701\ttest: 0.8798701\ttest1: 0.8745010\tbest: 0.8745010 (153)\ttotal: 29.8s\tremaining: 27.8s\n",
      "155:\tlearn: 0.8800018\ttest: 0.8800018\ttest1: 0.8745010\tbest: 0.8745010 (153)\ttotal: 30s\tremaining: 27.7s\n",
      "156:\tlearn: 0.8798701\ttest: 0.8798701\ttest1: 0.8742962\tbest: 0.8745010 (153)\ttotal: 30.2s\tremaining: 27.5s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "157:\tlearn: 0.8799140\ttest: 0.8799140\ttest1: 0.8742962\tbest: 0.8745010 (153)\ttotal: 30.4s\tremaining: 27.4s\n",
      "158:\tlearn: 0.8799140\ttest: 0.8799140\ttest1: 0.8745010\tbest: 0.8745010 (153)\ttotal: 30.6s\tremaining: 27.2s\n",
      "159:\tlearn: 0.8799140\ttest: 0.8799140\ttest1: 0.8747057\tbest: 0.8747057 (159)\ttotal: 30.9s\tremaining: 27s\n",
      "160:\tlearn: 0.8800018\ttest: 0.8800018\ttest1: 0.8742962\tbest: 0.8747057 (159)\ttotal: 31.1s\tremaining: 26.8s\n",
      "161:\tlearn: 0.8801334\ttest: 0.8801334\ttest1: 0.8745010\tbest: 0.8747057 (159)\ttotal: 31.3s\tremaining: 26.6s\n",
      "162:\tlearn: 0.8801334\ttest: 0.8801334\ttest1: 0.8743986\tbest: 0.8747057 (159)\ttotal: 31.5s\tremaining: 26.5s\n",
      "163:\tlearn: 0.8801334\ttest: 0.8801334\ttest1: 0.8742962\tbest: 0.8747057 (159)\ttotal: 31.7s\tremaining: 26.3s\n",
      "164:\tlearn: 0.8800456\ttest: 0.8800456\ttest1: 0.8742962\tbest: 0.8747057 (159)\ttotal: 31.9s\tremaining: 26.1s\n",
      "165:\tlearn: 0.8801773\ttest: 0.8801773\ttest1: 0.8742962\tbest: 0.8747057 (159)\ttotal: 32.2s\tremaining: 26s\n",
      "166:\tlearn: 0.8802211\ttest: 0.8802211\ttest1: 0.8742962\tbest: 0.8747057 (159)\ttotal: 32.4s\tremaining: 25.8s\n",
      "167:\tlearn: 0.8800456\ttest: 0.8800456\ttest1: 0.8743986\tbest: 0.8747057 (159)\ttotal: 32.6s\tremaining: 25.6s\n",
      "168:\tlearn: 0.8800018\ttest: 0.8800018\ttest1: 0.8745010\tbest: 0.8747057 (159)\ttotal: 32.9s\tremaining: 25.5s\n",
      "169:\tlearn: 0.8800018\ttest: 0.8800018\ttest1: 0.8743986\tbest: 0.8747057 (159)\ttotal: 33.1s\tremaining: 25.3s\n",
      "170:\tlearn: 0.8803089\ttest: 0.8803089\ttest1: 0.8743986\tbest: 0.8747057 (159)\ttotal: 33.3s\tremaining: 25.1s\n",
      "171:\tlearn: 0.8803089\ttest: 0.8803089\ttest1: 0.8743986\tbest: 0.8747057 (159)\ttotal: 33.4s\tremaining: 24.8s\n",
      "172:\tlearn: 0.8803089\ttest: 0.8803089\ttest1: 0.8746033\tbest: 0.8747057 (159)\ttotal: 33.6s\tremaining: 24.7s\n",
      "173:\tlearn: 0.8803089\ttest: 0.8803089\ttest1: 0.8746033\tbest: 0.8747057 (159)\ttotal: 33.7s\tremaining: 24.4s\n",
      "174:\tlearn: 0.8802650\ttest: 0.8802650\ttest1: 0.8745010\tbest: 0.8747057 (159)\ttotal: 33.9s\tremaining: 24.2s\n",
      "175:\tlearn: 0.8803089\ttest: 0.8803089\ttest1: 0.8745010\tbest: 0.8747057 (159)\ttotal: 34.1s\tremaining: 24s\n",
      "176:\tlearn: 0.8803089\ttest: 0.8803089\ttest1: 0.8743986\tbest: 0.8747057 (159)\ttotal: 34.3s\tremaining: 23.9s\n",
      "177:\tlearn: 0.8802211\ttest: 0.8802211\ttest1: 0.8743986\tbest: 0.8747057 (159)\ttotal: 34.5s\tremaining: 23.7s\n",
      "178:\tlearn: 0.8801773\ttest: 0.8801773\ttest1: 0.8743986\tbest: 0.8747057 (159)\ttotal: 34.8s\tremaining: 23.5s\n",
      "179:\tlearn: 0.8801773\ttest: 0.8801773\ttest1: 0.8743986\tbest: 0.8747057 (159)\ttotal: 34.8s\tremaining: 23.2s\n",
      "180:\tlearn: 0.8802211\ttest: 0.8802211\ttest1: 0.8745010\tbest: 0.8747057 (159)\ttotal: 35.1s\tremaining: 23.1s\n",
      "181:\tlearn: 0.8803089\ttest: 0.8803089\ttest1: 0.8740915\tbest: 0.8747057 (159)\ttotal: 35.3s\tremaining: 22.9s\n",
      "182:\tlearn: 0.8804405\ttest: 0.8804405\ttest1: 0.8739891\tbest: 0.8747057 (159)\ttotal: 35.5s\tremaining: 22.7s\n",
      "183:\tlearn: 0.8804844\ttest: 0.8804844\ttest1: 0.8738868\tbest: 0.8747057 (159)\ttotal: 35.8s\tremaining: 22.5s\n",
      "184:\tlearn: 0.8807476\ttest: 0.8807476\ttest1: 0.8739891\tbest: 0.8747057 (159)\ttotal: 36s\tremaining: 22.4s\n",
      "185:\tlearn: 0.8807476\ttest: 0.8807476\ttest1: 0.8741939\tbest: 0.8747057 (159)\ttotal: 36.2s\tremaining: 22.2s\n",
      "186:\tlearn: 0.8807476\ttest: 0.8807476\ttest1: 0.8741939\tbest: 0.8747057 (159)\ttotal: 36.3s\tremaining: 21.9s\n",
      "187:\tlearn: 0.8808354\ttest: 0.8808354\ttest1: 0.8741939\tbest: 0.8747057 (159)\ttotal: 36.5s\tremaining: 21.7s\n",
      "188:\tlearn: 0.8808354\ttest: 0.8808354\ttest1: 0.8741939\tbest: 0.8747057 (159)\ttotal: 36.7s\tremaining: 21.6s\n",
      "189:\tlearn: 0.8810986\ttest: 0.8810986\ttest1: 0.8746033\tbest: 0.8747057 (159)\ttotal: 37s\tremaining: 21.4s\n",
      "190:\tlearn: 0.8811425\ttest: 0.8811425\ttest1: 0.8745010\tbest: 0.8747057 (159)\ttotal: 37.2s\tremaining: 21.2s\n",
      "191:\tlearn: 0.8810109\ttest: 0.8810109\ttest1: 0.8743986\tbest: 0.8747057 (159)\ttotal: 37.5s\tremaining: 21.1s\n",
      "192:\tlearn: 0.8810109\ttest: 0.8810109\ttest1: 0.8741939\tbest: 0.8747057 (159)\ttotal: 37.8s\tremaining: 20.9s\n",
      "193:\tlearn: 0.8810986\ttest: 0.8810986\ttest1: 0.8741939\tbest: 0.8747057 (159)\ttotal: 38s\tremaining: 20.8s\n",
      "194:\tlearn: 0.8811425\ttest: 0.8811425\ttest1: 0.8747057\tbest: 0.8747057 (159)\ttotal: 38.3s\tremaining: 20.6s\n",
      "195:\tlearn: 0.8815813\ttest: 0.8815813\ttest1: 0.8742962\tbest: 0.8747057 (159)\ttotal: 38.5s\tremaining: 20.4s\n",
      "196:\tlearn: 0.8815813\ttest: 0.8815813\ttest1: 0.8742962\tbest: 0.8747057 (159)\ttotal: 38.7s\tremaining: 20.2s\n",
      "197:\tlearn: 0.8815813\ttest: 0.8815813\ttest1: 0.8742962\tbest: 0.8747057 (159)\ttotal: 38.9s\tremaining: 20.1s\n",
      "198:\tlearn: 0.8817129\ttest: 0.8817129\ttest1: 0.8743986\tbest: 0.8747057 (159)\ttotal: 39.2s\tremaining: 19.9s\n",
      "199:\tlearn: 0.8821516\ttest: 0.8821516\ttest1: 0.8745010\tbest: 0.8747057 (159)\ttotal: 39.4s\tremaining: 19.7s\n",
      "200:\tlearn: 0.8821516\ttest: 0.8821516\ttest1: 0.8745010\tbest: 0.8747057 (159)\ttotal: 39.6s\tremaining: 19.5s\n",
      "201:\tlearn: 0.8823710\ttest: 0.8823710\ttest1: 0.8745010\tbest: 0.8747057 (159)\ttotal: 39.8s\tremaining: 19.3s\n",
      "202:\tlearn: 0.8823271\ttest: 0.8823271\ttest1: 0.8743986\tbest: 0.8747057 (159)\ttotal: 40s\tremaining: 19.1s\n",
      "203:\tlearn: 0.8821955\ttest: 0.8821955\ttest1: 0.8742962\tbest: 0.8747057 (159)\ttotal: 40.3s\tremaining: 19s\n",
      "204:\tlearn: 0.8822833\ttest: 0.8822833\ttest1: 0.8746033\tbest: 0.8747057 (159)\ttotal: 40.5s\tremaining: 18.8s\n",
      "205:\tlearn: 0.8822833\ttest: 0.8822833\ttest1: 0.8743986\tbest: 0.8747057 (159)\ttotal: 40.7s\tremaining: 18.6s\n",
      "206:\tlearn: 0.8825026\ttest: 0.8825026\ttest1: 0.8745010\tbest: 0.8747057 (159)\ttotal: 40.9s\tremaining: 18.4s\n",
      "207:\tlearn: 0.8826781\ttest: 0.8826781\ttest1: 0.8743986\tbest: 0.8747057 (159)\ttotal: 41.2s\tremaining: 18.2s\n",
      "208:\tlearn: 0.8825026\ttest: 0.8825026\ttest1: 0.8743986\tbest: 0.8747057 (159)\ttotal: 41.4s\tremaining: 18s\n",
      "209:\tlearn: 0.8824588\ttest: 0.8824588\ttest1: 0.8745010\tbest: 0.8747057 (159)\ttotal: 41.6s\tremaining: 17.8s\n",
      "Stopped by overfitting detector  (50 iterations wait)\n",
      "\n",
      "bestTest = 0.8747057017\n",
      "bestIteration = 159\n",
      "\n",
      "Shrink model to first 160 iterations.\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<catboost.core.CatBoostClassifier at 0x1a28df85f8>"
      ]
     },
     "execution_count": 168,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "catbCV_grid.fit(X_train,y_train,eval_set=[(X_train,y_train), (X_test,y_test)],early_stopping_rounds=50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 169,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<catboost.core.CatBoostClassifier object at 0x1a28df85f8>\n",
      "confusion matrix of training data\n",
      "[[16429   862]\n",
      " [ 1875  3626]]\n",
      "confusion matrix of test data\n",
      "[[7057  372]\n",
      " [ 852 1488]]\n",
      "classification report of test data\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.89      0.95      0.92      7429\n",
      "           1       0.80      0.64      0.71      2340\n",
      "\n",
      "    accuracy                           0.87      9769\n",
      "   macro avg       0.85      0.79      0.81      9769\n",
      "weighted avg       0.87      0.87      0.87      9769\n",
      "\n",
      "Accuracy on test data: 87.470570%\n",
      "Accuracy on training data: 87.991400%\n",
      "Area under the ROC curve : 0.792912\n"
     ]
    }
   ],
   "source": [
    "show_classifier_metrics(catbCV_grid)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/debashis/anaconda3/lib/python3.6/site-packages/sklearn/externals/six.py:31: DeprecationWarning: The module is deprecated in version 0.21 and will be removed in version 0.23 since we've dropped support for Python 2.7. Please rely on the official version of six (https://pypi.org/project/six/).\n",
      "  \"(https://pypi.org/project/six/).\", DeprecationWarning)\n"
     ]
    }
   ],
   "source": [
    "from imblearn.over_sampling import RandomOverSampler"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.24135661635661634"
      ]
     },
     "execution_count": 125,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.sum(y_train)/y_train.shape[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 139,
   "metadata": {},
   "outputs": [],
   "source": [
    "ros = RandomOverSampler(random_state=1,sampling_strategy=0.8)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 140,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_resampled, y_resampled = ros.fit_resample(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 150,
   "metadata": {},
   "outputs": [],
   "source": [
    "catb_ros = CatBoostClassifier(learning_rate=0.1,iterations=400,reg_lambda=2,verbose=1,random_seed=10,eval_metric='Accuracy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 151,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0:\tlearn: 0.8032645\ttest: 0.8032645\ttest1: 0.7948613\tbest: 0.7948613 (0)\ttotal: 63.4ms\tremaining: 25.3s\n",
      "1:\tlearn: 0.8070559\ttest: 0.8070559\ttest1: 0.7912785\tbest: 0.7948613 (0)\ttotal: 127ms\tremaining: 25.4s\n",
      "2:\tlearn: 0.8115220\ttest: 0.8115220\ttest1: 0.7997748\tbest: 0.7997748 (2)\ttotal: 202ms\tremaining: 26.8s\n",
      "3:\tlearn: 0.8085018\ttest: 0.8085018\ttest1: 0.7948613\tbest: 0.7997748 (2)\ttotal: 261ms\tremaining: 25.9s\n",
      "4:\tlearn: 0.8138354\ttest: 0.8138354\ttest1: 0.8037670\tbest: 0.8037670 (4)\ttotal: 341ms\tremaining: 26.9s\n",
      "5:\tlearn: 0.8156990\ttest: 0.8156990\ttest1: 0.8057119\tbest: 0.8057119 (5)\ttotal: 398ms\tremaining: 26.1s\n",
      "6:\tlearn: 0.8169200\ttest: 0.8169200\ttest1: 0.8088852\tbest: 0.8088852 (6)\ttotal: 456ms\tremaining: 25.6s\n",
      "7:\tlearn: 0.8167272\ttest: 0.8167272\ttest1: 0.8075545\tbest: 0.8088852 (6)\ttotal: 514ms\tremaining: 25.2s\n",
      "8:\tlearn: 0.8186871\ttest: 0.8186871\ttest1: 0.8112396\tbest: 0.8112396 (8)\ttotal: 585ms\tremaining: 25.4s\n",
      "9:\tlearn: 0.8201009\ttest: 0.8201009\ttest1: 0.8128775\tbest: 0.8128775 (9)\ttotal: 644ms\tremaining: 25.1s\n",
      "10:\tlearn: 0.8215146\ttest: 0.8215146\ttest1: 0.8132869\tbest: 0.8132869 (10)\ttotal: 737ms\tremaining: 26.1s\n",
      "11:\tlearn: 0.8218359\ttest: 0.8218359\ttest1: 0.8142082\tbest: 0.8142082 (11)\ttotal: 801ms\tremaining: 25.9s\n",
      "12:\tlearn: 0.8235067\ttest: 0.8235067\ttest1: 0.8162555\tbest: 0.8162555 (12)\ttotal: 872ms\tremaining: 25.9s\n",
      "13:\tlearn: 0.8245028\ttest: 0.8245028\ttest1: 0.8161531\tbest: 0.8162555 (12)\ttotal: 930ms\tremaining: 25.6s\n",
      "14:\tlearn: 0.8242136\ttest: 0.8242136\ttest1: 0.8165626\tbest: 0.8165626 (14)\ttotal: 984ms\tremaining: 25.3s\n",
      "15:\tlearn: 0.8251133\ttest: 0.8251133\ttest1: 0.8168697\tbest: 0.8168697 (15)\ttotal: 1.06s\tremaining: 25.5s\n",
      "16:\tlearn: 0.8267519\ttest: 0.8267519\ttest1: 0.8202477\tbest: 0.8202477 (16)\ttotal: 1.19s\tremaining: 26.9s\n",
      "17:\tlearn: 0.8276194\ttest: 0.8276194\ttest1: 0.8217832\tbest: 0.8217832 (17)\ttotal: 1.35s\tremaining: 28.8s\n",
      "18:\tlearn: 0.8279086\ttest: 0.8279086\ttest1: 0.8234210\tbest: 0.8234210 (18)\ttotal: 1.52s\tremaining: 30.6s\n",
      "19:\tlearn: 0.8282942\ttest: 0.8282942\ttest1: 0.8246494\tbest: 0.8246494 (19)\ttotal: 1.61s\tremaining: 30.5s\n",
      "20:\tlearn: 0.8296758\ttest: 0.8296758\ttest1: 0.8253660\tbest: 0.8253660 (20)\ttotal: 1.67s\tremaining: 30.1s\n",
      "21:\tlearn: 0.8298365\ttest: 0.8298365\ttest1: 0.8251612\tbest: 0.8253660 (20)\ttotal: 1.73s\tremaining: 29.6s\n",
      "22:\tlearn: 0.8302863\ttest: 0.8302863\ttest1: 0.8258778\tbest: 0.8258778 (22)\ttotal: 1.89s\tremaining: 31s\n",
      "23:\tlearn: 0.8309289\ttest: 0.8309289\ttest1: 0.8260825\tbest: 0.8260825 (23)\ttotal: 2.01s\tremaining: 31.4s\n",
      "24:\tlearn: 0.8328889\ttest: 0.8328889\ttest1: 0.8299724\tbest: 0.8299724 (24)\ttotal: 2.08s\tremaining: 31.2s\n",
      "25:\tlearn: 0.8329210\ttest: 0.8329210\ttest1: 0.8299724\tbest: 0.8299724 (24)\ttotal: 2.16s\tremaining: 31.1s\n",
      "26:\tlearn: 0.8330495\ttest: 0.8330495\ttest1: 0.8296653\tbest: 0.8299724 (24)\ttotal: 2.22s\tremaining: 30.7s\n",
      "27:\tlearn: 0.8337885\ttest: 0.8337885\ttest1: 0.8290511\tbest: 0.8299724 (24)\ttotal: 2.28s\tremaining: 30.3s\n",
      "28:\tlearn: 0.8340777\ttest: 0.8340777\ttest1: 0.8301771\tbest: 0.8301771 (28)\ttotal: 2.46s\tremaining: 31.4s\n",
      "29:\tlearn: 0.8347524\ttest: 0.8347524\ttest1: 0.8293582\tbest: 0.8301771 (28)\ttotal: 2.63s\tremaining: 32.5s\n",
      "30:\tlearn: 0.8350737\ttest: 0.8350737\ttest1: 0.8298700\tbest: 0.8301771 (28)\ttotal: 2.8s\tremaining: 33.3s\n",
      "31:\tlearn: 0.8350095\ttest: 0.8350095\ttest1: 0.8330433\tbest: 0.8330433 (31)\ttotal: 2.94s\tremaining: 33.8s\n",
      "32:\tlearn: 0.8352023\ttest: 0.8352023\ttest1: 0.8328386\tbest: 0.8330433 (31)\ttotal: 3s\tremaining: 33.3s\n",
      "33:\tlearn: 0.8351380\ttest: 0.8351380\ttest1: 0.8331457\tbest: 0.8331457 (33)\ttotal: 3.07s\tremaining: 33.1s\n",
      "34:\tlearn: 0.8349773\ttest: 0.8349773\ttest1: 0.8339646\tbest: 0.8339646 (34)\ttotal: 3.14s\tremaining: 32.7s\n",
      "35:\tlearn: 0.8351380\ttest: 0.8351380\ttest1: 0.8342717\tbest: 0.8342717 (35)\ttotal: 3.2s\tremaining: 32.3s\n",
      "36:\tlearn: 0.8359734\ttest: 0.8359734\ttest1: 0.8343740\tbest: 0.8343740 (36)\ttotal: 3.25s\tremaining: 31.9s\n",
      "37:\tlearn: 0.8359091\ttest: 0.8359091\ttest1: 0.8338622\tbest: 0.8343740 (36)\ttotal: 3.33s\tremaining: 31.8s\n",
      "38:\tlearn: 0.8363590\ttest: 0.8363590\ttest1: 0.8348859\tbest: 0.8348859 (38)\ttotal: 3.39s\tremaining: 31.4s\n",
      "39:\tlearn: 0.8363590\ttest: 0.8363590\ttest1: 0.8337599\tbest: 0.8348859 (38)\ttotal: 3.45s\tremaining: 31s\n",
      "40:\tlearn: 0.8368409\ttest: 0.8368409\ttest1: 0.8346811\tbest: 0.8348859 (38)\ttotal: 3.51s\tremaining: 30.7s\n",
      "41:\tlearn: 0.8370980\ttest: 0.8370980\ttest1: 0.8347835\tbest: 0.8348859 (38)\ttotal: 3.58s\tremaining: 30.6s\n",
      "42:\tlearn: 0.8369052\ttest: 0.8369052\ttest1: 0.8356024\tbest: 0.8356024 (42)\ttotal: 3.64s\tremaining: 30.2s\n",
      "43:\tlearn: 0.8369052\ttest: 0.8369052\ttest1: 0.8352953\tbest: 0.8356024 (42)\ttotal: 3.7s\tremaining: 29.9s\n",
      "44:\tlearn: 0.8375157\ttest: 0.8375157\ttest1: 0.8352953\tbest: 0.8356024 (42)\ttotal: 3.76s\tremaining: 29.6s\n",
      "45:\tlearn: 0.8382547\ttest: 0.8382547\ttest1: 0.8372402\tbest: 0.8372402 (45)\ttotal: 3.83s\tremaining: 29.4s\n",
      "46:\tlearn: 0.8386724\ttest: 0.8386724\ttest1: 0.8369332\tbest: 0.8372402 (45)\ttotal: 3.88s\tremaining: 29.2s\n",
      "47:\tlearn: 0.8389615\ttest: 0.8389615\ttest1: 0.8375473\tbest: 0.8375473 (47)\ttotal: 3.94s\tremaining: 28.9s\n",
      "48:\tlearn: 0.8391543\ttest: 0.8391543\ttest1: 0.8380592\tbest: 0.8380592 (48)\ttotal: 3.99s\tremaining: 28.6s\n",
      "49:\tlearn: 0.8388973\ttest: 0.8388973\ttest1: 0.8377521\tbest: 0.8380592 (48)\ttotal: 4.06s\tremaining: 28.5s\n",
      "50:\tlearn: 0.8391222\ttest: 0.8391222\ttest1: 0.8373426\tbest: 0.8380592 (48)\ttotal: 4.12s\tremaining: 28.2s\n",
      "51:\tlearn: 0.8407287\ttest: 0.8407287\ttest1: 0.8387757\tbest: 0.8387757 (51)\ttotal: 4.17s\tremaining: 27.9s\n",
      "52:\tlearn: 0.8406966\ttest: 0.8406966\ttest1: 0.8385710\tbest: 0.8387757 (51)\ttotal: 4.23s\tremaining: 27.7s\n",
      "53:\tlearn: 0.8403432\ttest: 0.8403432\ttest1: 0.8391852\tbest: 0.8391852 (53)\ttotal: 4.29s\tremaining: 27.5s\n",
      "54:\tlearn: 0.8404717\ttest: 0.8404717\ttest1: 0.8392875\tbest: 0.8392875 (54)\ttotal: 4.37s\tremaining: 27.4s\n",
      "55:\tlearn: 0.8405681\ttest: 0.8405681\ttest1: 0.8387757\tbest: 0.8392875 (54)\ttotal: 4.43s\tremaining: 27.2s\n",
      "56:\tlearn: 0.8415962\ttest: 0.8415962\ttest1: 0.8400041\tbest: 0.8400041 (56)\ttotal: 4.48s\tremaining: 27s\n",
      "57:\tlearn: 0.8416605\ttest: 0.8416605\ttest1: 0.8405159\tbest: 0.8405159 (57)\ttotal: 4.55s\tremaining: 26.9s\n",
      "58:\tlearn: 0.8412749\ttest: 0.8412749\ttest1: 0.8412325\tbest: 0.8412325 (58)\ttotal: 4.61s\tremaining: 26.6s\n",
      "59:\tlearn: 0.8412428\ttest: 0.8412428\ttest1: 0.8416419\tbest: 0.8416419 (59)\ttotal: 4.66s\tremaining: 26.4s\n",
      "60:\tlearn: 0.8416926\ttest: 0.8416926\ttest1: 0.8416419\tbest: 0.8416419 (59)\ttotal: 4.72s\tremaining: 26.2s\n",
      "61:\tlearn: 0.8423353\ttest: 0.8423353\ttest1: 0.8410277\tbest: 0.8416419 (59)\ttotal: 4.77s\tremaining: 26s\n",
      "62:\tlearn: 0.8428493\ttest: 0.8428493\ttest1: 0.8420514\tbest: 0.8420514 (62)\ttotal: 4.83s\tremaining: 25.8s\n",
      "63:\tlearn: 0.8432670\ttest: 0.8432670\ttest1: 0.8409254\tbest: 0.8420514 (62)\ttotal: 4.89s\tremaining: 25.7s\n",
      "64:\tlearn: 0.8428815\ttest: 0.8428815\ttest1: 0.8421538\tbest: 0.8421538 (64)\ttotal: 4.95s\tremaining: 25.5s\n",
      "65:\tlearn: 0.8432028\ttest: 0.8432028\ttest1: 0.8408230\tbest: 0.8421538 (64)\ttotal: 5.04s\tremaining: 25.5s\n",
      "66:\tlearn: 0.8438133\ttest: 0.8438133\ttest1: 0.8414372\tbest: 0.8421538 (64)\ttotal: 5.09s\tremaining: 25.3s\n",
      "67:\tlearn: 0.8438454\ttest: 0.8438454\ttest1: 0.8416419\tbest: 0.8421538 (64)\ttotal: 5.29s\tremaining: 25.8s\n",
      "68:\tlearn: 0.8439739\ttest: 0.8439739\ttest1: 0.8419490\tbest: 0.8421538 (64)\ttotal: 5.46s\tremaining: 26.2s\n",
      "69:\tlearn: 0.8443595\ttest: 0.8443595\ttest1: 0.8418467\tbest: 0.8421538 (64)\ttotal: 5.62s\tremaining: 26.5s\n",
      "70:\tlearn: 0.8443273\ttest: 0.8443273\ttest1: 0.8419490\tbest: 0.8421538 (64)\ttotal: 5.79s\tremaining: 26.8s\n",
      "71:\tlearn: 0.8445523\ttest: 0.8445523\ttest1: 0.8417443\tbest: 0.8421538 (64)\ttotal: 5.96s\tremaining: 27.2s\n",
      "72:\tlearn: 0.8441667\ttest: 0.8441667\ttest1: 0.8423585\tbest: 0.8423585 (72)\ttotal: 6.14s\tremaining: 27.5s\n",
      "73:\tlearn: 0.8443595\ttest: 0.8443595\ttest1: 0.8429727\tbest: 0.8429727 (73)\ttotal: 6.26s\tremaining: 27.6s\n",
      "74:\tlearn: 0.8442952\ttest: 0.8442952\ttest1: 0.8426656\tbest: 0.8429727 (73)\ttotal: 6.32s\tremaining: 27.4s\n",
      "75:\tlearn: 0.8447772\ttest: 0.8447772\ttest1: 0.8424608\tbest: 0.8429727 (73)\ttotal: 6.38s\tremaining: 27.2s\n",
      "76:\tlearn: 0.8451949\ttest: 0.8451949\ttest1: 0.8434845\tbest: 0.8434845 (76)\ttotal: 6.43s\tremaining: 27s\n",
      "77:\tlearn: 0.8451306\ttest: 0.8451306\ttest1: 0.8435869\tbest: 0.8435869 (77)\ttotal: 6.49s\tremaining: 26.8s\n",
      "78:\tlearn: 0.8453234\ttest: 0.8453234\ttest1: 0.8440987\tbest: 0.8440987 (78)\ttotal: 6.55s\tremaining: 26.6s\n",
      "79:\tlearn: 0.8451949\ttest: 0.8451949\ttest1: 0.8439963\tbest: 0.8440987 (78)\ttotal: 6.61s\tremaining: 26.4s\n",
      "80:\tlearn: 0.8455483\ttest: 0.8455483\ttest1: 0.8437916\tbest: 0.8440987 (78)\ttotal: 6.66s\tremaining: 26.2s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "81:\tlearn: 0.8458375\ttest: 0.8458375\ttest1: 0.8435869\tbest: 0.8440987 (78)\ttotal: 6.73s\tremaining: 26.1s\n",
      "82:\tlearn: 0.8460945\ttest: 0.8460945\ttest1: 0.8444058\tbest: 0.8444058 (82)\ttotal: 6.78s\tremaining: 25.9s\n",
      "83:\tlearn: 0.8459017\ttest: 0.8459017\ttest1: 0.8445081\tbest: 0.8445081 (83)\ttotal: 6.84s\tremaining: 25.7s\n",
      "84:\tlearn: 0.8461588\ttest: 0.8461588\ttest1: 0.8454294\tbest: 0.8454294 (84)\ttotal: 6.89s\tremaining: 25.5s\n",
      "85:\tlearn: 0.8457732\ttest: 0.8457732\ttest1: 0.8456341\tbest: 0.8456341 (85)\ttotal: 6.96s\tremaining: 25.4s\n",
      "86:\tlearn: 0.8459981\ttest: 0.8459981\ttest1: 0.8456341\tbest: 0.8456341 (85)\ttotal: 7s\tremaining: 25.2s\n",
      "87:\tlearn: 0.8462552\ttest: 0.8462552\ttest1: 0.8449176\tbest: 0.8456341 (85)\ttotal: 7.1s\tremaining: 25.2s\n",
      "88:\tlearn: 0.8465122\ttest: 0.8465122\ttest1: 0.8452247\tbest: 0.8456341 (85)\ttotal: 7.2s\tremaining: 25.1s\n",
      "89:\tlearn: 0.8464801\ttest: 0.8464801\ttest1: 0.8455318\tbest: 0.8456341 (85)\ttotal: 7.29s\tremaining: 25.1s\n",
      "90:\tlearn: 0.8464480\ttest: 0.8464480\ttest1: 0.8457365\tbest: 0.8457365 (90)\ttotal: 7.39s\tremaining: 25.1s\n",
      "91:\tlearn: 0.8466086\ttest: 0.8466086\ttest1: 0.8457365\tbest: 0.8457365 (90)\ttotal: 7.48s\tremaining: 25s\n",
      "92:\tlearn: 0.8467050\ttest: 0.8467050\ttest1: 0.8458389\tbest: 0.8458389 (92)\ttotal: 7.54s\tremaining: 24.9s\n",
      "93:\tlearn: 0.8470584\ttest: 0.8470584\ttest1: 0.8455318\tbest: 0.8458389 (92)\ttotal: 7.63s\tremaining: 24.8s\n",
      "94:\tlearn: 0.8472191\ttest: 0.8472191\ttest1: 0.8451223\tbest: 0.8458389 (92)\ttotal: 7.68s\tremaining: 24.7s\n",
      "95:\tlearn: 0.8477653\ttest: 0.8477653\ttest1: 0.8461460\tbest: 0.8461460 (95)\ttotal: 7.74s\tremaining: 24.5s\n",
      "96:\tlearn: 0.8477653\ttest: 0.8477653\ttest1: 0.8460436\tbest: 0.8461460 (95)\ttotal: 7.8s\tremaining: 24.4s\n",
      "97:\tlearn: 0.8480866\ttest: 0.8480866\ttest1: 0.8467602\tbest: 0.8467602 (97)\ttotal: 7.96s\tremaining: 24.5s\n",
      "98:\tlearn: 0.8482794\ttest: 0.8482794\ttest1: 0.8465554\tbest: 0.8467602 (97)\ttotal: 8.12s\tremaining: 24.7s\n",
      "99:\tlearn: 0.8486328\ttest: 0.8486328\ttest1: 0.8465554\tbest: 0.8467602 (97)\ttotal: 8.2s\tremaining: 24.6s\n",
      "100:\tlearn: 0.8484722\ttest: 0.8484722\ttest1: 0.8467602\tbest: 0.8467602 (97)\ttotal: 8.26s\tremaining: 24.5s\n",
      "101:\tlearn: 0.8492433\ttest: 0.8492433\ttest1: 0.8467602\tbest: 0.8467602 (97)\ttotal: 8.32s\tremaining: 24.3s\n",
      "102:\tlearn: 0.8493076\ttest: 0.8493076\ttest1: 0.8464531\tbest: 0.8467602 (97)\ttotal: 8.37s\tremaining: 24.1s\n",
      "103:\tlearn: 0.8494682\ttest: 0.8494682\ttest1: 0.8468625\tbest: 0.8468625 (103)\ttotal: 8.44s\tremaining: 24s\n",
      "104:\tlearn: 0.8493718\ttest: 0.8493718\ttest1: 0.8470673\tbest: 0.8470673 (104)\ttotal: 8.49s\tremaining: 23.9s\n",
      "105:\tlearn: 0.8494040\ttest: 0.8494040\ttest1: 0.8470673\tbest: 0.8470673 (104)\ttotal: 8.54s\tremaining: 23.7s\n",
      "106:\tlearn: 0.8495646\ttest: 0.8495646\ttest1: 0.8471696\tbest: 0.8471696 (106)\ttotal: 8.6s\tremaining: 23.5s\n",
      "107:\tlearn: 0.8497253\ttest: 0.8497253\ttest1: 0.8470673\tbest: 0.8471696 (106)\ttotal: 8.65s\tremaining: 23.4s\n",
      "108:\tlearn: 0.8500787\ttest: 0.8500787\ttest1: 0.8461460\tbest: 0.8471696 (106)\ttotal: 8.71s\tremaining: 23.3s\n",
      "109:\tlearn: 0.8500787\ttest: 0.8500787\ttest1: 0.8460436\tbest: 0.8471696 (106)\ttotal: 8.77s\tremaining: 23.1s\n",
      "110:\tlearn: 0.8501109\ttest: 0.8501109\ttest1: 0.8461460\tbest: 0.8471696 (106)\ttotal: 8.83s\tremaining: 23s\n",
      "111:\tlearn: 0.8506571\ttest: 0.8506571\ttest1: 0.8459412\tbest: 0.8471696 (106)\ttotal: 8.9s\tremaining: 22.9s\n",
      "112:\tlearn: 0.8505928\ttest: 0.8505928\ttest1: 0.8458389\tbest: 0.8471696 (106)\ttotal: 8.96s\tremaining: 22.8s\n",
      "113:\tlearn: 0.8506249\ttest: 0.8506249\ttest1: 0.8457365\tbest: 0.8471696 (106)\ttotal: 9.01s\tremaining: 22.6s\n",
      "114:\tlearn: 0.8506571\ttest: 0.8506571\ttest1: 0.8457365\tbest: 0.8471696 (106)\ttotal: 9.07s\tremaining: 22.5s\n",
      "115:\tlearn: 0.8506249\ttest: 0.8506249\ttest1: 0.8457365\tbest: 0.8471696 (106)\ttotal: 9.13s\tremaining: 22.4s\n",
      "116:\tlearn: 0.8509784\ttest: 0.8509784\ttest1: 0.8464531\tbest: 0.8471696 (106)\ttotal: 9.19s\tremaining: 22.2s\n",
      "117:\tlearn: 0.8510105\ttest: 0.8510105\ttest1: 0.8466578\tbest: 0.8471696 (106)\ttotal: 9.25s\tremaining: 22.1s\n",
      "118:\tlearn: 0.8512354\ttest: 0.8512354\ttest1: 0.8467602\tbest: 0.8471696 (106)\ttotal: 9.31s\tremaining: 22s\n",
      "119:\tlearn: 0.8514603\ttest: 0.8514603\ttest1: 0.8466578\tbest: 0.8471696 (106)\ttotal: 9.38s\tremaining: 21.9s\n",
      "120:\tlearn: 0.8516531\ttest: 0.8516531\ttest1: 0.8465554\tbest: 0.8471696 (106)\ttotal: 9.44s\tremaining: 21.8s\n",
      "121:\tlearn: 0.8517816\ttest: 0.8517816\ttest1: 0.8464531\tbest: 0.8471696 (106)\ttotal: 9.49s\tremaining: 21.6s\n",
      "122:\tlearn: 0.8517816\ttest: 0.8517816\ttest1: 0.8466578\tbest: 0.8471696 (106)\ttotal: 9.6s\tremaining: 21.6s\n",
      "123:\tlearn: 0.8518780\ttest: 0.8518780\ttest1: 0.8466578\tbest: 0.8471696 (106)\ttotal: 9.75s\tremaining: 21.7s\n",
      "124:\tlearn: 0.8520708\ttest: 0.8520708\ttest1: 0.8470673\tbest: 0.8471696 (106)\ttotal: 9.81s\tremaining: 21.6s\n",
      "125:\tlearn: 0.8518459\ttest: 0.8518459\ttest1: 0.8472720\tbest: 0.8472720 (125)\ttotal: 9.88s\tremaining: 21.5s\n",
      "126:\tlearn: 0.8516210\ttest: 0.8516210\ttest1: 0.8476814\tbest: 0.8476814 (126)\ttotal: 9.94s\tremaining: 21.4s\n",
      "127:\tlearn: 0.8516531\ttest: 0.8516531\ttest1: 0.8478862\tbest: 0.8478862 (127)\ttotal: 10.1s\tremaining: 21.4s\n",
      "128:\tlearn: 0.8519744\ttest: 0.8519744\ttest1: 0.8474767\tbest: 0.8478862 (127)\ttotal: 10.3s\tremaining: 21.5s\n",
      "129:\tlearn: 0.8520708\ttest: 0.8520708\ttest1: 0.8475791\tbest: 0.8478862 (127)\ttotal: 10.4s\tremaining: 21.5s\n",
      "130:\tlearn: 0.8520066\ttest: 0.8520066\ttest1: 0.8475791\tbest: 0.8478862 (127)\ttotal: 10.4s\tremaining: 21.4s\n",
      "131:\tlearn: 0.8516852\ttest: 0.8516852\ttest1: 0.8478862\tbest: 0.8478862 (127)\ttotal: 10.5s\tremaining: 21.3s\n",
      "132:\tlearn: 0.8517174\ttest: 0.8517174\ttest1: 0.8479885\tbest: 0.8479885 (132)\ttotal: 10.5s\tremaining: 21.2s\n",
      "133:\tlearn: 0.8517495\ttest: 0.8517495\ttest1: 0.8476814\tbest: 0.8479885 (132)\ttotal: 10.6s\tremaining: 21.1s\n",
      "134:\tlearn: 0.8517816\ttest: 0.8517816\ttest1: 0.8479885\tbest: 0.8479885 (132)\ttotal: 10.7s\tremaining: 21s\n",
      "135:\tlearn: 0.8518138\ttest: 0.8518138\ttest1: 0.8482956\tbest: 0.8482956 (135)\ttotal: 10.7s\tremaining: 20.8s\n",
      "136:\tlearn: 0.8518780\ttest: 0.8518780\ttest1: 0.8482956\tbest: 0.8482956 (135)\ttotal: 10.8s\tremaining: 20.7s\n",
      "137:\tlearn: 0.8517495\ttest: 0.8517495\ttest1: 0.8482956\tbest: 0.8482956 (135)\ttotal: 11s\tremaining: 20.8s\n",
      "138:\tlearn: 0.8523600\ttest: 0.8523600\ttest1: 0.8483980\tbest: 0.8483980 (138)\ttotal: 11.1s\tremaining: 20.9s\n",
      "139:\tlearn: 0.8525528\ttest: 0.8525528\ttest1: 0.8482956\tbest: 0.8483980 (138)\ttotal: 11.2s\tremaining: 20.8s\n",
      "140:\tlearn: 0.8529062\ttest: 0.8529062\ttest1: 0.8487051\tbest: 0.8487051 (140)\ttotal: 11.2s\tremaining: 20.7s\n",
      "141:\tlearn: 0.8530347\ttest: 0.8530347\ttest1: 0.8485004\tbest: 0.8487051 (140)\ttotal: 11.3s\tremaining: 20.5s\n",
      "142:\tlearn: 0.8529383\ttest: 0.8529383\ttest1: 0.8487051\tbest: 0.8487051 (140)\ttotal: 11.4s\tremaining: 20.4s\n",
      "143:\tlearn: 0.8531311\ttest: 0.8531311\ttest1: 0.8492169\tbest: 0.8492169 (143)\ttotal: 11.4s\tremaining: 20.4s\n",
      "144:\tlearn: 0.8530990\ttest: 0.8530990\ttest1: 0.8493193\tbest: 0.8493193 (144)\ttotal: 11.6s\tremaining: 20.3s\n",
      "145:\tlearn: 0.8530026\ttest: 0.8530026\ttest1: 0.8494216\tbest: 0.8494216 (145)\ttotal: 11.6s\tremaining: 20.2s\n",
      "146:\tlearn: 0.8530347\ttest: 0.8530347\ttest1: 0.8494216\tbest: 0.8494216 (145)\ttotal: 11.7s\tremaining: 20.2s\n",
      "147:\tlearn: 0.8530990\ttest: 0.8530990\ttest1: 0.8495240\tbest: 0.8495240 (147)\ttotal: 11.8s\tremaining: 20.1s\n",
      "148:\tlearn: 0.8532596\ttest: 0.8532596\ttest1: 0.8495240\tbest: 0.8495240 (147)\ttotal: 11.9s\tremaining: 20s\n",
      "149:\tlearn: 0.8528741\ttest: 0.8528741\ttest1: 0.8491145\tbest: 0.8495240 (147)\ttotal: 12s\tremaining: 20s\n",
      "150:\tlearn: 0.8531633\ttest: 0.8531633\ttest1: 0.8489098\tbest: 0.8495240 (147)\ttotal: 12s\tremaining: 19.8s\n",
      "151:\tlearn: 0.8530347\ttest: 0.8530347\ttest1: 0.8488075\tbest: 0.8495240 (147)\ttotal: 12.1s\tremaining: 19.7s\n",
      "152:\tlearn: 0.8531311\ttest: 0.8531311\ttest1: 0.8488075\tbest: 0.8495240 (147)\ttotal: 12.1s\tremaining: 19.6s\n",
      "153:\tlearn: 0.8533882\ttest: 0.8533882\ttest1: 0.8488075\tbest: 0.8495240 (147)\ttotal: 12.3s\tremaining: 19.6s\n",
      "154:\tlearn: 0.8535488\ttest: 0.8535488\ttest1: 0.8488075\tbest: 0.8495240 (147)\ttotal: 12.4s\tremaining: 19.6s\n",
      "155:\tlearn: 0.8535810\ttest: 0.8535810\ttest1: 0.8488075\tbest: 0.8495240 (147)\ttotal: 12.5s\tremaining: 19.5s\n",
      "156:\tlearn: 0.8541914\ttest: 0.8541914\ttest1: 0.8491145\tbest: 0.8495240 (147)\ttotal: 12.5s\tremaining: 19.4s\n",
      "157:\tlearn: 0.8543521\ttest: 0.8543521\ttest1: 0.8493193\tbest: 0.8495240 (147)\ttotal: 12.6s\tremaining: 19.3s\n",
      "158:\tlearn: 0.8541593\ttest: 0.8541593\ttest1: 0.8497287\tbest: 0.8497287 (158)\ttotal: 12.6s\tremaining: 19.2s\n",
      "159:\tlearn: 0.8544806\ttest: 0.8544806\ttest1: 0.8494216\tbest: 0.8497287 (158)\ttotal: 12.7s\tremaining: 19.1s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "160:\tlearn: 0.8545127\ttest: 0.8545127\ttest1: 0.8494216\tbest: 0.8497287 (158)\ttotal: 12.8s\tremaining: 19s\n",
      "161:\tlearn: 0.8545449\ttest: 0.8545449\ttest1: 0.8494216\tbest: 0.8497287 (158)\ttotal: 12.9s\tremaining: 18.9s\n",
      "162:\tlearn: 0.8548983\ttest: 0.8548983\ttest1: 0.8495240\tbest: 0.8497287 (158)\ttotal: 13s\tremaining: 18.9s\n",
      "163:\tlearn: 0.8549304\ttest: 0.8549304\ttest1: 0.8497287\tbest: 0.8497287 (158)\ttotal: 13.1s\tremaining: 18.8s\n",
      "164:\tlearn: 0.8549947\ttest: 0.8549947\ttest1: 0.8495240\tbest: 0.8497287 (158)\ttotal: 13.2s\tremaining: 18.8s\n",
      "165:\tlearn: 0.8549947\ttest: 0.8549947\ttest1: 0.8495240\tbest: 0.8497287 (158)\ttotal: 13.2s\tremaining: 18.7s\n",
      "166:\tlearn: 0.8557658\ttest: 0.8557658\ttest1: 0.8494216\tbest: 0.8497287 (158)\ttotal: 13.3s\tremaining: 18.6s\n",
      "167:\tlearn: 0.8559265\ttest: 0.8559265\ttest1: 0.8498311\tbest: 0.8498311 (167)\ttotal: 13.4s\tremaining: 18.5s\n",
      "168:\tlearn: 0.8558944\ttest: 0.8558944\ttest1: 0.8498311\tbest: 0.8498311 (167)\ttotal: 13.4s\tremaining: 18.3s\n",
      "169:\tlearn: 0.8559265\ttest: 0.8559265\ttest1: 0.8498311\tbest: 0.8498311 (167)\ttotal: 13.5s\tremaining: 18.3s\n",
      "170:\tlearn: 0.8567619\ttest: 0.8567619\ttest1: 0.8497287\tbest: 0.8498311 (167)\ttotal: 13.7s\tremaining: 18.3s\n",
      "171:\tlearn: 0.8568904\ttest: 0.8568904\ttest1: 0.8498311\tbest: 0.8498311 (167)\ttotal: 13.8s\tremaining: 18.2s\n",
      "172:\tlearn: 0.8569225\ttest: 0.8569225\ttest1: 0.8499335\tbest: 0.8499335 (172)\ttotal: 13.8s\tremaining: 18.1s\n",
      "173:\tlearn: 0.8566012\ttest: 0.8566012\ttest1: 0.8496264\tbest: 0.8499335 (172)\ttotal: 13.9s\tremaining: 18.1s\n",
      "174:\tlearn: 0.8566334\ttest: 0.8566334\ttest1: 0.8495240\tbest: 0.8499335 (172)\ttotal: 14s\tremaining: 18s\n",
      "175:\tlearn: 0.8566012\ttest: 0.8566012\ttest1: 0.8496264\tbest: 0.8499335 (172)\ttotal: 14.1s\tremaining: 17.9s\n",
      "176:\tlearn: 0.8568583\ttest: 0.8568583\ttest1: 0.8498311\tbest: 0.8499335 (172)\ttotal: 14.1s\tremaining: 17.8s\n",
      "177:\tlearn: 0.8567619\ttest: 0.8567619\ttest1: 0.8497287\tbest: 0.8499335 (172)\ttotal: 14.2s\tremaining: 17.7s\n",
      "178:\tlearn: 0.8567619\ttest: 0.8567619\ttest1: 0.8497287\tbest: 0.8499335 (172)\ttotal: 14.3s\tremaining: 17.6s\n",
      "179:\tlearn: 0.8571796\ttest: 0.8571796\ttest1: 0.8497287\tbest: 0.8499335 (172)\ttotal: 14.3s\tremaining: 17.5s\n",
      "180:\tlearn: 0.8571474\ttest: 0.8571474\ttest1: 0.8499335\tbest: 0.8499335 (172)\ttotal: 14.4s\tremaining: 17.4s\n",
      "181:\tlearn: 0.8572117\ttest: 0.8572117\ttest1: 0.8499335\tbest: 0.8499335 (172)\ttotal: 14.5s\tremaining: 17.3s\n",
      "182:\tlearn: 0.8573081\ttest: 0.8573081\ttest1: 0.8498311\tbest: 0.8499335 (172)\ttotal: 14.5s\tremaining: 17.2s\n",
      "183:\tlearn: 0.8573724\ttest: 0.8573724\ttest1: 0.8498311\tbest: 0.8499335 (172)\ttotal: 14.6s\tremaining: 17.1s\n",
      "184:\tlearn: 0.8572760\ttest: 0.8572760\ttest1: 0.8498311\tbest: 0.8499335 (172)\ttotal: 14.6s\tremaining: 17s\n",
      "185:\tlearn: 0.8573402\ttest: 0.8573402\ttest1: 0.8497287\tbest: 0.8499335 (172)\ttotal: 14.7s\tremaining: 16.9s\n",
      "186:\tlearn: 0.8573402\ttest: 0.8573402\ttest1: 0.8498311\tbest: 0.8499335 (172)\ttotal: 14.7s\tremaining: 16.8s\n",
      "187:\tlearn: 0.8574366\ttest: 0.8574366\ttest1: 0.8499335\tbest: 0.8499335 (172)\ttotal: 14.8s\tremaining: 16.7s\n",
      "188:\tlearn: 0.8575009\ttest: 0.8575009\ttest1: 0.8499335\tbest: 0.8499335 (172)\ttotal: 15s\tremaining: 16.7s\n",
      "189:\tlearn: 0.8574688\ttest: 0.8574688\ttest1: 0.8501382\tbest: 0.8501382 (189)\ttotal: 15.1s\tremaining: 16.6s\n",
      "190:\tlearn: 0.8579507\ttest: 0.8579507\ttest1: 0.8496264\tbest: 0.8501382 (189)\ttotal: 15.1s\tremaining: 16.5s\n",
      "191:\tlearn: 0.8583041\ttest: 0.8583041\ttest1: 0.8497287\tbest: 0.8501382 (189)\ttotal: 15.2s\tremaining: 16.4s\n",
      "192:\tlearn: 0.8584005\ttest: 0.8584005\ttest1: 0.8496264\tbest: 0.8501382 (189)\ttotal: 15.2s\tremaining: 16.3s\n",
      "193:\tlearn: 0.8582720\ttest: 0.8582720\ttest1: 0.8497287\tbest: 0.8501382 (189)\ttotal: 15.3s\tremaining: 16.2s\n",
      "194:\tlearn: 0.8587540\ttest: 0.8587540\ttest1: 0.8497287\tbest: 0.8501382 (189)\ttotal: 15.3s\tremaining: 16.1s\n",
      "195:\tlearn: 0.8589468\ttest: 0.8589468\ttest1: 0.8496264\tbest: 0.8501382 (189)\ttotal: 15.4s\tremaining: 16s\n",
      "196:\tlearn: 0.8592038\ttest: 0.8592038\ttest1: 0.8496264\tbest: 0.8501382 (189)\ttotal: 15.5s\tremaining: 15.9s\n",
      "197:\tlearn: 0.8592359\ttest: 0.8592359\ttest1: 0.8495240\tbest: 0.8501382 (189)\ttotal: 15.5s\tremaining: 15.8s\n",
      "198:\tlearn: 0.8592681\ttest: 0.8592681\ttest1: 0.8493193\tbest: 0.8501382 (189)\ttotal: 15.6s\tremaining: 15.7s\n",
      "199:\tlearn: 0.8593645\ttest: 0.8593645\ttest1: 0.8494216\tbest: 0.8501382 (189)\ttotal: 15.6s\tremaining: 15.6s\n",
      "200:\tlearn: 0.8594287\ttest: 0.8594287\ttest1: 0.8496264\tbest: 0.8501382 (189)\ttotal: 15.7s\tremaining: 15.5s\n",
      "201:\tlearn: 0.8594608\ttest: 0.8594608\ttest1: 0.8496264\tbest: 0.8501382 (189)\ttotal: 15.8s\tremaining: 15.5s\n",
      "202:\tlearn: 0.8592359\ttest: 0.8592359\ttest1: 0.8494216\tbest: 0.8501382 (189)\ttotal: 15.8s\tremaining: 15.4s\n",
      "203:\tlearn: 0.8593323\ttest: 0.8593323\ttest1: 0.8497287\tbest: 0.8501382 (189)\ttotal: 15.9s\tremaining: 15.3s\n",
      "204:\tlearn: 0.8593323\ttest: 0.8593323\ttest1: 0.8501382\tbest: 0.8501382 (189)\ttotal: 15.9s\tremaining: 15.2s\n",
      "205:\tlearn: 0.8599107\ttest: 0.8599107\ttest1: 0.8500358\tbest: 0.8501382 (189)\ttotal: 16s\tremaining: 15.1s\n",
      "206:\tlearn: 0.8600071\ttest: 0.8600071\ttest1: 0.8500358\tbest: 0.8501382 (189)\ttotal: 16.1s\tremaining: 15s\n",
      "207:\tlearn: 0.8601356\ttest: 0.8601356\ttest1: 0.8501382\tbest: 0.8501382 (189)\ttotal: 16.2s\tremaining: 14.9s\n",
      "208:\tlearn: 0.8601035\ttest: 0.8601035\ttest1: 0.8502406\tbest: 0.8502406 (208)\ttotal: 16.3s\tremaining: 14.9s\n",
      "209:\tlearn: 0.8601035\ttest: 0.8601035\ttest1: 0.8502406\tbest: 0.8502406 (208)\ttotal: 16.4s\tremaining: 14.8s\n",
      "210:\tlearn: 0.8598464\ttest: 0.8598464\ttest1: 0.8502406\tbest: 0.8502406 (208)\ttotal: 16.5s\tremaining: 14.7s\n",
      "211:\tlearn: 0.8598464\ttest: 0.8598464\ttest1: 0.8502406\tbest: 0.8502406 (208)\ttotal: 16.5s\tremaining: 14.6s\n",
      "212:\tlearn: 0.8596858\ttest: 0.8596858\ttest1: 0.8506500\tbest: 0.8506500 (212)\ttotal: 16.6s\tremaining: 14.6s\n",
      "213:\tlearn: 0.8595894\ttest: 0.8595894\ttest1: 0.8506500\tbest: 0.8506500 (212)\ttotal: 16.6s\tremaining: 14.5s\n",
      "214:\tlearn: 0.8595251\ttest: 0.8595251\ttest1: 0.8506500\tbest: 0.8506500 (212)\ttotal: 16.7s\tremaining: 14.4s\n",
      "215:\tlearn: 0.8595894\ttest: 0.8595894\ttest1: 0.8507524\tbest: 0.8507524 (215)\ttotal: 16.7s\tremaining: 14.3s\n",
      "216:\tlearn: 0.8596215\ttest: 0.8596215\ttest1: 0.8504453\tbest: 0.8507524 (215)\ttotal: 16.8s\tremaining: 14.2s\n",
      "217:\tlearn: 0.8596858\ttest: 0.8596858\ttest1: 0.8506500\tbest: 0.8507524 (215)\ttotal: 16.9s\tremaining: 14.1s\n",
      "218:\tlearn: 0.8599428\ttest: 0.8599428\ttest1: 0.8506500\tbest: 0.8507524 (215)\ttotal: 16.9s\tremaining: 14s\n",
      "219:\tlearn: 0.8598143\ttest: 0.8598143\ttest1: 0.8504453\tbest: 0.8507524 (215)\ttotal: 17s\tremaining: 13.9s\n",
      "220:\tlearn: 0.8599749\ttest: 0.8599749\ttest1: 0.8504453\tbest: 0.8507524 (215)\ttotal: 17.1s\tremaining: 13.9s\n",
      "221:\tlearn: 0.8600713\ttest: 0.8600713\ttest1: 0.8506500\tbest: 0.8507524 (215)\ttotal: 17.2s\tremaining: 13.8s\n",
      "222:\tlearn: 0.8602641\ttest: 0.8602641\ttest1: 0.8505477\tbest: 0.8507524 (215)\ttotal: 17.3s\tremaining: 13.7s\n",
      "223:\tlearn: 0.8604890\ttest: 0.8604890\ttest1: 0.8506500\tbest: 0.8507524 (215)\ttotal: 17.3s\tremaining: 13.6s\n",
      "224:\tlearn: 0.8605212\ttest: 0.8605212\ttest1: 0.8505477\tbest: 0.8507524 (215)\ttotal: 17.4s\tremaining: 13.5s\n",
      "225:\tlearn: 0.8609067\ttest: 0.8609067\ttest1: 0.8507524\tbest: 0.8507524 (215)\ttotal: 17.5s\tremaining: 13.4s\n",
      "226:\tlearn: 0.8610352\ttest: 0.8610352\ttest1: 0.8507524\tbest: 0.8507524 (215)\ttotal: 17.5s\tremaining: 13.4s\n",
      "227:\tlearn: 0.8609389\ttest: 0.8609389\ttest1: 0.8505477\tbest: 0.8507524 (215)\ttotal: 17.6s\tremaining: 13.3s\n",
      "228:\tlearn: 0.8611316\ttest: 0.8611316\ttest1: 0.8506500\tbest: 0.8507524 (215)\ttotal: 17.6s\tremaining: 13.2s\n",
      "229:\tlearn: 0.8612280\ttest: 0.8612280\ttest1: 0.8505477\tbest: 0.8507524 (215)\ttotal: 17.7s\tremaining: 13.1s\n",
      "230:\tlearn: 0.8609389\ttest: 0.8609389\ttest1: 0.8511618\tbest: 0.8511618 (230)\ttotal: 17.8s\tremaining: 13s\n",
      "231:\tlearn: 0.8612280\ttest: 0.8612280\ttest1: 0.8511618\tbest: 0.8511618 (230)\ttotal: 17.8s\tremaining: 12.9s\n",
      "232:\tlearn: 0.8612280\ttest: 0.8612280\ttest1: 0.8510595\tbest: 0.8511618 (230)\ttotal: 17.9s\tremaining: 12.8s\n",
      "233:\tlearn: 0.8614529\ttest: 0.8614529\ttest1: 0.8507524\tbest: 0.8511618 (230)\ttotal: 17.9s\tremaining: 12.7s\n",
      "234:\tlearn: 0.8617421\ttest: 0.8617421\ttest1: 0.8509571\tbest: 0.8511618 (230)\ttotal: 18s\tremaining: 12.6s\n",
      "235:\tlearn: 0.8614851\ttest: 0.8614851\ttest1: 0.8508547\tbest: 0.8511618 (230)\ttotal: 18s\tremaining: 12.5s\n",
      "236:\tlearn: 0.8617421\ttest: 0.8617421\ttest1: 0.8505477\tbest: 0.8511618 (230)\ttotal: 18.1s\tremaining: 12.5s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "237:\tlearn: 0.8617421\ttest: 0.8617421\ttest1: 0.8505477\tbest: 0.8511618 (230)\ttotal: 18.2s\tremaining: 12.4s\n",
      "238:\tlearn: 0.8619028\ttest: 0.8619028\ttest1: 0.8502406\tbest: 0.8511618 (230)\ttotal: 18.3s\tremaining: 12.3s\n",
      "239:\tlearn: 0.8615815\ttest: 0.8615815\ttest1: 0.8504453\tbest: 0.8511618 (230)\ttotal: 18.4s\tremaining: 12.2s\n",
      "240:\tlearn: 0.8614529\ttest: 0.8614529\ttest1: 0.8503429\tbest: 0.8511618 (230)\ttotal: 18.5s\tremaining: 12.2s\n",
      "241:\tlearn: 0.8614529\ttest: 0.8614529\ttest1: 0.8505477\tbest: 0.8511618 (230)\ttotal: 18.6s\tremaining: 12.2s\n",
      "242:\tlearn: 0.8614529\ttest: 0.8614529\ttest1: 0.8502406\tbest: 0.8511618 (230)\ttotal: 18.7s\tremaining: 12.1s\n",
      "243:\tlearn: 0.8616136\ttest: 0.8616136\ttest1: 0.8503429\tbest: 0.8511618 (230)\ttotal: 18.8s\tremaining: 12s\n",
      "244:\tlearn: 0.8618706\ttest: 0.8618706\ttest1: 0.8507524\tbest: 0.8511618 (230)\ttotal: 18.8s\tremaining: 11.9s\n",
      "245:\tlearn: 0.8620313\ttest: 0.8620313\ttest1: 0.8511618\tbest: 0.8511618 (230)\ttotal: 18.9s\tremaining: 11.8s\n",
      "246:\tlearn: 0.8620313\ttest: 0.8620313\ttest1: 0.8513666\tbest: 0.8513666 (246)\ttotal: 19s\tremaining: 11.8s\n",
      "247:\tlearn: 0.8620956\ttest: 0.8620956\ttest1: 0.8511618\tbest: 0.8513666 (246)\ttotal: 19.1s\tremaining: 11.7s\n",
      "248:\tlearn: 0.8617100\ttest: 0.8617100\ttest1: 0.8512642\tbest: 0.8513666 (246)\ttotal: 19.3s\tremaining: 11.7s\n",
      "249:\tlearn: 0.8619670\ttest: 0.8619670\ttest1: 0.8513666\tbest: 0.8513666 (246)\ttotal: 19.5s\tremaining: 11.7s\n",
      "250:\tlearn: 0.8618385\ttest: 0.8618385\ttest1: 0.8518784\tbest: 0.8518784 (250)\ttotal: 19.6s\tremaining: 11.6s\n",
      "251:\tlearn: 0.8619992\ttest: 0.8619992\ttest1: 0.8518784\tbest: 0.8518784 (250)\ttotal: 19.6s\tremaining: 11.5s\n",
      "252:\tlearn: 0.8619349\ttest: 0.8619349\ttest1: 0.8515713\tbest: 0.8518784 (250)\ttotal: 19.7s\tremaining: 11.4s\n",
      "253:\tlearn: 0.8621277\ttest: 0.8621277\ttest1: 0.8515713\tbest: 0.8518784 (250)\ttotal: 19.8s\tremaining: 11.4s\n",
      "254:\tlearn: 0.8624169\ttest: 0.8624169\ttest1: 0.8515713\tbest: 0.8518784 (250)\ttotal: 19.9s\tremaining: 11.3s\n",
      "255:\tlearn: 0.8624490\ttest: 0.8624490\ttest1: 0.8516737\tbest: 0.8518784 (250)\ttotal: 19.9s\tremaining: 11.2s\n",
      "256:\tlearn: 0.8624490\ttest: 0.8624490\ttest1: 0.8517760\tbest: 0.8518784 (250)\ttotal: 20s\tremaining: 11.1s\n",
      "257:\tlearn: 0.8624811\ttest: 0.8624811\ttest1: 0.8517760\tbest: 0.8518784 (250)\ttotal: 20s\tremaining: 11s\n",
      "258:\tlearn: 0.8624490\ttest: 0.8624490\ttest1: 0.8517760\tbest: 0.8518784 (250)\ttotal: 20.1s\tremaining: 10.9s\n",
      "259:\tlearn: 0.8623526\ttest: 0.8623526\ttest1: 0.8515713\tbest: 0.8518784 (250)\ttotal: 20.3s\tremaining: 10.9s\n",
      "260:\tlearn: 0.8623847\ttest: 0.8623847\ttest1: 0.8513666\tbest: 0.8518784 (250)\ttotal: 20.6s\tremaining: 11s\n",
      "261:\tlearn: 0.8624169\ttest: 0.8624169\ttest1: 0.8512642\tbest: 0.8518784 (250)\ttotal: 20.7s\tremaining: 10.9s\n",
      "262:\tlearn: 0.8625133\ttest: 0.8625133\ttest1: 0.8518784\tbest: 0.8518784 (250)\ttotal: 20.8s\tremaining: 10.9s\n",
      "263:\tlearn: 0.8626418\ttest: 0.8626418\ttest1: 0.8518784\tbest: 0.8518784 (250)\ttotal: 20.9s\tremaining: 10.8s\n",
      "264:\tlearn: 0.8628988\ttest: 0.8628988\ttest1: 0.8516737\tbest: 0.8518784 (250)\ttotal: 21s\tremaining: 10.7s\n",
      "265:\tlearn: 0.8630273\ttest: 0.8630273\ttest1: 0.8516737\tbest: 0.8518784 (250)\ttotal: 21.1s\tremaining: 10.6s\n",
      "266:\tlearn: 0.8629952\ttest: 0.8629952\ttest1: 0.8516737\tbest: 0.8518784 (250)\ttotal: 21.1s\tremaining: 10.5s\n",
      "267:\tlearn: 0.8629952\ttest: 0.8629952\ttest1: 0.8516737\tbest: 0.8518784 (250)\ttotal: 21.2s\tremaining: 10.4s\n",
      "268:\tlearn: 0.8630595\ttest: 0.8630595\ttest1: 0.8516737\tbest: 0.8518784 (250)\ttotal: 21.3s\tremaining: 10.4s\n",
      "269:\tlearn: 0.8631237\ttest: 0.8631237\ttest1: 0.8517760\tbest: 0.8518784 (250)\ttotal: 21.3s\tremaining: 10.3s\n",
      "270:\tlearn: 0.8631559\ttest: 0.8631559\ttest1: 0.8517760\tbest: 0.8518784 (250)\ttotal: 21.4s\tremaining: 10.2s\n",
      "271:\tlearn: 0.8631559\ttest: 0.8631559\ttest1: 0.8516737\tbest: 0.8518784 (250)\ttotal: 21.5s\tremaining: 10.1s\n",
      "272:\tlearn: 0.8631237\ttest: 0.8631237\ttest1: 0.8516737\tbest: 0.8518784 (250)\ttotal: 21.6s\tremaining: 10s\n",
      "273:\tlearn: 0.8631559\ttest: 0.8631559\ttest1: 0.8515713\tbest: 0.8518784 (250)\ttotal: 21.6s\tremaining: 9.95s\n",
      "274:\tlearn: 0.8628667\ttest: 0.8628667\ttest1: 0.8517760\tbest: 0.8518784 (250)\ttotal: 21.7s\tremaining: 9.86s\n",
      "275:\tlearn: 0.8629310\ttest: 0.8629310\ttest1: 0.8517760\tbest: 0.8518784 (250)\ttotal: 21.8s\tremaining: 9.79s\n",
      "276:\tlearn: 0.8629952\ttest: 0.8629952\ttest1: 0.8515713\tbest: 0.8518784 (250)\ttotal: 21.8s\tremaining: 9.7s\n",
      "277:\tlearn: 0.8629952\ttest: 0.8629952\ttest1: 0.8518784\tbest: 0.8518784 (250)\ttotal: 21.9s\tremaining: 9.62s\n",
      "278:\tlearn: 0.8630273\ttest: 0.8630273\ttest1: 0.8518784\tbest: 0.8518784 (250)\ttotal: 22s\tremaining: 9.54s\n",
      "279:\tlearn: 0.8630273\ttest: 0.8630273\ttest1: 0.8518784\tbest: 0.8518784 (250)\ttotal: 22s\tremaining: 9.45s\n",
      "280:\tlearn: 0.8629952\ttest: 0.8629952\ttest1: 0.8519808\tbest: 0.8519808 (280)\ttotal: 22.1s\tremaining: 9.37s\n",
      "281:\tlearn: 0.8631559\ttest: 0.8631559\ttest1: 0.8517760\tbest: 0.8519808 (280)\ttotal: 22.2s\tremaining: 9.29s\n",
      "282:\tlearn: 0.8631559\ttest: 0.8631559\ttest1: 0.8513666\tbest: 0.8519808 (280)\ttotal: 22.3s\tremaining: 9.21s\n",
      "283:\tlearn: 0.8632844\ttest: 0.8632844\ttest1: 0.8513666\tbest: 0.8519808 (280)\ttotal: 22.4s\tremaining: 9.14s\n",
      "284:\tlearn: 0.8632844\ttest: 0.8632844\ttest1: 0.8511618\tbest: 0.8519808 (280)\ttotal: 22.5s\tremaining: 9.08s\n",
      "285:\tlearn: 0.8632844\ttest: 0.8632844\ttest1: 0.8512642\tbest: 0.8519808 (280)\ttotal: 22.6s\tremaining: 9.02s\n",
      "286:\tlearn: 0.8631880\ttest: 0.8631880\ttest1: 0.8512642\tbest: 0.8519808 (280)\ttotal: 22.7s\tremaining: 8.94s\n",
      "287:\tlearn: 0.8633486\ttest: 0.8633486\ttest1: 0.8512642\tbest: 0.8519808 (280)\ttotal: 22.8s\tremaining: 8.87s\n",
      "288:\tlearn: 0.8635093\ttest: 0.8635093\ttest1: 0.8515713\tbest: 0.8519808 (280)\ttotal: 22.9s\tremaining: 8.79s\n",
      "289:\tlearn: 0.8635093\ttest: 0.8635093\ttest1: 0.8516737\tbest: 0.8519808 (280)\ttotal: 23s\tremaining: 8.72s\n",
      "290:\tlearn: 0.8640877\ttest: 0.8640877\ttest1: 0.8515713\tbest: 0.8519808 (280)\ttotal: 23.1s\tremaining: 8.65s\n",
      "291:\tlearn: 0.8643447\ttest: 0.8643447\ttest1: 0.8515713\tbest: 0.8519808 (280)\ttotal: 23.2s\tremaining: 8.58s\n",
      "292:\tlearn: 0.8644411\ttest: 0.8644411\ttest1: 0.8516737\tbest: 0.8519808 (280)\ttotal: 23.3s\tremaining: 8.51s\n",
      "293:\tlearn: 0.8644411\ttest: 0.8644411\ttest1: 0.8516737\tbest: 0.8519808 (280)\ttotal: 23.4s\tremaining: 8.43s\n",
      "294:\tlearn: 0.8646339\ttest: 0.8646339\ttest1: 0.8517760\tbest: 0.8519808 (280)\ttotal: 23.5s\tremaining: 8.35s\n",
      "295:\tlearn: 0.8646660\ttest: 0.8646660\ttest1: 0.8520831\tbest: 0.8520831 (295)\ttotal: 23.6s\tremaining: 8.28s\n",
      "296:\tlearn: 0.8646981\ttest: 0.8646981\ttest1: 0.8519808\tbest: 0.8520831 (295)\ttotal: 23.6s\tremaining: 8.19s\n",
      "297:\tlearn: 0.8646017\ttest: 0.8646017\ttest1: 0.8518784\tbest: 0.8520831 (295)\ttotal: 23.7s\tremaining: 8.13s\n",
      "298:\tlearn: 0.8645375\ttest: 0.8645375\ttest1: 0.8521855\tbest: 0.8521855 (298)\ttotal: 23.9s\tremaining: 8.08s\n",
      "299:\tlearn: 0.8647303\ttest: 0.8647303\ttest1: 0.8521855\tbest: 0.8521855 (298)\ttotal: 24.1s\tremaining: 8.02s\n",
      "300:\tlearn: 0.8647945\ttest: 0.8647945\ttest1: 0.8520831\tbest: 0.8521855 (298)\ttotal: 24.2s\tremaining: 7.95s\n",
      "301:\tlearn: 0.8648267\ttest: 0.8648267\ttest1: 0.8519808\tbest: 0.8521855 (298)\ttotal: 24.3s\tremaining: 7.87s\n",
      "302:\tlearn: 0.8647945\ttest: 0.8647945\ttest1: 0.8518784\tbest: 0.8521855 (298)\ttotal: 24.3s\tremaining: 7.79s\n",
      "303:\tlearn: 0.8648909\ttest: 0.8648909\ttest1: 0.8521855\tbest: 0.8521855 (298)\ttotal: 24.4s\tremaining: 7.71s\n",
      "304:\tlearn: 0.8648909\ttest: 0.8648909\ttest1: 0.8521855\tbest: 0.8521855 (298)\ttotal: 24.5s\tremaining: 7.62s\n",
      "305:\tlearn: 0.8646981\ttest: 0.8646981\ttest1: 0.8521855\tbest: 0.8521855 (298)\ttotal: 24.5s\tremaining: 7.53s\n",
      "306:\tlearn: 0.8648909\ttest: 0.8648909\ttest1: 0.8519808\tbest: 0.8521855 (298)\ttotal: 24.6s\tremaining: 7.45s\n",
      "307:\tlearn: 0.8646339\ttest: 0.8646339\ttest1: 0.8520831\tbest: 0.8521855 (298)\ttotal: 24.7s\tremaining: 7.37s\n",
      "308:\tlearn: 0.8646339\ttest: 0.8646339\ttest1: 0.8520831\tbest: 0.8521855 (298)\ttotal: 24.7s\tremaining: 7.29s\n",
      "309:\tlearn: 0.8649230\ttest: 0.8649230\ttest1: 0.8520831\tbest: 0.8521855 (298)\ttotal: 24.8s\tremaining: 7.2s\n",
      "310:\tlearn: 0.8649230\ttest: 0.8649230\ttest1: 0.8520831\tbest: 0.8521855 (298)\ttotal: 24.9s\tremaining: 7.12s\n",
      "311:\tlearn: 0.8649230\ttest: 0.8649230\ttest1: 0.8521855\tbest: 0.8521855 (298)\ttotal: 25s\tremaining: 7.04s\n",
      "312:\tlearn: 0.8649230\ttest: 0.8649230\ttest1: 0.8521855\tbest: 0.8521855 (298)\ttotal: 25.1s\tremaining: 6.97s\n",
      "313:\tlearn: 0.8649552\ttest: 0.8649552\ttest1: 0.8521855\tbest: 0.8521855 (298)\ttotal: 25.2s\tremaining: 6.89s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "314:\tlearn: 0.8650837\ttest: 0.8650837\ttest1: 0.8521855\tbest: 0.8521855 (298)\ttotal: 25.3s\tremaining: 6.84s\n",
      "315:\tlearn: 0.8649552\ttest: 0.8649552\ttest1: 0.8526973\tbest: 0.8526973 (315)\ttotal: 25.4s\tremaining: 6.75s\n",
      "316:\tlearn: 0.8649552\ttest: 0.8649552\ttest1: 0.8527997\tbest: 0.8527997 (316)\ttotal: 25.5s\tremaining: 6.67s\n",
      "317:\tlearn: 0.8649552\ttest: 0.8649552\ttest1: 0.8527997\tbest: 0.8527997 (316)\ttotal: 25.6s\tremaining: 6.59s\n",
      "318:\tlearn: 0.8651801\ttest: 0.8651801\ttest1: 0.8526973\tbest: 0.8527997 (316)\ttotal: 25.6s\tremaining: 6.51s\n",
      "319:\tlearn: 0.8651801\ttest: 0.8651801\ttest1: 0.8526973\tbest: 0.8527997 (316)\ttotal: 25.7s\tremaining: 6.43s\n",
      "320:\tlearn: 0.8654693\ttest: 0.8654693\ttest1: 0.8527997\tbest: 0.8527997 (316)\ttotal: 25.8s\tremaining: 6.34s\n",
      "321:\tlearn: 0.8654371\ttest: 0.8654371\ttest1: 0.8529020\tbest: 0.8529020 (321)\ttotal: 25.8s\tremaining: 6.26s\n",
      "322:\tlearn: 0.8653729\ttest: 0.8653729\ttest1: 0.8529020\tbest: 0.8529020 (321)\ttotal: 25.9s\tremaining: 6.17s\n",
      "323:\tlearn: 0.8654050\ttest: 0.8654050\ttest1: 0.8530044\tbest: 0.8530044 (323)\ttotal: 25.9s\tremaining: 6.08s\n",
      "324:\tlearn: 0.8651801\ttest: 0.8651801\ttest1: 0.8524926\tbest: 0.8530044 (323)\ttotal: 26s\tremaining: 6s\n",
      "325:\tlearn: 0.8651801\ttest: 0.8651801\ttest1: 0.8524926\tbest: 0.8530044 (323)\ttotal: 26.1s\tremaining: 5.92s\n",
      "326:\tlearn: 0.8649873\ttest: 0.8649873\ttest1: 0.8533115\tbest: 0.8533115 (326)\ttotal: 26.2s\tremaining: 5.84s\n",
      "327:\tlearn: 0.8649873\ttest: 0.8649873\ttest1: 0.8534139\tbest: 0.8534139 (327)\ttotal: 26.3s\tremaining: 5.78s\n",
      "328:\tlearn: 0.8649873\ttest: 0.8649873\ttest1: 0.8534139\tbest: 0.8534139 (327)\ttotal: 26.4s\tremaining: 5.7s\n",
      "329:\tlearn: 0.8652122\ttest: 0.8652122\ttest1: 0.8533115\tbest: 0.8534139 (327)\ttotal: 26.6s\tremaining: 5.63s\n",
      "330:\tlearn: 0.8652122\ttest: 0.8652122\ttest1: 0.8533115\tbest: 0.8534139 (327)\ttotal: 26.7s\tremaining: 5.57s\n",
      "331:\tlearn: 0.8652122\ttest: 0.8652122\ttest1: 0.8533115\tbest: 0.8534139 (327)\ttotal: 26.8s\tremaining: 5.49s\n",
      "332:\tlearn: 0.8652444\ttest: 0.8652444\ttest1: 0.8533115\tbest: 0.8534139 (327)\ttotal: 27.3s\tremaining: 5.49s\n",
      "333:\tlearn: 0.8650194\ttest: 0.8650194\ttest1: 0.8532091\tbest: 0.8534139 (327)\ttotal: 27.4s\tremaining: 5.42s\n",
      "334:\tlearn: 0.8650194\ttest: 0.8650194\ttest1: 0.8532091\tbest: 0.8534139 (327)\ttotal: 27.5s\tremaining: 5.34s\n",
      "335:\tlearn: 0.8652122\ttest: 0.8652122\ttest1: 0.8533115\tbest: 0.8534139 (327)\ttotal: 27.7s\tremaining: 5.27s\n",
      "336:\tlearn: 0.8651801\ttest: 0.8651801\ttest1: 0.8533115\tbest: 0.8534139 (327)\ttotal: 27.8s\tremaining: 5.19s\n",
      "337:\tlearn: 0.8655335\ttest: 0.8655335\ttest1: 0.8531068\tbest: 0.8534139 (327)\ttotal: 27.9s\tremaining: 5.12s\n",
      "338:\tlearn: 0.8657584\ttest: 0.8657584\ttest1: 0.8531068\tbest: 0.8534139 (327)\ttotal: 28s\tremaining: 5.04s\n",
      "339:\tlearn: 0.8657584\ttest: 0.8657584\ttest1: 0.8532091\tbest: 0.8534139 (327)\ttotal: 28.1s\tremaining: 4.96s\n",
      "340:\tlearn: 0.8657906\ttest: 0.8657906\ttest1: 0.8532091\tbest: 0.8534139 (327)\ttotal: 28.3s\tremaining: 4.89s\n",
      "341:\tlearn: 0.8657906\ttest: 0.8657906\ttest1: 0.8532091\tbest: 0.8534139 (327)\ttotal: 28.4s\tremaining: 4.81s\n",
      "342:\tlearn: 0.8653086\ttest: 0.8653086\ttest1: 0.8533115\tbest: 0.8534139 (327)\ttotal: 28.4s\tremaining: 4.73s\n",
      "343:\tlearn: 0.8653407\ttest: 0.8653407\ttest1: 0.8534139\tbest: 0.8534139 (327)\ttotal: 28.5s\tremaining: 4.64s\n",
      "344:\tlearn: 0.8653407\ttest: 0.8653407\ttest1: 0.8533115\tbest: 0.8534139 (327)\ttotal: 28.6s\tremaining: 4.56s\n",
      "345:\tlearn: 0.8653086\ttest: 0.8653086\ttest1: 0.8533115\tbest: 0.8534139 (327)\ttotal: 28.7s\tremaining: 4.48s\n",
      "346:\tlearn: 0.8653407\ttest: 0.8653407\ttest1: 0.8531068\tbest: 0.8534139 (327)\ttotal: 28.9s\tremaining: 4.42s\n",
      "347:\tlearn: 0.8653407\ttest: 0.8653407\ttest1: 0.8531068\tbest: 0.8534139 (327)\ttotal: 29.1s\tremaining: 4.35s\n",
      "348:\tlearn: 0.8653407\ttest: 0.8653407\ttest1: 0.8533115\tbest: 0.8534139 (327)\ttotal: 29.4s\tremaining: 4.29s\n",
      "349:\tlearn: 0.8656621\ttest: 0.8656621\ttest1: 0.8530044\tbest: 0.8534139 (327)\ttotal: 29.7s\tremaining: 4.24s\n",
      "350:\tlearn: 0.8656299\ttest: 0.8656299\ttest1: 0.8530044\tbest: 0.8534139 (327)\ttotal: 29.9s\tremaining: 4.18s\n",
      "351:\tlearn: 0.8658227\ttest: 0.8658227\ttest1: 0.8532091\tbest: 0.8534139 (327)\ttotal: 30.2s\tremaining: 4.11s\n",
      "352:\tlearn: 0.8658548\ttest: 0.8658548\ttest1: 0.8534139\tbest: 0.8534139 (327)\ttotal: 30.3s\tremaining: 4.04s\n",
      "353:\tlearn: 0.8658227\ttest: 0.8658227\ttest1: 0.8532091\tbest: 0.8534139 (327)\ttotal: 30.7s\tremaining: 3.99s\n",
      "354:\tlearn: 0.8657584\ttest: 0.8657584\ttest1: 0.8533115\tbest: 0.8534139 (327)\ttotal: 30.8s\tremaining: 3.9s\n",
      "355:\tlearn: 0.8657584\ttest: 0.8657584\ttest1: 0.8534139\tbest: 0.8534139 (327)\ttotal: 31s\tremaining: 3.84s\n",
      "356:\tlearn: 0.8658870\ttest: 0.8658870\ttest1: 0.8535162\tbest: 0.8535162 (356)\ttotal: 31.7s\tremaining: 3.82s\n",
      "357:\tlearn: 0.8659191\ttest: 0.8659191\ttest1: 0.8536186\tbest: 0.8536186 (357)\ttotal: 32s\tremaining: 3.76s\n",
      "358:\tlearn: 0.8661440\ttest: 0.8661440\ttest1: 0.8535162\tbest: 0.8536186 (357)\ttotal: 32.1s\tremaining: 3.67s\n",
      "359:\tlearn: 0.8662725\ttest: 0.8662725\ttest1: 0.8535162\tbest: 0.8536186 (357)\ttotal: 32.3s\tremaining: 3.59s\n",
      "360:\tlearn: 0.8669473\ttest: 0.8669473\ttest1: 0.8534139\tbest: 0.8536186 (357)\ttotal: 32.4s\tremaining: 3.5s\n",
      "361:\tlearn: 0.8669794\ttest: 0.8669794\ttest1: 0.8534139\tbest: 0.8536186 (357)\ttotal: 32.5s\tremaining: 3.41s\n",
      "362:\tlearn: 0.8664974\ttest: 0.8664974\ttest1: 0.8532091\tbest: 0.8536186 (357)\ttotal: 32.6s\tremaining: 3.32s\n",
      "363:\tlearn: 0.8666260\ttest: 0.8666260\ttest1: 0.8532091\tbest: 0.8536186 (357)\ttotal: 32.7s\tremaining: 3.23s\n",
      "364:\tlearn: 0.8665938\ttest: 0.8665938\ttest1: 0.8532091\tbest: 0.8536186 (357)\ttotal: 32.7s\tremaining: 3.14s\n",
      "365:\tlearn: 0.8665617\ttest: 0.8665617\ttest1: 0.8533115\tbest: 0.8536186 (357)\ttotal: 32.8s\tremaining: 3.05s\n",
      "366:\tlearn: 0.8666260\ttest: 0.8666260\ttest1: 0.8532091\tbest: 0.8536186 (357)\ttotal: 32.9s\tremaining: 2.96s\n",
      "367:\tlearn: 0.8666581\ttest: 0.8666581\ttest1: 0.8534139\tbest: 0.8536186 (357)\ttotal: 33s\tremaining: 2.87s\n",
      "368:\tlearn: 0.8666581\ttest: 0.8666581\ttest1: 0.8534139\tbest: 0.8536186 (357)\ttotal: 33.1s\tremaining: 2.78s\n",
      "369:\tlearn: 0.8666902\ttest: 0.8666902\ttest1: 0.8533115\tbest: 0.8536186 (357)\ttotal: 33.2s\tremaining: 2.69s\n",
      "370:\tlearn: 0.8666902\ttest: 0.8666902\ttest1: 0.8533115\tbest: 0.8536186 (357)\ttotal: 33.2s\tremaining: 2.6s\n",
      "371:\tlearn: 0.8666581\ttest: 0.8666581\ttest1: 0.8534139\tbest: 0.8536186 (357)\ttotal: 33.4s\tremaining: 2.51s\n",
      "372:\tlearn: 0.8665938\ttest: 0.8665938\ttest1: 0.8531068\tbest: 0.8536186 (357)\ttotal: 33.5s\tremaining: 2.42s\n",
      "373:\tlearn: 0.8664974\ttest: 0.8664974\ttest1: 0.8530044\tbest: 0.8536186 (357)\ttotal: 33.6s\tremaining: 2.33s\n",
      "374:\tlearn: 0.8665617\ttest: 0.8665617\ttest1: 0.8529020\tbest: 0.8536186 (357)\ttotal: 33.6s\tremaining: 2.24s\n",
      "375:\tlearn: 0.8665617\ttest: 0.8665617\ttest1: 0.8529020\tbest: 0.8536186 (357)\ttotal: 33.7s\tremaining: 2.15s\n",
      "376:\tlearn: 0.8665617\ttest: 0.8665617\ttest1: 0.8529020\tbest: 0.8536186 (357)\ttotal: 33.8s\tremaining: 2.06s\n",
      "377:\tlearn: 0.8665617\ttest: 0.8665617\ttest1: 0.8530044\tbest: 0.8536186 (357)\ttotal: 33.8s\tremaining: 1.97s\n",
      "378:\tlearn: 0.8665617\ttest: 0.8665617\ttest1: 0.8530044\tbest: 0.8536186 (357)\ttotal: 33.9s\tremaining: 1.88s\n",
      "379:\tlearn: 0.8667545\ttest: 0.8667545\ttest1: 0.8530044\tbest: 0.8536186 (357)\ttotal: 33.9s\tremaining: 1.79s\n",
      "380:\tlearn: 0.8665617\ttest: 0.8665617\ttest1: 0.8530044\tbest: 0.8536186 (357)\ttotal: 34s\tremaining: 1.7s\n",
      "381:\tlearn: 0.8668830\ttest: 0.8668830\ttest1: 0.8531068\tbest: 0.8536186 (357)\ttotal: 34.2s\tremaining: 1.61s\n",
      "382:\tlearn: 0.8668830\ttest: 0.8668830\ttest1: 0.8530044\tbest: 0.8536186 (357)\ttotal: 34.4s\tremaining: 1.52s\n",
      "383:\tlearn: 0.8672364\ttest: 0.8672364\ttest1: 0.8529020\tbest: 0.8536186 (357)\ttotal: 34.4s\tremaining: 1.43s\n",
      "384:\tlearn: 0.8672686\ttest: 0.8672686\ttest1: 0.8527997\tbest: 0.8536186 (357)\ttotal: 34.5s\tremaining: 1.34s\n",
      "385:\tlearn: 0.8672686\ttest: 0.8672686\ttest1: 0.8531068\tbest: 0.8536186 (357)\ttotal: 34.5s\tremaining: 1.25s\n",
      "386:\tlearn: 0.8673971\ttest: 0.8673971\ttest1: 0.8531068\tbest: 0.8536186 (357)\ttotal: 34.6s\tremaining: 1.16s\n",
      "387:\tlearn: 0.8674614\ttest: 0.8674614\ttest1: 0.8532091\tbest: 0.8536186 (357)\ttotal: 34.7s\tremaining: 1.07s\n",
      "388:\tlearn: 0.8677184\ttest: 0.8677184\ttest1: 0.8532091\tbest: 0.8536186 (357)\ttotal: 34.7s\tremaining: 982ms\n",
      "389:\tlearn: 0.8676541\ttest: 0.8676541\ttest1: 0.8532091\tbest: 0.8536186 (357)\ttotal: 34.8s\tremaining: 892ms\n",
      "390:\tlearn: 0.8676541\ttest: 0.8676541\ttest1: 0.8532091\tbest: 0.8536186 (357)\ttotal: 34.9s\tremaining: 802ms\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "391:\tlearn: 0.8676541\ttest: 0.8676541\ttest1: 0.8532091\tbest: 0.8536186 (357)\ttotal: 35.1s\tremaining: 716ms\n",
      "392:\tlearn: 0.8682968\ttest: 0.8682968\ttest1: 0.8532091\tbest: 0.8536186 (357)\ttotal: 35.2s\tremaining: 627ms\n",
      "393:\tlearn: 0.8683289\ttest: 0.8683289\ttest1: 0.8532091\tbest: 0.8536186 (357)\ttotal: 35.2s\tremaining: 537ms\n",
      "394:\tlearn: 0.8686823\ttest: 0.8686823\ttest1: 0.8533115\tbest: 0.8536186 (357)\ttotal: 35.3s\tremaining: 447ms\n",
      "395:\tlearn: 0.8686502\ttest: 0.8686502\ttest1: 0.8533115\tbest: 0.8536186 (357)\ttotal: 35.4s\tremaining: 357ms\n",
      "396:\tlearn: 0.8685859\ttest: 0.8685859\ttest1: 0.8533115\tbest: 0.8536186 (357)\ttotal: 35.4s\tremaining: 268ms\n",
      "397:\tlearn: 0.8685538\ttest: 0.8685538\ttest1: 0.8533115\tbest: 0.8536186 (357)\ttotal: 35.5s\tremaining: 178ms\n",
      "Stopped by overfitting detector  (40 iterations wait)\n",
      "\n",
      "bestTest = 0.8536185894\n",
      "bestIteration = 357\n",
      "\n",
      "Shrink model to first 358 iterations.\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<catboost.core.CatBoostClassifier at 0x1a2d1e9358>"
      ]
     },
     "execution_count": 151,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "catb_ros.fit(X_resampled,y_resampled,eval_set=[(X_resampled,y_resampled), (X_test,y_test)],early_stopping_rounds=40)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 152,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy on test data: 85.361859%\n",
      "Accuracy on training data: 86.591910%\n",
      "Area under the ROC curve : 0.843451\n"
     ]
    }
   ],
   "source": [
    "print('Accuracy on test data: %f%%' % (metrics.accuracy_score(y_test, catb_ros.predict(X_test))*100))\n",
    "print('Accuracy on training data: %f%%' % (metrics.accuracy_score(y_resampled, catb_ros.predict(X_resampled))*100))\n",
    "print('Area under the ROC curve : %f' % (metrics.roc_auc_score(y_test, catb_ros.predict(X_test))))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### SMOTE"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 144,
   "metadata": {},
   "outputs": [],
   "source": [
    "from imblearn.over_sampling import SMOTE"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 145,
   "metadata": {},
   "outputs": [],
   "source": [
    "smt = SMOTE(random_state=10,sampling_strategy=0.7)\n",
    "X_train_smt, y_train_smt = smt.fit_sample(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 146,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    17291\n",
       "1     5501\n",
       "Name: income, dtype: int64"
      ]
     },
     "execution_count": 146,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 147,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([17291, 12103])"
      ]
     },
     "execution_count": 147,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.bincount(y_train_smt)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 153,
   "metadata": {},
   "outputs": [],
   "source": [
    "catb_smote = CatBoostClassifier(learning_rate=0.1,iterations=400,reg_lambda=2,verbose=1,random_seed=10,eval_metric='Accuracy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 154,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0:\tlearn: 0.7904334\ttest: 0.7904334\ttest1: 0.8200430\tbest: 0.8200430 (0)\ttotal: 88.1ms\tremaining: 35.2s\n",
      "1:\tlearn: 0.7959788\ttest: 0.7959788\ttest1: 0.8298700\tbest: 0.8298700 (1)\ttotal: 155ms\tremaining: 30.8s\n",
      "2:\tlearn: 0.8081921\ttest: 0.8081921\ttest1: 0.8108302\tbest: 0.8298700 (1)\ttotal: 215ms\tremaining: 28.5s\n",
      "3:\tlearn: 0.8081581\ttest: 0.8081581\ttest1: 0.8088852\tbest: 0.8298700 (1)\ttotal: 281ms\tremaining: 27.8s\n",
      "4:\tlearn: 0.8137715\ttest: 0.8137715\ttest1: 0.8182004\tbest: 0.8298700 (1)\ttotal: 378ms\tremaining: 29.9s\n",
      "5:\tlearn: 0.8138055\ttest: 0.8138055\ttest1: 0.8178933\tbest: 0.8298700 (1)\ttotal: 439ms\tremaining: 28.8s\n",
      "6:\tlearn: 0.8161189\ttest: 0.8161189\ttest1: 0.8207595\tbest: 0.8298700 (1)\ttotal: 497ms\tremaining: 27.9s\n",
      "7:\tlearn: 0.8181942\ttest: 0.8181942\ttest1: 0.8215785\tbest: 0.8298700 (1)\ttotal: 557ms\tremaining: 27.3s\n",
      "8:\tlearn: 0.8228550\ttest: 0.8228550\ttest1: 0.8218856\tbest: 0.8298700 (1)\ttotal: 638ms\tremaining: 27.7s\n",
      "9:\tlearn: 0.8243179\ttest: 0.8243179\ttest1: 0.8227045\tbest: 0.8298700 (1)\ttotal: 695ms\tremaining: 27.1s\n",
      "10:\tlearn: 0.8245901\ttest: 0.8245901\ttest1: 0.8240352\tbest: 0.8298700 (1)\ttotal: 754ms\tremaining: 26.7s\n",
      "11:\tlearn: 0.8290808\ttest: 0.8290808\ttest1: 0.8253660\tbest: 0.8298700 (1)\ttotal: 809ms\tremaining: 26.2s\n",
      "12:\tlearn: 0.8310540\ttest: 0.8310540\ttest1: 0.8303818\tbest: 0.8303818 (12)\ttotal: 906ms\tremaining: 27s\n",
      "13:\tlearn: 0.8331973\ttest: 0.8331973\ttest1: 0.8324291\tbest: 0.8324291 (13)\ttotal: 1.08s\tremaining: 29.8s\n",
      "14:\tlearn: 0.8349663\ttest: 0.8349663\ttest1: 0.8328386\tbest: 0.8328386 (14)\ttotal: 1.25s\tremaining: 32.1s\n",
      "15:\tlearn: 0.8355447\ttest: 0.8355447\ttest1: 0.8326338\tbest: 0.8328386 (14)\ttotal: 1.31s\tremaining: 31.4s\n",
      "16:\tlearn: 0.8376880\ttest: 0.8376880\ttest1: 0.8344764\tbest: 0.8344764 (16)\ttotal: 1.42s\tremaining: 32s\n",
      "17:\tlearn: 0.8384024\ttest: 0.8384024\ttest1: 0.8356024\tbest: 0.8356024 (17)\ttotal: 1.66s\tremaining: 35.3s\n",
      "18:\tlearn: 0.8402055\ttest: 0.8402055\ttest1: 0.8358071\tbest: 0.8358071 (18)\ttotal: 1.84s\tremaining: 36.9s\n",
      "19:\tlearn: 0.8427230\ttest: 0.8427230\ttest1: 0.8347835\tbest: 0.8358071 (18)\ttotal: 2.02s\tremaining: 38.4s\n",
      "20:\tlearn: 0.8427910\ttest: 0.8427910\ttest1: 0.8344764\tbest: 0.8358071 (18)\ttotal: 2.2s\tremaining: 39.7s\n",
      "21:\tlearn: 0.8433694\ttest: 0.8433694\ttest1: 0.8350906\tbest: 0.8358071 (18)\ttotal: 2.46s\tremaining: 42.4s\n",
      "22:\tlearn: 0.8442199\ttest: 0.8442199\ttest1: 0.8371379\tbest: 0.8371379 (22)\ttotal: 2.65s\tremaining: 43.5s\n",
      "23:\tlearn: 0.8445261\ttest: 0.8445261\ttest1: 0.8373426\tbest: 0.8373426 (23)\ttotal: 2.83s\tremaining: 44.4s\n",
      "24:\tlearn: 0.8453086\ttest: 0.8453086\ttest1: 0.8372402\tbest: 0.8373426 (23)\ttotal: 3.01s\tremaining: 45.1s\n",
      "25:\tlearn: 0.8467714\ttest: 0.8467714\ttest1: 0.8376497\tbest: 0.8376497 (25)\ttotal: 3.11s\tremaining: 44.7s\n",
      "26:\tlearn: 0.8460910\ttest: 0.8460910\ttest1: 0.8386734\tbest: 0.8386734 (26)\ttotal: 3.18s\tremaining: 43.9s\n",
      "27:\tlearn: 0.8468735\ttest: 0.8468735\ttest1: 0.8386734\tbest: 0.8386734 (26)\ttotal: 3.3s\tremaining: 43.8s\n",
      "28:\tlearn: 0.8475539\ttest: 0.8475539\ttest1: 0.8378544\tbest: 0.8386734 (26)\ttotal: 3.4s\tremaining: 43.5s\n",
      "29:\tlearn: 0.8480302\ttest: 0.8480302\ttest1: 0.8379568\tbest: 0.8386734 (26)\ttotal: 3.5s\tremaining: 43.2s\n",
      "30:\tlearn: 0.8482343\ttest: 0.8482343\ttest1: 0.8390828\tbest: 0.8390828 (30)\ttotal: 3.59s\tremaining: 42.7s\n",
      "31:\tlearn: 0.8494931\ttest: 0.8494931\ttest1: 0.8388781\tbest: 0.8390828 (30)\ttotal: 3.67s\tremaining: 42.2s\n",
      "32:\tlearn: 0.8504116\ttest: 0.8504116\ttest1: 0.8390828\tbest: 0.8390828 (30)\ttotal: 3.75s\tremaining: 41.6s\n",
      "33:\tlearn: 0.8515003\ttest: 0.8515003\ttest1: 0.8400041\tbest: 0.8400041 (33)\ttotal: 3.81s\tremaining: 41s\n",
      "34:\tlearn: 0.8513642\ttest: 0.8513642\ttest1: 0.8405159\tbest: 0.8405159 (34)\ttotal: 3.87s\tremaining: 40.3s\n",
      "35:\tlearn: 0.8517385\ttest: 0.8517385\ttest1: 0.8407206\tbest: 0.8407206 (35)\ttotal: 3.99s\tremaining: 40.4s\n",
      "36:\tlearn: 0.8521467\ttest: 0.8521467\ttest1: 0.8412325\tbest: 0.8412325 (36)\ttotal: 4.05s\tremaining: 39.7s\n",
      "37:\tlearn: 0.8519086\ttest: 0.8519086\ttest1: 0.8421538\tbest: 0.8421538 (37)\ttotal: 4.11s\tremaining: 39.2s\n",
      "38:\tlearn: 0.8515683\ttest: 0.8515683\ttest1: 0.8423585\tbest: 0.8423585 (38)\ttotal: 4.17s\tremaining: 38.6s\n",
      "39:\tlearn: 0.8524869\ttest: 0.8524869\ttest1: 0.8435869\tbest: 0.8435869 (39)\ttotal: 4.25s\tremaining: 38.2s\n",
      "40:\tlearn: 0.8524869\ttest: 0.8524869\ttest1: 0.8436892\tbest: 0.8436892 (40)\ttotal: 4.31s\tremaining: 37.7s\n",
      "41:\tlearn: 0.8530312\ttest: 0.8530312\ttest1: 0.8434845\tbest: 0.8436892 (40)\ttotal: 4.37s\tremaining: 37.2s\n",
      "42:\tlearn: 0.8537457\ttest: 0.8537457\ttest1: 0.8434845\tbest: 0.8436892 (40)\ttotal: 4.42s\tremaining: 36.7s\n",
      "43:\tlearn: 0.8537457\ttest: 0.8537457\ttest1: 0.8434845\tbest: 0.8436892 (40)\ttotal: 4.51s\tremaining: 36.5s\n",
      "44:\tlearn: 0.8543580\ttest: 0.8543580\ttest1: 0.8439963\tbest: 0.8439963 (44)\ttotal: 4.59s\tremaining: 36.2s\n",
      "45:\tlearn: 0.8543580\ttest: 0.8543580\ttest1: 0.8433821\tbest: 0.8439963 (44)\ttotal: 4.65s\tremaining: 35.8s\n",
      "46:\tlearn: 0.8551745\ttest: 0.8551745\ttest1: 0.8437916\tbest: 0.8439963 (44)\ttotal: 4.71s\tremaining: 35.4s\n",
      "47:\tlearn: 0.8554127\ttest: 0.8554127\ttest1: 0.8429727\tbest: 0.8439963 (44)\ttotal: 4.8s\tremaining: 35.2s\n",
      "48:\tlearn: 0.8562292\ttest: 0.8562292\ttest1: 0.8430750\tbest: 0.8439963 (44)\ttotal: 4.86s\tremaining: 34.8s\n",
      "49:\tlearn: 0.8565353\ttest: 0.8565353\ttest1: 0.8437916\tbest: 0.8439963 (44)\ttotal: 4.92s\tremaining: 34.4s\n",
      "50:\tlearn: 0.8563312\ttest: 0.8563312\ttest1: 0.8436892\tbest: 0.8439963 (44)\ttotal: 4.98s\tremaining: 34.1s\n",
      "51:\tlearn: 0.8565013\ttest: 0.8565013\ttest1: 0.8446105\tbest: 0.8446105 (51)\ttotal: 5.09s\tremaining: 34.1s\n",
      "52:\tlearn: 0.8570116\ttest: 0.8570116\ttest1: 0.8447129\tbest: 0.8447129 (52)\ttotal: 5.15s\tremaining: 33.7s\n",
      "53:\tlearn: 0.8575900\ttest: 0.8575900\ttest1: 0.8450200\tbest: 0.8450200 (53)\ttotal: 5.22s\tremaining: 33.4s\n",
      "54:\tlearn: 0.8574879\ttest: 0.8574879\ttest1: 0.8449176\tbest: 0.8450200 (53)\ttotal: 5.28s\tremaining: 33.1s\n",
      "55:\tlearn: 0.8577261\ttest: 0.8577261\ttest1: 0.8451223\tbest: 0.8451223 (55)\ttotal: 5.38s\tremaining: 33.1s\n",
      "56:\tlearn: 0.8581003\ttest: 0.8581003\ttest1: 0.8456341\tbest: 0.8456341 (56)\ttotal: 5.45s\tremaining: 32.8s\n",
      "57:\tlearn: 0.8589168\ttest: 0.8589168\ttest1: 0.8454294\tbest: 0.8456341 (56)\ttotal: 5.58s\tremaining: 32.9s\n",
      "58:\tlearn: 0.8589508\ttest: 0.8589508\ttest1: 0.8456341\tbest: 0.8456341 (56)\ttotal: 5.71s\tremaining: 33s\n",
      "59:\tlearn: 0.8595632\ttest: 0.8595632\ttest1: 0.8455318\tbest: 0.8456341 (56)\ttotal: 5.82s\tremaining: 33s\n",
      "60:\tlearn: 0.8594611\ttest: 0.8594611\ttest1: 0.8463507\tbest: 0.8463507 (60)\ttotal: 5.96s\tremaining: 33.1s\n",
      "61:\tlearn: 0.8592910\ttest: 0.8592910\ttest1: 0.8463507\tbest: 0.8463507 (60)\ttotal: 6.02s\tremaining: 32.8s\n",
      "62:\tlearn: 0.8593591\ttest: 0.8593591\ttest1: 0.8461460\tbest: 0.8463507 (60)\ttotal: 6.24s\tremaining: 33.4s\n",
      "63:\tlearn: 0.8599374\ttest: 0.8599374\ttest1: 0.8471696\tbest: 0.8471696 (63)\ttotal: 6.52s\tremaining: 34.2s\n",
      "64:\tlearn: 0.8599374\ttest: 0.8599374\ttest1: 0.8472720\tbest: 0.8472720 (64)\ttotal: 6.6s\tremaining: 34s\n",
      "65:\tlearn: 0.8606518\ttest: 0.8606518\ttest1: 0.8471696\tbest: 0.8472720 (64)\ttotal: 6.7s\tremaining: 33.9s\n",
      "66:\tlearn: 0.8602096\ttest: 0.8602096\ttest1: 0.8468625\tbest: 0.8472720 (64)\ttotal: 6.8s\tremaining: 33.8s\n",
      "67:\tlearn: 0.8601755\ttest: 0.8601755\ttest1: 0.8472720\tbest: 0.8472720 (64)\ttotal: 6.87s\tremaining: 33.5s\n",
      "68:\tlearn: 0.8601755\ttest: 0.8601755\ttest1: 0.8468625\tbest: 0.8472720 (64)\ttotal: 6.95s\tremaining: 33.3s\n",
      "69:\tlearn: 0.8602776\ttest: 0.8602776\ttest1: 0.8468625\tbest: 0.8472720 (64)\ttotal: 7.06s\tremaining: 33.3s\n",
      "70:\tlearn: 0.8615023\ttest: 0.8615023\ttest1: 0.8468625\tbest: 0.8472720 (64)\ttotal: 7.14s\tremaining: 33.1s\n",
      "71:\tlearn: 0.8618766\ttest: 0.8618766\ttest1: 0.8473743\tbest: 0.8473743 (71)\ttotal: 7.2s\tremaining: 32.8s\n",
      "72:\tlearn: 0.8617405\ttest: 0.8617405\ttest1: 0.8472720\tbest: 0.8473743 (71)\ttotal: 7.24s\tremaining: 32.5s\n",
      "73:\tlearn: 0.8625910\ttest: 0.8625910\ttest1: 0.8486027\tbest: 0.8486027 (73)\ttotal: 7.33s\tremaining: 32.3s\n",
      "74:\tlearn: 0.8628632\ttest: 0.8628632\ttest1: 0.8487051\tbest: 0.8487051 (74)\ttotal: 7.39s\tremaining: 32s\n",
      "75:\tlearn: 0.8630673\ttest: 0.8630673\ttest1: 0.8487051\tbest: 0.8487051 (74)\ttotal: 7.45s\tremaining: 31.8s\n",
      "76:\tlearn: 0.8632374\ttest: 0.8632374\ttest1: 0.8485004\tbest: 0.8487051 (74)\ttotal: 7.5s\tremaining: 31.5s\n",
      "77:\tlearn: 0.8632714\ttest: 0.8632714\ttest1: 0.8498311\tbest: 0.8498311 (77)\ttotal: 7.57s\tremaining: 31.3s\n",
      "78:\tlearn: 0.8633054\ttest: 0.8633054\ttest1: 0.8497287\tbest: 0.8498311 (77)\ttotal: 7.63s\tremaining: 31s\n",
      "79:\tlearn: 0.8637817\ttest: 0.8637817\ttest1: 0.8495240\tbest: 0.8498311 (77)\ttotal: 7.69s\tremaining: 30.8s\n",
      "80:\tlearn: 0.8643601\ttest: 0.8643601\ttest1: 0.8498311\tbest: 0.8498311 (77)\ttotal: 7.75s\tremaining: 30.5s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "81:\tlearn: 0.8646322\ttest: 0.8646322\ttest1: 0.8500358\tbest: 0.8500358 (81)\ttotal: 7.83s\tremaining: 30.4s\n",
      "82:\tlearn: 0.8646322\ttest: 0.8646322\ttest1: 0.8500358\tbest: 0.8500358 (81)\ttotal: 7.92s\tremaining: 30.2s\n",
      "83:\tlearn: 0.8651085\ttest: 0.8651085\ttest1: 0.8499335\tbest: 0.8500358 (81)\ttotal: 8.01s\tremaining: 30.1s\n",
      "84:\tlearn: 0.8649044\ttest: 0.8649044\ttest1: 0.8496264\tbest: 0.8500358 (81)\ttotal: 8.11s\tremaining: 30.1s\n",
      "85:\tlearn: 0.8658230\ttest: 0.8658230\ttest1: 0.8495240\tbest: 0.8500358 (81)\ttotal: 8.18s\tremaining: 29.9s\n",
      "86:\tlearn: 0.8662652\ttest: 0.8662652\ttest1: 0.8494216\tbest: 0.8500358 (81)\ttotal: 8.29s\tremaining: 29.8s\n",
      "87:\tlearn: 0.8661972\ttest: 0.8661972\ttest1: 0.8494216\tbest: 0.8500358 (81)\ttotal: 8.43s\tremaining: 29.9s\n",
      "88:\tlearn: 0.8663673\ttest: 0.8663673\ttest1: 0.8494216\tbest: 0.8500358 (81)\ttotal: 8.52s\tremaining: 29.8s\n",
      "89:\tlearn: 0.8660951\ttest: 0.8660951\ttest1: 0.8492169\tbest: 0.8500358 (81)\ttotal: 8.61s\tremaining: 29.7s\n",
      "90:\tlearn: 0.8665374\ttest: 0.8665374\ttest1: 0.8493193\tbest: 0.8500358 (81)\ttotal: 8.73s\tremaining: 29.7s\n",
      "91:\tlearn: 0.8670477\ttest: 0.8670477\ttest1: 0.8499335\tbest: 0.8500358 (81)\ttotal: 8.83s\tremaining: 29.6s\n",
      "92:\tlearn: 0.8682384\ttest: 0.8682384\ttest1: 0.8513666\tbest: 0.8513666 (92)\ttotal: 8.93s\tremaining: 29.5s\n",
      "93:\tlearn: 0.8682724\ttest: 0.8682724\ttest1: 0.8522878\tbest: 0.8522878 (93)\ttotal: 9.02s\tremaining: 29.4s\n",
      "94:\tlearn: 0.8681704\ttest: 0.8681704\ttest1: 0.8526973\tbest: 0.8526973 (94)\ttotal: 9.17s\tremaining: 29.4s\n",
      "95:\tlearn: 0.8682044\ttest: 0.8682044\ttest1: 0.8523902\tbest: 0.8526973 (94)\ttotal: 9.34s\tremaining: 29.6s\n",
      "96:\tlearn: 0.8681704\ttest: 0.8681704\ttest1: 0.8523902\tbest: 0.8526973 (94)\ttotal: 9.42s\tremaining: 29.4s\n",
      "97:\tlearn: 0.8685786\ttest: 0.8685786\ttest1: 0.8526973\tbest: 0.8526973 (94)\ttotal: 9.53s\tremaining: 29.4s\n",
      "98:\tlearn: 0.8688848\ttest: 0.8688848\ttest1: 0.8533115\tbest: 0.8533115 (98)\ttotal: 9.64s\tremaining: 29.3s\n",
      "99:\tlearn: 0.8690889\ttest: 0.8690889\ttest1: 0.8530044\tbest: 0.8533115 (98)\ttotal: 9.7s\tremaining: 29.1s\n",
      "100:\tlearn: 0.8689869\ttest: 0.8689869\ttest1: 0.8532091\tbest: 0.8533115 (98)\ttotal: 9.75s\tremaining: 28.9s\n",
      "101:\tlearn: 0.8691570\ttest: 0.8691570\ttest1: 0.8526973\tbest: 0.8533115 (98)\ttotal: 9.81s\tremaining: 28.6s\n",
      "102:\tlearn: 0.8693951\ttest: 0.8693951\ttest1: 0.8532091\tbest: 0.8533115 (98)\ttotal: 9.87s\tremaining: 28.5s\n",
      "103:\tlearn: 0.8692590\ttest: 0.8692590\ttest1: 0.8532091\tbest: 0.8533115 (98)\ttotal: 9.93s\tremaining: 28.3s\n",
      "104:\tlearn: 0.8690549\ttest: 0.8690549\ttest1: 0.8529020\tbest: 0.8533115 (98)\ttotal: 10s\tremaining: 28.1s\n",
      "105:\tlearn: 0.8698034\ttest: 0.8698034\ttest1: 0.8526973\tbest: 0.8533115 (98)\ttotal: 10.1s\tremaining: 27.9s\n",
      "106:\tlearn: 0.8701436\ttest: 0.8701436\ttest1: 0.8524926\tbest: 0.8533115 (98)\ttotal: 10.1s\tremaining: 27.8s\n",
      "107:\tlearn: 0.8702796\ttest: 0.8702796\ttest1: 0.8524926\tbest: 0.8533115 (98)\ttotal: 10.2s\tremaining: 27.6s\n",
      "108:\tlearn: 0.8709601\ttest: 0.8709601\ttest1: 0.8525949\tbest: 0.8533115 (98)\ttotal: 10.3s\tremaining: 27.5s\n",
      "109:\tlearn: 0.8711642\ttest: 0.8711642\ttest1: 0.8526973\tbest: 0.8533115 (98)\ttotal: 10.4s\tremaining: 27.4s\n",
      "110:\tlearn: 0.8710281\ttest: 0.8710281\ttest1: 0.8527997\tbest: 0.8533115 (98)\ttotal: 10.5s\tremaining: 27.3s\n",
      "111:\tlearn: 0.8706539\ttest: 0.8706539\ttest1: 0.8531068\tbest: 0.8533115 (98)\ttotal: 10.6s\tremaining: 27.2s\n",
      "112:\tlearn: 0.8708580\ttest: 0.8708580\ttest1: 0.8533115\tbest: 0.8533115 (98)\ttotal: 10.7s\tremaining: 27.1s\n",
      "113:\tlearn: 0.8709601\ttest: 0.8709601\ttest1: 0.8533115\tbest: 0.8533115 (98)\ttotal: 10.8s\tremaining: 27s\n",
      "114:\tlearn: 0.8709601\ttest: 0.8709601\ttest1: 0.8533115\tbest: 0.8533115 (98)\ttotal: 10.8s\tremaining: 26.8s\n",
      "115:\tlearn: 0.8710961\ttest: 0.8710961\ttest1: 0.8527997\tbest: 0.8533115 (98)\ttotal: 10.9s\tremaining: 26.6s\n",
      "116:\tlearn: 0.8718106\ttest: 0.8718106\ttest1: 0.8530044\tbest: 0.8533115 (98)\ttotal: 10.9s\tremaining: 26.5s\n",
      "117:\tlearn: 0.8716745\ttest: 0.8716745\ttest1: 0.8531068\tbest: 0.8533115 (98)\ttotal: 11s\tremaining: 26.3s\n",
      "118:\tlearn: 0.8717425\ttest: 0.8717425\ttest1: 0.8533115\tbest: 0.8533115 (98)\ttotal: 11.1s\tremaining: 26.1s\n",
      "119:\tlearn: 0.8716405\ttest: 0.8716405\ttest1: 0.8534139\tbest: 0.8534139 (119)\ttotal: 11.1s\tremaining: 26s\n",
      "120:\tlearn: 0.8717085\ttest: 0.8717085\ttest1: 0.8534139\tbest: 0.8534139 (119)\ttotal: 11.2s\tremaining: 25.9s\n",
      "121:\tlearn: 0.8718446\ttest: 0.8718446\ttest1: 0.8531068\tbest: 0.8534139 (119)\ttotal: 11.3s\tremaining: 25.7s\n",
      "122:\tlearn: 0.8724229\ttest: 0.8724229\ttest1: 0.8532091\tbest: 0.8534139 (119)\ttotal: 11.4s\tremaining: 25.6s\n",
      "123:\tlearn: 0.8726271\ttest: 0.8726271\ttest1: 0.8535162\tbest: 0.8535162 (123)\ttotal: 11.4s\tremaining: 25.4s\n",
      "124:\tlearn: 0.8727631\ttest: 0.8727631\ttest1: 0.8534139\tbest: 0.8535162 (123)\ttotal: 11.5s\tremaining: 25.3s\n",
      "125:\tlearn: 0.8725930\ttest: 0.8725930\ttest1: 0.8533115\tbest: 0.8535162 (123)\ttotal: 11.6s\tremaining: 25.1s\n",
      "126:\tlearn: 0.8731374\ttest: 0.8731374\ttest1: 0.8535162\tbest: 0.8535162 (123)\ttotal: 11.6s\tremaining: 25s\n",
      "127:\tlearn: 0.8734095\ttest: 0.8734095\ttest1: 0.8537210\tbest: 0.8537210 (127)\ttotal: 11.7s\tremaining: 24.9s\n",
      "128:\tlearn: 0.8739198\ttest: 0.8739198\ttest1: 0.8538233\tbest: 0.8538233 (128)\ttotal: 11.8s\tremaining: 24.8s\n",
      "129:\tlearn: 0.8742601\ttest: 0.8742601\ttest1: 0.8537210\tbest: 0.8538233 (128)\ttotal: 11.9s\tremaining: 24.8s\n",
      "130:\tlearn: 0.8743621\ttest: 0.8743621\ttest1: 0.8534139\tbest: 0.8538233 (128)\ttotal: 12s\tremaining: 24.6s\n",
      "131:\tlearn: 0.8743961\ttest: 0.8743961\ttest1: 0.8534139\tbest: 0.8538233 (128)\ttotal: 12.1s\tremaining: 24.5s\n",
      "132:\tlearn: 0.8746343\ttest: 0.8746343\ttest1: 0.8535162\tbest: 0.8538233 (128)\ttotal: 12.1s\tremaining: 24.3s\n",
      "133:\tlearn: 0.8744642\ttest: 0.8744642\ttest1: 0.8532091\tbest: 0.8538233 (128)\ttotal: 12.2s\tremaining: 24.2s\n",
      "134:\tlearn: 0.8742601\ttest: 0.8742601\ttest1: 0.8538233\tbest: 0.8538233 (128)\ttotal: 12.3s\tremaining: 24.1s\n",
      "135:\tlearn: 0.8747023\ttest: 0.8747023\ttest1: 0.8537210\tbest: 0.8538233 (128)\ttotal: 12.4s\tremaining: 24.1s\n",
      "136:\tlearn: 0.8746003\ttest: 0.8746003\ttest1: 0.8539257\tbest: 0.8539257 (136)\ttotal: 12.5s\tremaining: 24s\n",
      "137:\tlearn: 0.8745322\ttest: 0.8745322\ttest1: 0.8540280\tbest: 0.8540280 (137)\ttotal: 12.6s\tremaining: 23.9s\n",
      "138:\tlearn: 0.8746003\ttest: 0.8746003\ttest1: 0.8538233\tbest: 0.8540280 (137)\ttotal: 12.6s\tremaining: 23.7s\n",
      "139:\tlearn: 0.8749064\ttest: 0.8749064\ttest1: 0.8539257\tbest: 0.8540280 (137)\ttotal: 12.7s\tremaining: 23.6s\n",
      "140:\tlearn: 0.8748724\ttest: 0.8748724\ttest1: 0.8540280\tbest: 0.8540280 (137)\ttotal: 12.8s\tremaining: 23.6s\n",
      "141:\tlearn: 0.8754848\ttest: 0.8754848\ttest1: 0.8541304\tbest: 0.8541304 (141)\ttotal: 13s\tremaining: 23.7s\n",
      "142:\tlearn: 0.8756889\ttest: 0.8756889\ttest1: 0.8537210\tbest: 0.8541304 (141)\ttotal: 13.2s\tremaining: 23.7s\n",
      "143:\tlearn: 0.8757570\ttest: 0.8757570\ttest1: 0.8534139\tbest: 0.8541304 (141)\ttotal: 13.2s\tremaining: 23.5s\n",
      "144:\tlearn: 0.8760291\ttest: 0.8760291\ttest1: 0.8535162\tbest: 0.8541304 (141)\ttotal: 13.3s\tremaining: 23.4s\n",
      "145:\tlearn: 0.8762673\ttest: 0.8762673\ttest1: 0.8536186\tbest: 0.8541304 (141)\ttotal: 13.3s\tremaining: 23.2s\n",
      "146:\tlearn: 0.8760631\ttest: 0.8760631\ttest1: 0.8535162\tbest: 0.8541304 (141)\ttotal: 13.4s\tremaining: 23.1s\n",
      "147:\tlearn: 0.8760972\ttest: 0.8760972\ttest1: 0.8534139\tbest: 0.8541304 (141)\ttotal: 13.5s\tremaining: 23s\n",
      "148:\tlearn: 0.8761652\ttest: 0.8761652\ttest1: 0.8536186\tbest: 0.8541304 (141)\ttotal: 13.5s\tremaining: 22.8s\n",
      "149:\tlearn: 0.8760631\ttest: 0.8760631\ttest1: 0.8535162\tbest: 0.8541304 (141)\ttotal: 13.6s\tremaining: 22.7s\n",
      "150:\tlearn: 0.8760972\ttest: 0.8760972\ttest1: 0.8536186\tbest: 0.8541304 (141)\ttotal: 13.7s\tremaining: 22.6s\n",
      "151:\tlearn: 0.8765735\ttest: 0.8765735\ttest1: 0.8536186\tbest: 0.8541304 (141)\ttotal: 13.7s\tremaining: 22.4s\n",
      "152:\tlearn: 0.8764374\ttest: 0.8764374\ttest1: 0.8536186\tbest: 0.8541304 (141)\ttotal: 13.8s\tremaining: 22.3s\n",
      "153:\tlearn: 0.8763693\ttest: 0.8763693\ttest1: 0.8536186\tbest: 0.8541304 (141)\ttotal: 13.9s\tremaining: 22.1s\n",
      "154:\tlearn: 0.8771858\ttest: 0.8771858\ttest1: 0.8538233\tbest: 0.8541304 (141)\ttotal: 13.9s\tremaining: 22s\n",
      "155:\tlearn: 0.8770838\ttest: 0.8770838\ttest1: 0.8529020\tbest: 0.8541304 (141)\ttotal: 14s\tremaining: 21.9s\n",
      "156:\tlearn: 0.8770838\ttest: 0.8770838\ttest1: 0.8530044\tbest: 0.8541304 (141)\ttotal: 14.1s\tremaining: 21.8s\n",
      "157:\tlearn: 0.8771858\ttest: 0.8771858\ttest1: 0.8530044\tbest: 0.8541304 (141)\ttotal: 14.1s\tremaining: 21.6s\n",
      "158:\tlearn: 0.8771518\ttest: 0.8771518\ttest1: 0.8530044\tbest: 0.8541304 (141)\ttotal: 14.2s\tremaining: 21.5s\n",
      "159:\tlearn: 0.8770497\ttest: 0.8770497\ttest1: 0.8531068\tbest: 0.8541304 (141)\ttotal: 14.2s\tremaining: 21.4s\n",
      "160:\tlearn: 0.8772539\ttest: 0.8772539\ttest1: 0.8531068\tbest: 0.8541304 (141)\ttotal: 14.3s\tremaining: 21.2s\n",
      "161:\tlearn: 0.8768796\ttest: 0.8768796\ttest1: 0.8532091\tbest: 0.8541304 (141)\ttotal: 14.3s\tremaining: 21.1s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "162:\tlearn: 0.8773219\ttest: 0.8773219\ttest1: 0.8536186\tbest: 0.8541304 (141)\ttotal: 14.4s\tremaining: 21s\n",
      "163:\tlearn: 0.8773559\ttest: 0.8773559\ttest1: 0.8538233\tbest: 0.8541304 (141)\ttotal: 14.5s\tremaining: 20.8s\n",
      "164:\tlearn: 0.8775941\ttest: 0.8775941\ttest1: 0.8542328\tbest: 0.8542328 (164)\ttotal: 14.5s\tremaining: 20.7s\n",
      "165:\tlearn: 0.8779003\ttest: 0.8779003\ttest1: 0.8541304\tbest: 0.8542328 (164)\ttotal: 14.6s\tremaining: 20.6s\n",
      "166:\tlearn: 0.8780363\ttest: 0.8780363\ttest1: 0.8543351\tbest: 0.8543351 (166)\ttotal: 14.8s\tremaining: 20.7s\n",
      "167:\tlearn: 0.8782405\ttest: 0.8782405\ttest1: 0.8543351\tbest: 0.8543351 (166)\ttotal: 14.9s\tremaining: 20.6s\n",
      "168:\tlearn: 0.8780704\ttest: 0.8780704\ttest1: 0.8539257\tbest: 0.8543351 (166)\ttotal: 15s\tremaining: 20.5s\n",
      "169:\tlearn: 0.8782064\ttest: 0.8782064\ttest1: 0.8539257\tbest: 0.8543351 (166)\ttotal: 15.1s\tremaining: 20.4s\n",
      "170:\tlearn: 0.8782745\ttest: 0.8782745\ttest1: 0.8543351\tbest: 0.8543351 (166)\ttotal: 15.1s\tremaining: 20.2s\n",
      "171:\tlearn: 0.8783425\ttest: 0.8783425\ttest1: 0.8545399\tbest: 0.8545399 (171)\ttotal: 15.2s\tremaining: 20.1s\n",
      "172:\tlearn: 0.8785126\ttest: 0.8785126\ttest1: 0.8549493\tbest: 0.8549493 (172)\ttotal: 15.2s\tremaining: 20s\n",
      "173:\tlearn: 0.8787508\ttest: 0.8787508\ttest1: 0.8547446\tbest: 0.8549493 (172)\ttotal: 15.3s\tremaining: 19.9s\n",
      "174:\tlearn: 0.8785466\ttest: 0.8785466\ttest1: 0.8550517\tbest: 0.8550517 (174)\ttotal: 15.4s\tremaining: 19.8s\n",
      "175:\tlearn: 0.8786827\ttest: 0.8786827\ttest1: 0.8553588\tbest: 0.8553588 (175)\ttotal: 15.4s\tremaining: 19.7s\n",
      "176:\tlearn: 0.8790570\ttest: 0.8790570\ttest1: 0.8553588\tbest: 0.8553588 (175)\ttotal: 15.5s\tremaining: 19.5s\n",
      "177:\tlearn: 0.8789889\ttest: 0.8789889\ttest1: 0.8551541\tbest: 0.8553588 (175)\ttotal: 15.6s\tremaining: 19.4s\n",
      "178:\tlearn: 0.8793291\ttest: 0.8793291\ttest1: 0.8554612\tbest: 0.8554612 (178)\ttotal: 15.6s\tremaining: 19.3s\n",
      "179:\tlearn: 0.8794652\ttest: 0.8794652\ttest1: 0.8553588\tbest: 0.8554612 (178)\ttotal: 15.7s\tremaining: 19.2s\n",
      "180:\tlearn: 0.8792271\ttest: 0.8792271\ttest1: 0.8554612\tbest: 0.8554612 (178)\ttotal: 15.8s\tremaining: 19.1s\n",
      "181:\tlearn: 0.8792611\ttest: 0.8792611\ttest1: 0.8554612\tbest: 0.8554612 (178)\ttotal: 15.8s\tremaining: 18.9s\n",
      "182:\tlearn: 0.8792611\ttest: 0.8792611\ttest1: 0.8556659\tbest: 0.8556659 (182)\ttotal: 15.9s\tremaining: 18.8s\n",
      "183:\tlearn: 0.8792271\ttest: 0.8792271\ttest1: 0.8558706\tbest: 0.8558706 (183)\ttotal: 15.9s\tremaining: 18.7s\n",
      "184:\tlearn: 0.8792951\ttest: 0.8792951\ttest1: 0.8558706\tbest: 0.8558706 (183)\ttotal: 16s\tremaining: 18.6s\n",
      "185:\tlearn: 0.8792611\ttest: 0.8792611\ttest1: 0.8560753\tbest: 0.8560753 (185)\ttotal: 16.1s\tremaining: 18.5s\n",
      "186:\tlearn: 0.8792271\ttest: 0.8792271\ttest1: 0.8561777\tbest: 0.8561777 (186)\ttotal: 16.1s\tremaining: 18.4s\n",
      "187:\tlearn: 0.8796013\ttest: 0.8796013\ttest1: 0.8555635\tbest: 0.8561777 (186)\ttotal: 16.2s\tremaining: 18.2s\n",
      "188:\tlearn: 0.8796693\ttest: 0.8796693\ttest1: 0.8555635\tbest: 0.8561777 (186)\ttotal: 16.2s\tremaining: 18.1s\n",
      "189:\tlearn: 0.8798394\ttest: 0.8798394\ttest1: 0.8553588\tbest: 0.8561777 (186)\ttotal: 16.3s\tremaining: 18s\n",
      "190:\tlearn: 0.8798394\ttest: 0.8798394\ttest1: 0.8554612\tbest: 0.8561777 (186)\ttotal: 16.4s\tremaining: 17.9s\n",
      "191:\tlearn: 0.8795673\ttest: 0.8795673\ttest1: 0.8553588\tbest: 0.8561777 (186)\ttotal: 16.4s\tremaining: 17.8s\n",
      "192:\tlearn: 0.8797714\ttest: 0.8797714\ttest1: 0.8558706\tbest: 0.8561777 (186)\ttotal: 16.5s\tremaining: 17.7s\n",
      "193:\tlearn: 0.8798394\ttest: 0.8798394\ttest1: 0.8557682\tbest: 0.8561777 (186)\ttotal: 16.6s\tremaining: 17.6s\n",
      "194:\tlearn: 0.8799075\ttest: 0.8799075\ttest1: 0.8557682\tbest: 0.8561777 (186)\ttotal: 16.7s\tremaining: 17.5s\n",
      "195:\tlearn: 0.8798734\ttest: 0.8798734\ttest1: 0.8560753\tbest: 0.8561777 (186)\ttotal: 16.8s\tremaining: 17.5s\n",
      "196:\tlearn: 0.8800095\ttest: 0.8800095\ttest1: 0.8561777\tbest: 0.8561777 (186)\ttotal: 16.9s\tremaining: 17.4s\n",
      "197:\tlearn: 0.8804518\ttest: 0.8804518\ttest1: 0.8562801\tbest: 0.8562801 (197)\ttotal: 17s\tremaining: 17.3s\n",
      "198:\tlearn: 0.8804518\ttest: 0.8804518\ttest1: 0.8561777\tbest: 0.8562801 (197)\ttotal: 17.1s\tremaining: 17.2s\n",
      "199:\tlearn: 0.8805539\ttest: 0.8805539\ttest1: 0.8562801\tbest: 0.8562801 (197)\ttotal: 17.1s\tremaining: 17.1s\n",
      "200:\tlearn: 0.8802817\ttest: 0.8802817\ttest1: 0.8558706\tbest: 0.8562801 (197)\ttotal: 17.2s\tremaining: 17.1s\n",
      "201:\tlearn: 0.8807580\ttest: 0.8807580\ttest1: 0.8561777\tbest: 0.8562801 (197)\ttotal: 17.3s\tremaining: 17s\n",
      "202:\tlearn: 0.8806899\ttest: 0.8806899\ttest1: 0.8561777\tbest: 0.8562801 (197)\ttotal: 17.3s\tremaining: 16.8s\n",
      "203:\tlearn: 0.8805539\ttest: 0.8805539\ttest1: 0.8556659\tbest: 0.8562801 (197)\ttotal: 17.4s\tremaining: 16.7s\n",
      "204:\tlearn: 0.8804858\ttest: 0.8804858\ttest1: 0.8554612\tbest: 0.8562801 (197)\ttotal: 17.5s\tremaining: 16.6s\n",
      "205:\tlearn: 0.8805539\ttest: 0.8805539\ttest1: 0.8556659\tbest: 0.8562801 (197)\ttotal: 17.5s\tremaining: 16.5s\n",
      "206:\tlearn: 0.8806899\ttest: 0.8806899\ttest1: 0.8560753\tbest: 0.8562801 (197)\ttotal: 17.6s\tremaining: 16.4s\n",
      "207:\tlearn: 0.8806219\ttest: 0.8806219\ttest1: 0.8561777\tbest: 0.8562801 (197)\ttotal: 17.6s\tremaining: 16.3s\n",
      "208:\tlearn: 0.8807920\ttest: 0.8807920\ttest1: 0.8564848\tbest: 0.8564848 (208)\ttotal: 17.7s\tremaining: 16.2s\n",
      "209:\tlearn: 0.8809281\ttest: 0.8809281\ttest1: 0.8562801\tbest: 0.8564848 (208)\ttotal: 17.8s\tremaining: 16.1s\n",
      "210:\tlearn: 0.8809281\ttest: 0.8809281\ttest1: 0.8561777\tbest: 0.8564848 (208)\ttotal: 17.8s\tremaining: 16s\n",
      "211:\tlearn: 0.8812343\ttest: 0.8812343\ttest1: 0.8560753\tbest: 0.8564848 (208)\ttotal: 17.9s\tremaining: 15.8s\n",
      "212:\tlearn: 0.8812343\ttest: 0.8812343\ttest1: 0.8563824\tbest: 0.8564848 (208)\ttotal: 17.9s\tremaining: 15.7s\n",
      "213:\tlearn: 0.8812683\ttest: 0.8812683\ttest1: 0.8562801\tbest: 0.8564848 (208)\ttotal: 18s\tremaining: 15.6s\n",
      "214:\tlearn: 0.8813023\ttest: 0.8813023\ttest1: 0.8563824\tbest: 0.8564848 (208)\ttotal: 18s\tremaining: 15.5s\n",
      "215:\tlearn: 0.8812002\ttest: 0.8812002\ttest1: 0.8561777\tbest: 0.8564848 (208)\ttotal: 18.1s\tremaining: 15.4s\n",
      "216:\tlearn: 0.8814044\ttest: 0.8814044\ttest1: 0.8564848\tbest: 0.8564848 (208)\ttotal: 18.2s\tremaining: 15.3s\n",
      "217:\tlearn: 0.8813023\ttest: 0.8813023\ttest1: 0.8563824\tbest: 0.8564848 (208)\ttotal: 18.2s\tremaining: 15.2s\n",
      "218:\tlearn: 0.8813023\ttest: 0.8813023\ttest1: 0.8563824\tbest: 0.8564848 (208)\ttotal: 18.3s\tremaining: 15.1s\n",
      "219:\tlearn: 0.8814044\ttest: 0.8814044\ttest1: 0.8563824\tbest: 0.8564848 (208)\ttotal: 18.3s\tremaining: 15s\n",
      "220:\tlearn: 0.8814044\ttest: 0.8814044\ttest1: 0.8563824\tbest: 0.8564848 (208)\ttotal: 18.4s\tremaining: 14.9s\n",
      "221:\tlearn: 0.8815745\ttest: 0.8815745\ttest1: 0.8563824\tbest: 0.8564848 (208)\ttotal: 18.4s\tremaining: 14.8s\n",
      "222:\tlearn: 0.8815405\ttest: 0.8815405\ttest1: 0.8565872\tbest: 0.8565872 (222)\ttotal: 18.5s\tremaining: 14.7s\n",
      "223:\tlearn: 0.8815745\ttest: 0.8815745\ttest1: 0.8567919\tbest: 0.8567919 (223)\ttotal: 18.6s\tremaining: 14.6s\n",
      "224:\tlearn: 0.8814724\ttest: 0.8814724\ttest1: 0.8569966\tbest: 0.8569966 (224)\ttotal: 18.7s\tremaining: 14.5s\n",
      "225:\tlearn: 0.8815064\ttest: 0.8815064\ttest1: 0.8568943\tbest: 0.8569966 (224)\ttotal: 18.7s\tremaining: 14.4s\n",
      "226:\tlearn: 0.8815405\ttest: 0.8815405\ttest1: 0.8569966\tbest: 0.8569966 (224)\ttotal: 18.8s\tremaining: 14.3s\n",
      "227:\tlearn: 0.8816765\ttest: 0.8816765\ttest1: 0.8570990\tbest: 0.8570990 (227)\ttotal: 18.8s\tremaining: 14.2s\n",
      "228:\tlearn: 0.8818466\ttest: 0.8818466\ttest1: 0.8561777\tbest: 0.8570990 (227)\ttotal: 18.9s\tremaining: 14.1s\n",
      "229:\tlearn: 0.8817786\ttest: 0.8817786\ttest1: 0.8563824\tbest: 0.8570990 (227)\ttotal: 18.9s\tremaining: 14s\n",
      "230:\tlearn: 0.8819147\ttest: 0.8819147\ttest1: 0.8562801\tbest: 0.8570990 (227)\ttotal: 19s\tremaining: 13.9s\n",
      "231:\tlearn: 0.8818807\ttest: 0.8818807\ttest1: 0.8560753\tbest: 0.8570990 (227)\ttotal: 19s\tremaining: 13.8s\n",
      "232:\tlearn: 0.8818807\ttest: 0.8818807\ttest1: 0.8562801\tbest: 0.8570990 (227)\ttotal: 19.1s\tremaining: 13.7s\n",
      "233:\tlearn: 0.8819827\ttest: 0.8819827\ttest1: 0.8563824\tbest: 0.8570990 (227)\ttotal: 19.2s\tremaining: 13.6s\n",
      "234:\tlearn: 0.8821188\ttest: 0.8821188\ttest1: 0.8569966\tbest: 0.8570990 (227)\ttotal: 19.2s\tremaining: 13.5s\n",
      "235:\tlearn: 0.8821188\ttest: 0.8821188\ttest1: 0.8573037\tbest: 0.8573037 (235)\ttotal: 19.3s\tremaining: 13.4s\n",
      "236:\tlearn: 0.8820848\ttest: 0.8820848\ttest1: 0.8573037\tbest: 0.8573037 (235)\ttotal: 19.3s\tremaining: 13.3s\n",
      "237:\tlearn: 0.8822209\ttest: 0.8822209\ttest1: 0.8574061\tbest: 0.8574061 (237)\ttotal: 19.4s\tremaining: 13.2s\n",
      "238:\tlearn: 0.8822209\ttest: 0.8822209\ttest1: 0.8573037\tbest: 0.8574061 (237)\ttotal: 19.4s\tremaining: 13.1s\n",
      "239:\tlearn: 0.8824250\ttest: 0.8824250\ttest1: 0.8570990\tbest: 0.8574061 (237)\ttotal: 19.5s\tremaining: 13s\n",
      "240:\tlearn: 0.8825611\ttest: 0.8825611\ttest1: 0.8569966\tbest: 0.8574061 (237)\ttotal: 19.6s\tremaining: 12.9s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "241:\tlearn: 0.8830374\ttest: 0.8830374\ttest1: 0.8574061\tbest: 0.8574061 (237)\ttotal: 19.6s\tremaining: 12.8s\n",
      "242:\tlearn: 0.8830033\ttest: 0.8830033\ttest1: 0.8572014\tbest: 0.8574061 (237)\ttotal: 19.7s\tremaining: 12.7s\n",
      "243:\tlearn: 0.8830374\ttest: 0.8830374\ttest1: 0.8577132\tbest: 0.8577132 (243)\ttotal: 19.7s\tremaining: 12.6s\n",
      "244:\tlearn: 0.8829013\ttest: 0.8829013\ttest1: 0.8577132\tbest: 0.8577132 (243)\ttotal: 19.8s\tremaining: 12.5s\n",
      "245:\tlearn: 0.8830033\ttest: 0.8830033\ttest1: 0.8577132\tbest: 0.8577132 (243)\ttotal: 19.8s\tremaining: 12.4s\n",
      "246:\tlearn: 0.8832755\ttest: 0.8832755\ttest1: 0.8574061\tbest: 0.8577132 (243)\ttotal: 19.9s\tremaining: 12.3s\n",
      "247:\tlearn: 0.8831054\ttest: 0.8831054\ttest1: 0.8573037\tbest: 0.8577132 (243)\ttotal: 20s\tremaining: 12.2s\n",
      "248:\tlearn: 0.8830714\ttest: 0.8830714\ttest1: 0.8574061\tbest: 0.8577132 (243)\ttotal: 20s\tremaining: 12.1s\n",
      "249:\tlearn: 0.8832755\ttest: 0.8832755\ttest1: 0.8574061\tbest: 0.8577132 (243)\ttotal: 20.1s\tremaining: 12.1s\n",
      "250:\tlearn: 0.8832755\ttest: 0.8832755\ttest1: 0.8574061\tbest: 0.8577132 (243)\ttotal: 20.1s\tremaining: 12s\n",
      "251:\tlearn: 0.8833435\ttest: 0.8833435\ttest1: 0.8574061\tbest: 0.8577132 (243)\ttotal: 20.2s\tremaining: 11.9s\n",
      "252:\tlearn: 0.8834116\ttest: 0.8834116\ttest1: 0.8573037\tbest: 0.8577132 (243)\ttotal: 20.2s\tremaining: 11.8s\n",
      "253:\tlearn: 0.8833435\ttest: 0.8833435\ttest1: 0.8575084\tbest: 0.8577132 (243)\ttotal: 20.3s\tremaining: 11.7s\n",
      "254:\tlearn: 0.8833435\ttest: 0.8833435\ttest1: 0.8576108\tbest: 0.8577132 (243)\ttotal: 20.4s\tremaining: 11.6s\n",
      "255:\tlearn: 0.8833435\ttest: 0.8833435\ttest1: 0.8575084\tbest: 0.8577132 (243)\ttotal: 20.4s\tremaining: 11.5s\n",
      "256:\tlearn: 0.8833776\ttest: 0.8833776\ttest1: 0.8576108\tbest: 0.8577132 (243)\ttotal: 20.5s\tremaining: 11.4s\n",
      "257:\tlearn: 0.8835477\ttest: 0.8835477\ttest1: 0.8576108\tbest: 0.8577132 (243)\ttotal: 20.5s\tremaining: 11.3s\n",
      "258:\tlearn: 0.8835817\ttest: 0.8835817\ttest1: 0.8575084\tbest: 0.8577132 (243)\ttotal: 20.6s\tremaining: 11.2s\n",
      "259:\tlearn: 0.8836157\ttest: 0.8836157\ttest1: 0.8572014\tbest: 0.8577132 (243)\ttotal: 20.7s\tremaining: 11.1s\n",
      "260:\tlearn: 0.8834456\ttest: 0.8834456\ttest1: 0.8573037\tbest: 0.8577132 (243)\ttotal: 20.7s\tremaining: 11s\n",
      "261:\tlearn: 0.8839219\ttest: 0.8839219\ttest1: 0.8579179\tbest: 0.8579179 (261)\ttotal: 20.8s\tremaining: 11s\n",
      "262:\tlearn: 0.8836157\ttest: 0.8836157\ttest1: 0.8574061\tbest: 0.8579179 (261)\ttotal: 20.9s\tremaining: 10.9s\n",
      "263:\tlearn: 0.8836497\ttest: 0.8836497\ttest1: 0.8574061\tbest: 0.8579179 (261)\ttotal: 21s\tremaining: 10.8s\n",
      "264:\tlearn: 0.8836837\ttest: 0.8836837\ttest1: 0.8574061\tbest: 0.8579179 (261)\ttotal: 21.1s\tremaining: 10.7s\n",
      "265:\tlearn: 0.8838198\ttest: 0.8838198\ttest1: 0.8575084\tbest: 0.8579179 (261)\ttotal: 21.1s\tremaining: 10.6s\n",
      "266:\tlearn: 0.8837858\ttest: 0.8837858\ttest1: 0.8575084\tbest: 0.8579179 (261)\ttotal: 21.1s\tremaining: 10.5s\n",
      "267:\tlearn: 0.8838198\ttest: 0.8838198\ttest1: 0.8574061\tbest: 0.8579179 (261)\ttotal: 21.2s\tremaining: 10.4s\n",
      "268:\tlearn: 0.8838198\ttest: 0.8838198\ttest1: 0.8574061\tbest: 0.8579179 (261)\ttotal: 21.2s\tremaining: 10.3s\n",
      "269:\tlearn: 0.8840580\ttest: 0.8840580\ttest1: 0.8572014\tbest: 0.8579179 (261)\ttotal: 21.3s\tremaining: 10.3s\n",
      "270:\tlearn: 0.8840240\ttest: 0.8840240\ttest1: 0.8575084\tbest: 0.8579179 (261)\ttotal: 21.4s\tremaining: 10.2s\n",
      "271:\tlearn: 0.8839219\ttest: 0.8839219\ttest1: 0.8573037\tbest: 0.8579179 (261)\ttotal: 21.4s\tremaining: 10.1s\n",
      "272:\tlearn: 0.8838879\ttest: 0.8838879\ttest1: 0.8573037\tbest: 0.8579179 (261)\ttotal: 21.5s\tremaining: 10s\n",
      "273:\tlearn: 0.8838879\ttest: 0.8838879\ttest1: 0.8575084\tbest: 0.8579179 (261)\ttotal: 21.7s\tremaining: 9.97s\n",
      "274:\tlearn: 0.8840920\ttest: 0.8840920\ttest1: 0.8575084\tbest: 0.8579179 (261)\ttotal: 21.8s\tremaining: 9.89s\n",
      "275:\tlearn: 0.8840580\ttest: 0.8840580\ttest1: 0.8576108\tbest: 0.8579179 (261)\ttotal: 21.8s\tremaining: 9.8s\n",
      "276:\tlearn: 0.8840920\ttest: 0.8840920\ttest1: 0.8575084\tbest: 0.8579179 (261)\ttotal: 21.9s\tremaining: 9.71s\n",
      "277:\tlearn: 0.8840920\ttest: 0.8840920\ttest1: 0.8574061\tbest: 0.8579179 (261)\ttotal: 22s\tremaining: 9.63s\n",
      "278:\tlearn: 0.8842621\ttest: 0.8842621\ttest1: 0.8576108\tbest: 0.8579179 (261)\ttotal: 22s\tremaining: 9.55s\n",
      "279:\tlearn: 0.8843301\ttest: 0.8843301\ttest1: 0.8577132\tbest: 0.8579179 (261)\ttotal: 22.1s\tremaining: 9.46s\n",
      "280:\tlearn: 0.8842961\ttest: 0.8842961\ttest1: 0.8576108\tbest: 0.8579179 (261)\ttotal: 22.1s\tremaining: 9.37s\n",
      "281:\tlearn: 0.8842281\ttest: 0.8842281\ttest1: 0.8576108\tbest: 0.8579179 (261)\ttotal: 22.2s\tremaining: 9.28s\n",
      "282:\tlearn: 0.8839219\ttest: 0.8839219\ttest1: 0.8577132\tbest: 0.8579179 (261)\ttotal: 22.2s\tremaining: 9.19s\n",
      "283:\tlearn: 0.8839219\ttest: 0.8839219\ttest1: 0.8579179\tbest: 0.8579179 (261)\ttotal: 22.3s\tremaining: 9.1s\n",
      "284:\tlearn: 0.8840240\ttest: 0.8840240\ttest1: 0.8576108\tbest: 0.8579179 (261)\ttotal: 22.3s\tremaining: 9.02s\n",
      "285:\tlearn: 0.8841260\ttest: 0.8841260\ttest1: 0.8578155\tbest: 0.8579179 (261)\ttotal: 22.4s\tremaining: 8.93s\n",
      "286:\tlearn: 0.8842961\ttest: 0.8842961\ttest1: 0.8579179\tbest: 0.8579179 (261)\ttotal: 22.5s\tremaining: 8.84s\n",
      "287:\tlearn: 0.8846023\ttest: 0.8846023\ttest1: 0.8584297\tbest: 0.8584297 (287)\ttotal: 22.5s\tremaining: 8.76s\n",
      "288:\tlearn: 0.8846023\ttest: 0.8846023\ttest1: 0.8581226\tbest: 0.8584297 (287)\ttotal: 22.6s\tremaining: 8.67s\n",
      "289:\tlearn: 0.8849085\ttest: 0.8849085\ttest1: 0.8584297\tbest: 0.8584297 (287)\ttotal: 22.6s\tremaining: 8.58s\n",
      "290:\tlearn: 0.8849085\ttest: 0.8849085\ttest1: 0.8584297\tbest: 0.8584297 (287)\ttotal: 22.7s\tremaining: 8.5s\n",
      "291:\tlearn: 0.8850446\ttest: 0.8850446\ttest1: 0.8581226\tbest: 0.8584297 (287)\ttotal: 22.7s\tremaining: 8.41s\n",
      "292:\tlearn: 0.8850105\ttest: 0.8850105\ttest1: 0.8582250\tbest: 0.8584297 (287)\ttotal: 22.8s\tremaining: 8.32s\n",
      "293:\tlearn: 0.8849085\ttest: 0.8849085\ttest1: 0.8582250\tbest: 0.8584297 (287)\ttotal: 22.8s\tremaining: 8.24s\n",
      "294:\tlearn: 0.8848745\ttest: 0.8848745\ttest1: 0.8583274\tbest: 0.8584297 (287)\ttotal: 22.9s\tremaining: 8.15s\n",
      "295:\tlearn: 0.8850105\ttest: 0.8850105\ttest1: 0.8584297\tbest: 0.8584297 (287)\ttotal: 23s\tremaining: 8.07s\n",
      "296:\tlearn: 0.8850786\ttest: 0.8850786\ttest1: 0.8582250\tbest: 0.8584297 (287)\ttotal: 23s\tremaining: 7.98s\n",
      "297:\tlearn: 0.8851806\ttest: 0.8851806\ttest1: 0.8582250\tbest: 0.8584297 (287)\ttotal: 23.1s\tremaining: 7.9s\n",
      "298:\tlearn: 0.8851806\ttest: 0.8851806\ttest1: 0.8583274\tbest: 0.8584297 (287)\ttotal: 23.1s\tremaining: 7.81s\n",
      "299:\tlearn: 0.8851806\ttest: 0.8851806\ttest1: 0.8582250\tbest: 0.8584297 (287)\ttotal: 23.2s\tremaining: 7.72s\n",
      "300:\tlearn: 0.8851806\ttest: 0.8851806\ttest1: 0.8581226\tbest: 0.8584297 (287)\ttotal: 23.2s\tremaining: 7.64s\n",
      "301:\tlearn: 0.8852487\ttest: 0.8852487\ttest1: 0.8583274\tbest: 0.8584297 (287)\ttotal: 23.3s\tremaining: 7.56s\n",
      "302:\tlearn: 0.8851806\ttest: 0.8851806\ttest1: 0.8582250\tbest: 0.8584297 (287)\ttotal: 23.3s\tremaining: 7.47s\n",
      "303:\tlearn: 0.8852147\ttest: 0.8852147\ttest1: 0.8584297\tbest: 0.8584297 (287)\ttotal: 23.4s\tremaining: 7.39s\n",
      "304:\tlearn: 0.8852827\ttest: 0.8852827\ttest1: 0.8580203\tbest: 0.8584297 (287)\ttotal: 23.5s\tremaining: 7.31s\n",
      "305:\tlearn: 0.8851806\ttest: 0.8851806\ttest1: 0.8577132\tbest: 0.8584297 (287)\ttotal: 23.5s\tremaining: 7.23s\n",
      "306:\tlearn: 0.8852147\ttest: 0.8852147\ttest1: 0.8578155\tbest: 0.8584297 (287)\ttotal: 23.6s\tremaining: 7.14s\n",
      "307:\tlearn: 0.8852827\ttest: 0.8852827\ttest1: 0.8578155\tbest: 0.8584297 (287)\ttotal: 23.6s\tremaining: 7.06s\n",
      "308:\tlearn: 0.8853167\ttest: 0.8853167\ttest1: 0.8578155\tbest: 0.8584297 (287)\ttotal: 23.7s\tremaining: 6.98s\n",
      "309:\tlearn: 0.8853167\ttest: 0.8853167\ttest1: 0.8579179\tbest: 0.8584297 (287)\ttotal: 23.8s\tremaining: 6.9s\n",
      "310:\tlearn: 0.8854188\ttest: 0.8854188\ttest1: 0.8577132\tbest: 0.8584297 (287)\ttotal: 23.8s\tremaining: 6.82s\n",
      "311:\tlearn: 0.8852827\ttest: 0.8852827\ttest1: 0.8579179\tbest: 0.8584297 (287)\ttotal: 23.9s\tremaining: 6.73s\n",
      "312:\tlearn: 0.8850446\ttest: 0.8850446\ttest1: 0.8577132\tbest: 0.8584297 (287)\ttotal: 23.9s\tremaining: 6.65s\n",
      "313:\tlearn: 0.8851466\ttest: 0.8851466\ttest1: 0.8578155\tbest: 0.8584297 (287)\ttotal: 24s\tremaining: 6.57s\n",
      "314:\tlearn: 0.8854868\ttest: 0.8854868\ttest1: 0.8579179\tbest: 0.8584297 (287)\ttotal: 24s\tremaining: 6.49s\n",
      "315:\tlearn: 0.8856229\ttest: 0.8856229\ttest1: 0.8576108\tbest: 0.8584297 (287)\ttotal: 24.1s\tremaining: 6.41s\n",
      "316:\tlearn: 0.8856569\ttest: 0.8856569\ttest1: 0.8577132\tbest: 0.8584297 (287)\ttotal: 24.1s\tremaining: 6.32s\n",
      "317:\tlearn: 0.8856569\ttest: 0.8856569\ttest1: 0.8579179\tbest: 0.8584297 (287)\ttotal: 24.2s\tremaining: 6.25s\n",
      "318:\tlearn: 0.8856229\ttest: 0.8856229\ttest1: 0.8578155\tbest: 0.8584297 (287)\ttotal: 24.3s\tremaining: 6.17s\n",
      "319:\tlearn: 0.8853508\ttest: 0.8853508\ttest1: 0.8577132\tbest: 0.8584297 (287)\ttotal: 24.3s\tremaining: 6.08s\n",
      "320:\tlearn: 0.8854868\ttest: 0.8854868\ttest1: 0.8577132\tbest: 0.8584297 (287)\ttotal: 24.4s\tremaining: 6s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "321:\tlearn: 0.8853508\ttest: 0.8853508\ttest1: 0.8574061\tbest: 0.8584297 (287)\ttotal: 24.5s\tremaining: 5.92s\n",
      "322:\tlearn: 0.8852147\ttest: 0.8852147\ttest1: 0.8575084\tbest: 0.8584297 (287)\ttotal: 24.5s\tremaining: 5.84s\n",
      "323:\tlearn: 0.8851466\ttest: 0.8851466\ttest1: 0.8573037\tbest: 0.8584297 (287)\ttotal: 24.6s\tremaining: 5.76s\n",
      "324:\tlearn: 0.8851126\ttest: 0.8851126\ttest1: 0.8573037\tbest: 0.8584297 (287)\ttotal: 24.6s\tremaining: 5.68s\n",
      "325:\tlearn: 0.8853848\ttest: 0.8853848\ttest1: 0.8578155\tbest: 0.8584297 (287)\ttotal: 24.7s\tremaining: 5.61s\n",
      "326:\tlearn: 0.8853167\ttest: 0.8853167\ttest1: 0.8578155\tbest: 0.8584297 (287)\ttotal: 24.7s\tremaining: 5.52s\n",
      "327:\tlearn: 0.8853167\ttest: 0.8853167\ttest1: 0.8580203\tbest: 0.8584297 (287)\ttotal: 24.8s\tremaining: 5.44s\n",
      "Stopped by overfitting detector  (40 iterations wait)\n",
      "\n",
      "bestTest = 0.8584297267\n",
      "bestIteration = 287\n",
      "\n",
      "Shrink model to first 288 iterations.\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<catboost.core.CatBoostClassifier at 0x1a2d1e95f8>"
      ]
     },
     "execution_count": 154,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "catb_smote.fit(X_train_smt,y_train_smt,eval_set=[(X_train_smt,y_train_smt), (X_test,y_test)],early_stopping_rounds=40)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 155,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.92      0.90      0.91      7429\n",
      "           1       0.69      0.74      0.71      2340\n",
      "\n",
      "    accuracy                           0.86      9769\n",
      "   macro avg       0.80      0.82      0.81      9769\n",
      "weighted avg       0.86      0.86      0.86      9769\n",
      "\n",
      "Accuracy on test data: 85.842973%\n",
      "Accuracy on training data: 88.460230%\n",
      "Area under the ROC curve : 0.843451\n"
     ]
    }
   ],
   "source": [
    "print(metrics.classification_report(y_test, catb_smote.predict(X_test)))\n",
    "print('Accuracy on test data: %f%%' % (metrics.accuracy_score(y_test, catb_smote.predict(X_test))*100))\n",
    "print('Accuracy on training data: %f%%' % (metrics.accuracy_score(y_train_smt, catb_smote.predict(X_train_smt))*100))\n",
    "print('Area under the ROC curve : %f' % (metrics.roc_auc_score(y_test, catb_ros.predict(X_test))))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### SMOTE with XGBoost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 159,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0]\tvalidation_0-error:0.190787\tvalidation_1-error:0.199406\n",
      "Multiple eval metrics have been passed: 'validation_1-error' will be used for early stopping.\n",
      "\n",
      "Will train until validation_1-error hasn't improved in 30 rounds.\n",
      "[1]\tvalidation_0-error:0.191468\tvalidation_1-error:0.19828\n",
      "[2]\tvalidation_0-error:0.190447\tvalidation_1-error:0.197666\n",
      "[3]\tvalidation_0-error:0.189392\tvalidation_1-error:0.198076\n",
      "[4]\tvalidation_0-error:0.188542\tvalidation_1-error:0.197052\n",
      "[5]\tvalidation_0-error:0.187113\tvalidation_1-error:0.197359\n",
      "[6]\tvalidation_0-error:0.176771\tvalidation_1-error:0.183437\n",
      "[7]\tvalidation_0-error:0.176124\tvalidation_1-error:0.178831\n",
      "[8]\tvalidation_0-error:0.174117\tvalidation_1-error:0.184052\n",
      "[9]\tvalidation_0-error:0.169082\tvalidation_1-error:0.176784\n",
      "[10]\tvalidation_0-error:0.16847\tvalidation_1-error:0.176579\n",
      "[11]\tvalidation_0-error:0.168232\tvalidation_1-error:0.176477\n",
      "[12]\tvalidation_0-error:0.169184\tvalidation_1-error:0.176886\n",
      "[13]\tvalidation_0-error:0.168096\tvalidation_1-error:0.176374\n",
      "[14]\tvalidation_0-error:0.167619\tvalidation_1-error:0.175453\n",
      "[15]\tvalidation_0-error:0.166463\tvalidation_1-error:0.174429\n",
      "[16]\tvalidation_0-error:0.166531\tvalidation_1-error:0.176374\n",
      "[17]\tvalidation_0-error:0.166565\tvalidation_1-error:0.176272\n",
      "[18]\tvalidation_0-error:0.164285\tvalidation_1-error:0.176374\n",
      "[19]\tvalidation_0-error:0.16306\tvalidation_1-error:0.174839\n",
      "[20]\tvalidation_0-error:0.1617\tvalidation_1-error:0.169618\n",
      "[21]\tvalidation_0-error:0.161393\tvalidation_1-error:0.172894\n",
      "[22]\tvalidation_0-error:0.159931\tvalidation_1-error:0.17187\n",
      "[23]\tvalidation_0-error:0.158808\tvalidation_1-error:0.17013\n",
      "[24]\tvalidation_0-error:0.15976\tvalidation_1-error:0.171154\n",
      "[25]\tvalidation_0-error:0.158502\tvalidation_1-error:0.16798\n",
      "[26]\tvalidation_0-error:0.157039\tvalidation_1-error:0.166854\n",
      "[27]\tvalidation_0-error:0.156222\tvalidation_1-error:0.165933\n",
      "[28]\tvalidation_0-error:0.15493\tvalidation_1-error:0.164295\n",
      "[29]\tvalidation_0-error:0.153365\tvalidation_1-error:0.163886\n",
      "[30]\tvalidation_0-error:0.153229\tvalidation_1-error:0.162657\n",
      "[31]\tvalidation_0-error:0.152378\tvalidation_1-error:0.161736\n",
      "[32]\tvalidation_0-error:0.153535\tvalidation_1-error:0.165217\n",
      "[33]\tvalidation_0-error:0.152548\tvalidation_1-error:0.163783\n",
      "[34]\tvalidation_0-error:0.151017\tvalidation_1-error:0.158972\n",
      "[35]\tvalidation_0-error:0.149554\tvalidation_1-error:0.159279\n",
      "[36]\tvalidation_0-error:0.149078\tvalidation_1-error:0.157949\n",
      "[37]\tvalidation_0-error:0.146492\tvalidation_1-error:0.157949\n",
      "[38]\tvalidation_0-error:0.146084\tvalidation_1-error:0.157949\n",
      "[39]\tvalidation_0-error:0.14588\tvalidation_1-error:0.157232\n",
      "[40]\tvalidation_0-error:0.145268\tvalidation_1-error:0.15672\n",
      "[41]\tvalidation_0-error:0.143567\tvalidation_1-error:0.155594\n",
      "[42]\tvalidation_0-error:0.142852\tvalidation_1-error:0.155492\n",
      "[43]\tvalidation_0-error:0.142818\tvalidation_1-error:0.15539\n",
      "[44]\tvalidation_0-error:0.14224\tvalidation_1-error:0.15539\n",
      "[45]\tvalidation_0-error:0.141832\tvalidation_1-error:0.155185\n",
      "[46]\tvalidation_0-error:0.141832\tvalidation_1-error:0.154571\n",
      "[47]\tvalidation_0-error:0.140539\tvalidation_1-error:0.153035\n",
      "[48]\tvalidation_0-error:0.141049\tvalidation_1-error:0.153752\n",
      "[49]\tvalidation_0-error:0.140233\tvalidation_1-error:0.153035\n",
      "[50]\tvalidation_0-error:0.140029\tvalidation_1-error:0.15283\n",
      "[51]\tvalidation_0-error:0.13928\tvalidation_1-error:0.151704\n",
      "[52]\tvalidation_0-error:0.139348\tvalidation_1-error:0.150885\n",
      "[53]\tvalidation_0-error:0.138736\tvalidation_1-error:0.150578\n",
      "[54]\tvalidation_0-error:0.138566\tvalidation_1-error:0.150271\n",
      "[55]\tvalidation_0-error:0.138702\tvalidation_1-error:0.150067\n",
      "[56]\tvalidation_0-error:0.137579\tvalidation_1-error:0.149555\n",
      "[57]\tvalidation_0-error:0.137613\tvalidation_1-error:0.149759\n",
      "[58]\tvalidation_0-error:0.137341\tvalidation_1-error:0.150169\n",
      "[59]\tvalidation_0-error:0.137171\tvalidation_1-error:0.150169\n",
      "[60]\tvalidation_0-error:0.136661\tvalidation_1-error:0.149452\n",
      "[61]\tvalidation_0-error:0.136627\tvalidation_1-error:0.149145\n",
      "[62]\tvalidation_0-error:0.13615\tvalidation_1-error:0.148736\n",
      "[63]\tvalidation_0-error:0.135878\tvalidation_1-error:0.148736\n",
      "[64]\tvalidation_0-error:0.135436\tvalidation_1-error:0.148122\n",
      "[65]\tvalidation_0-error:0.134891\tvalidation_1-error:0.148429\n",
      "[66]\tvalidation_0-error:0.134653\tvalidation_1-error:0.148122\n",
      "[67]\tvalidation_0-error:0.134551\tvalidation_1-error:0.148019\n",
      "[68]\tvalidation_0-error:0.134381\tvalidation_1-error:0.147098\n",
      "[69]\tvalidation_0-error:0.13319\tvalidation_1-error:0.14587\n",
      "[70]\tvalidation_0-error:0.132952\tvalidation_1-error:0.146484\n",
      "[71]\tvalidation_0-error:0.131898\tvalidation_1-error:0.146279\n",
      "[72]\tvalidation_0-error:0.131285\tvalidation_1-error:0.145767\n",
      "[73]\tvalidation_0-error:0.130877\tvalidation_1-error:0.145051\n",
      "[74]\tvalidation_0-error:0.130537\tvalidation_1-error:0.14587\n",
      "[75]\tvalidation_0-error:0.130197\tvalidation_1-error:0.146177\n",
      "[76]\tvalidation_0-error:0.130129\tvalidation_1-error:0.145665\n",
      "[77]\tvalidation_0-error:0.12972\tvalidation_1-error:0.145972\n",
      "[78]\tvalidation_0-error:0.128598\tvalidation_1-error:0.14587\n",
      "[79]\tvalidation_0-error:0.128223\tvalidation_1-error:0.146074\n",
      "[80]\tvalidation_0-error:0.127849\tvalidation_1-error:0.146279\n",
      "[81]\tvalidation_0-error:0.127713\tvalidation_1-error:0.14587\n",
      "[82]\tvalidation_0-error:0.127407\tvalidation_1-error:0.145665\n",
      "[83]\tvalidation_0-error:0.127407\tvalidation_1-error:0.145665\n",
      "[84]\tvalidation_0-error:0.127067\tvalidation_1-error:0.145358\n",
      "[85]\tvalidation_0-error:0.127237\tvalidation_1-error:0.14546\n",
      "[86]\tvalidation_0-error:0.126931\tvalidation_1-error:0.145562\n",
      "[87]\tvalidation_0-error:0.126693\tvalidation_1-error:0.146279\n",
      "[88]\tvalidation_0-error:0.126931\tvalidation_1-error:0.145767\n",
      "[89]\tvalidation_0-error:0.126556\tvalidation_1-error:0.145153\n",
      "[90]\tvalidation_0-error:0.126454\tvalidation_1-error:0.146074\n",
      "[91]\tvalidation_0-error:0.126114\tvalidation_1-error:0.146177\n",
      "[92]\tvalidation_0-error:0.12608\tvalidation_1-error:0.14546\n",
      "[93]\tvalidation_0-error:0.125842\tvalidation_1-error:0.145051\n",
      "[94]\tvalidation_0-error:0.125468\tvalidation_1-error:0.145255\n",
      "[95]\tvalidation_0-error:0.12523\tvalidation_1-error:0.14546\n",
      "[96]\tvalidation_0-error:0.124753\tvalidation_1-error:0.145153\n",
      "[97]\tvalidation_0-error:0.124617\tvalidation_1-error:0.145051\n",
      "[98]\tvalidation_0-error:0.124651\tvalidation_1-error:0.144539\n",
      "[99]\tvalidation_0-error:0.124345\tvalidation_1-error:0.144744\n",
      "[100]\tvalidation_0-error:0.124243\tvalidation_1-error:0.144232\n",
      "[101]\tvalidation_0-error:0.123801\tvalidation_1-error:0.144539\n",
      "[102]\tvalidation_0-error:0.123597\tvalidation_1-error:0.144129\n",
      "[103]\tvalidation_0-error:0.123393\tvalidation_1-error:0.144334\n",
      "[104]\tvalidation_0-error:0.122848\tvalidation_1-error:0.14372\n",
      "[105]\tvalidation_0-error:0.12261\tvalidation_1-error:0.143515\n",
      "[106]\tvalidation_0-error:0.122576\tvalidation_1-error:0.14372\n",
      "[107]\tvalidation_0-error:0.121998\tvalidation_1-error:0.14372\n",
      "[108]\tvalidation_0-error:0.121453\tvalidation_1-error:0.14372\n",
      "[109]\tvalidation_0-error:0.121249\tvalidation_1-error:0.143413\n",
      "[110]\tvalidation_0-error:0.121147\tvalidation_1-error:0.143618\n",
      "[111]\tvalidation_0-error:0.121011\tvalidation_1-error:0.143515\n",
      "[112]\tvalidation_0-error:0.120603\tvalidation_1-error:0.142901\n",
      "[113]\tvalidation_0-error:0.11999\tvalidation_1-error:0.142389\n",
      "[114]\tvalidation_0-error:0.119956\tvalidation_1-error:0.142082\n",
      "[115]\tvalidation_0-error:0.119956\tvalidation_1-error:0.142082\n",
      "[116]\tvalidation_0-error:0.119548\tvalidation_1-error:0.141877\n",
      "[117]\tvalidation_0-error:0.119412\tvalidation_1-error:0.14198\n",
      "[118]\tvalidation_0-error:0.11863\tvalidation_1-error:0.141877\n",
      "[119]\tvalidation_0-error:0.118766\tvalidation_1-error:0.141673\n",
      "[120]\tvalidation_0-error:0.118698\tvalidation_1-error:0.141673\n",
      "[121]\tvalidation_0-error:0.118187\tvalidation_1-error:0.142389\n",
      "[122]\tvalidation_0-error:0.117983\tvalidation_1-error:0.142594\n",
      "[123]\tvalidation_0-error:0.117643\tvalidation_1-error:0.142389\n",
      "[124]\tvalidation_0-error:0.117337\tvalidation_1-error:0.142082\n",
      "[125]\tvalidation_0-error:0.117235\tvalidation_1-error:0.14198\n",
      "[126]\tvalidation_0-error:0.117371\tvalidation_1-error:0.141877\n",
      "[127]\tvalidation_0-error:0.117337\tvalidation_1-error:0.141468\n",
      "[128]\tvalidation_0-error:0.116929\tvalidation_1-error:0.141673\n",
      "[129]\tvalidation_0-error:0.116895\tvalidation_1-error:0.141161\n",
      "[130]\tvalidation_0-error:0.11652\tvalidation_1-error:0.141161\n",
      "[131]\tvalidation_0-error:0.11652\tvalidation_1-error:0.140956\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[132]\tvalidation_0-error:0.116146\tvalidation_1-error:0.141058\n",
      "[133]\tvalidation_0-error:0.115738\tvalidation_1-error:0.141058\n",
      "[134]\tvalidation_0-error:0.115568\tvalidation_1-error:0.140751\n",
      "[135]\tvalidation_0-error:0.115806\tvalidation_1-error:0.141468\n",
      "[136]\tvalidation_0-error:0.11567\tvalidation_1-error:0.141263\n",
      "[137]\tvalidation_0-error:0.1155\tvalidation_1-error:0.141366\n",
      "[138]\tvalidation_0-error:0.115296\tvalidation_1-error:0.141366\n",
      "[139]\tvalidation_0-error:0.115126\tvalidation_1-error:0.141366\n",
      "[140]\tvalidation_0-error:0.114989\tvalidation_1-error:0.141263\n",
      "[141]\tvalidation_0-error:0.115126\tvalidation_1-error:0.141161\n",
      "[142]\tvalidation_0-error:0.114887\tvalidation_1-error:0.140751\n",
      "[143]\tvalidation_0-error:0.114683\tvalidation_1-error:0.140751\n",
      "[144]\tvalidation_0-error:0.114649\tvalidation_1-error:0.140547\n",
      "[145]\tvalidation_0-error:0.114377\tvalidation_1-error:0.140751\n",
      "[146]\tvalidation_0-error:0.114547\tvalidation_1-error:0.140547\n",
      "[147]\tvalidation_0-error:0.114275\tvalidation_1-error:0.141058\n",
      "[148]\tvalidation_0-error:0.114377\tvalidation_1-error:0.140854\n",
      "[149]\tvalidation_0-error:0.114173\tvalidation_1-error:0.140854\n",
      "[150]\tvalidation_0-error:0.113901\tvalidation_1-error:0.140547\n",
      "[151]\tvalidation_0-error:0.114003\tvalidation_1-error:0.140342\n",
      "[152]\tvalidation_0-error:0.113663\tvalidation_1-error:0.140547\n",
      "[153]\tvalidation_0-error:0.113833\tvalidation_1-error:0.140649\n",
      "[154]\tvalidation_0-error:0.113901\tvalidation_1-error:0.140649\n",
      "[155]\tvalidation_0-error:0.113799\tvalidation_1-error:0.140751\n",
      "[156]\tvalidation_0-error:0.113799\tvalidation_1-error:0.140649\n",
      "[157]\tvalidation_0-error:0.113595\tvalidation_1-error:0.140649\n",
      "[158]\tvalidation_0-error:0.113561\tvalidation_1-error:0.140649\n",
      "[159]\tvalidation_0-error:0.113322\tvalidation_1-error:0.140854\n",
      "[160]\tvalidation_0-error:0.113288\tvalidation_1-error:0.140854\n",
      "[161]\tvalidation_0-error:0.11322\tvalidation_1-error:0.141058\n",
      "[162]\tvalidation_0-error:0.112982\tvalidation_1-error:0.141161\n",
      "[163]\tvalidation_0-error:0.112846\tvalidation_1-error:0.141161\n",
      "[164]\tvalidation_0-error:0.112812\tvalidation_1-error:0.140956\n",
      "[165]\tvalidation_0-error:0.112676\tvalidation_1-error:0.140751\n",
      "[166]\tvalidation_0-error:0.111621\tvalidation_1-error:0.141058\n",
      "[167]\tvalidation_0-error:0.111826\tvalidation_1-error:0.141263\n",
      "[168]\tvalidation_0-error:0.111689\tvalidation_1-error:0.141058\n",
      "[169]\tvalidation_0-error:0.111621\tvalidation_1-error:0.141058\n",
      "[170]\tvalidation_0-error:0.111315\tvalidation_1-error:0.140956\n",
      "[171]\tvalidation_0-error:0.111179\tvalidation_1-error:0.141161\n",
      "[172]\tvalidation_0-error:0.111553\tvalidation_1-error:0.140956\n",
      "[173]\tvalidation_0-error:0.111655\tvalidation_1-error:0.141058\n",
      "[174]\tvalidation_0-error:0.111349\tvalidation_1-error:0.140854\n",
      "[175]\tvalidation_0-error:0.111553\tvalidation_1-error:0.140854\n",
      "[176]\tvalidation_0-error:0.111315\tvalidation_1-error:0.140342\n",
      "[177]\tvalidation_0-error:0.111315\tvalidation_1-error:0.140649\n",
      "[178]\tvalidation_0-error:0.111315\tvalidation_1-error:0.140649\n",
      "[179]\tvalidation_0-error:0.111077\tvalidation_1-error:0.140854\n",
      "[180]\tvalidation_0-error:0.110839\tvalidation_1-error:0.140444\n",
      "[181]\tvalidation_0-error:0.110567\tvalidation_1-error:0.140751\n",
      "Stopping. Best iteration:\n",
      "[151]\tvalidation_0-error:0.114003\tvalidation_1-error:0.140342\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=0.9,\n",
       "              colsample_bytree=0.9, gamma=0, learning_rate=0.1,\n",
       "              max_delta_step=0, max_depth=5, min_child_weight=1, missing=None,\n",
       "              n_estimators=400, n_jobs=1, nthread=None,\n",
       "              objective='binary:logistic', random_state=0, reg_alpha=0,\n",
       "              reg_lambda=1.5, scale_pos_weight=1, seed=10, silent=True,\n",
       "              subsample=0.9, verbosity=1)"
      ]
     },
     "execution_count": 159,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "param = {}\n",
    "param['learning_rate'] = 0.1\n",
    "param['verbosity'] = 1\n",
    "param['colsample_bylevel'] = 0.9\n",
    "param['colsample_bytree'] = 0.9\n",
    "param['subsample'] = 0.9\n",
    "param['reg_lambda']= 1.5\n",
    "param['max_depth'] = 5\n",
    "param['n_estimators'] = 400\n",
    "param['seed']=10\n",
    "xgb_smote= XGBClassifier(**param)\n",
    "xgb_smote.fit(X_train_smt, y_train_smt, eval_metric=['error'], eval_set=[(X_train_smt, y_train_smt),(X_test.values, y_test.values)],early_stopping_rounds=30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 166,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.92      0.89      0.91      7429\n",
      "           1       0.69      0.76      0.72      2340\n",
      "\n",
      "    accuracy                           0.86      9769\n",
      "   macro avg       0.80      0.83      0.81      9769\n",
      "weighted avg       0.87      0.86      0.86      9769\n",
      "\n",
      "Accuracy on test data: 85.965810%\n",
      "Accuracy on training data: 88.599714%\n",
      "Area under the ROC curve : 0.826490\n"
     ]
    }
   ],
   "source": [
    "print(metrics.classification_report(y_test, xgb_smote.predict(X_test.values)))\n",
    "print('Accuracy on test data: %f%%' % (metrics.accuracy_score(y_test, xgb_smote.predict(X_test.values))*100))\n",
    "print('Accuracy on training data: %f%%' % (metrics.accuracy_score(y_train_smt, xgb_smote.predict(X_train_smt))*100))\n",
    "print('Area under the ROC curve : %f' % (metrics.roc_auc_score(y_test, xgb_smote.predict(X_test.values))))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}