Untitled.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# we have two sets of data, labeled train data and unlabeled test data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "data=pd.read_csv('./ml5/train.csv',sep=';')\n",
    "data=data.set_index('id')\n",
    "\n",
    "X=data.iloc[:,:-1]\n",
    "y=data.iloc[:,-1]\n",
    "\n",
    "X_t=pd.read_csv('./ml5/test.csv',sep=';',index_col='id')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index([u'age', u'gender', u'height', u'weight', u'ap_hi', u'ap_lo',\n",
       "       u'cholesterol', u'gluc', u'smoke', u'alco', u'active', u'cardio'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.columns"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Check for data type, missing values, outliers, correlations etc..."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>gender</th>\n",
       "      <th>height</th>\n",
       "      <th>weight</th>\n",
       "      <th>ap_hi</th>\n",
       "      <th>ap_lo</th>\n",
       "      <th>cholesterol</th>\n",
       "      <th>gluc</th>\n",
       "      <th>smoke</th>\n",
       "      <th>alco</th>\n",
       "      <th>active</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>id</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>18393</td>\n",
       "      <td>2</td>\n",
       "      <td>168</td>\n",
       "      <td>62.0</td>\n",
       "      <td>110</td>\n",
       "      <td>80</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>20228</td>\n",
       "      <td>1</td>\n",
       "      <td>156</td>\n",
       "      <td>85.0</td>\n",
       "      <td>140</td>\n",
       "      <td>90</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>18857</td>\n",
       "      <td>1</td>\n",
       "      <td>165</td>\n",
       "      <td>64.0</td>\n",
       "      <td>130</td>\n",
       "      <td>70</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>17623</td>\n",
       "      <td>2</td>\n",
       "      <td>169</td>\n",
       "      <td>82.0</td>\n",
       "      <td>150</td>\n",
       "      <td>100</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>17474</td>\n",
       "      <td>1</td>\n",
       "      <td>156</td>\n",
       "      <td>56.0</td>\n",
       "      <td>100</td>\n",
       "      <td>60</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      age  gender  height  weight  ap_hi  ap_lo  cholesterol  gluc  smoke  \\\n",
       "id                                                                          \n",
       "0   18393       2     168    62.0    110     80            1     1      0   \n",
       "1   20228       1     156    85.0    140     90            3     1      0   \n",
       "2   18857       1     165    64.0    130     70            3     1      0   \n",
       "3   17623       2     169    82.0    150    100            1     1      0   \n",
       "4   17474       1     156    56.0    100     60            1     1      0   \n",
       "\n",
       "    alco  active  \n",
       "id                \n",
       "0      0       1  \n",
       "1      0       1  \n",
       "2      0       0  \n",
       "3      0       1  \n",
       "4      0       0  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "age              int64\n",
       "gender           int64\n",
       "height           int64\n",
       "weight         float64\n",
       "ap_hi            int64\n",
       "ap_lo            int64\n",
       "cholesterol      int64\n",
       "gluc             int64\n",
       "smoke            int64\n",
       "alco             int64\n",
       "active           int64\n",
       "dtype: object"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.dtypes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>gender</th>\n",
       "      <th>height</th>\n",
       "      <th>weight</th>\n",
       "      <th>ap_hi</th>\n",
       "      <th>ap_lo</th>\n",
       "      <th>cholesterol</th>\n",
       "      <th>gluc</th>\n",
       "      <th>smoke</th>\n",
       "      <th>alco</th>\n",
       "      <th>active</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>70000.000000</td>\n",
       "      <td>70000.000000</td>\n",
       "      <td>70000.000000</td>\n",
       "      <td>70000.000000</td>\n",
       "      <td>70000.000000</td>\n",
       "      <td>70000.000000</td>\n",
       "      <td>70000.000000</td>\n",
       "      <td>70000.000000</td>\n",
       "      <td>70000.000000</td>\n",
       "      <td>70000.000000</td>\n",
       "      <td>70000.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>19468.865814</td>\n",
       "      <td>1.349571</td>\n",
       "      <td>164.359229</td>\n",
       "      <td>74.205690</td>\n",
       "      <td>128.817286</td>\n",
       "      <td>96.630414</td>\n",
       "      <td>1.366871</td>\n",
       "      <td>1.226457</td>\n",
       "      <td>0.088129</td>\n",
       "      <td>0.053771</td>\n",
       "      <td>0.803729</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>2467.251667</td>\n",
       "      <td>0.476838</td>\n",
       "      <td>8.210126</td>\n",
       "      <td>14.395757</td>\n",
       "      <td>154.011419</td>\n",
       "      <td>188.472530</td>\n",
       "      <td>0.680250</td>\n",
       "      <td>0.572270</td>\n",
       "      <td>0.283484</td>\n",
       "      <td>0.225568</td>\n",
       "      <td>0.397179</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>10798.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>55.000000</td>\n",
       "      <td>10.000000</td>\n",
       "      <td>-150.000000</td>\n",
       "      <td>-70.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>17664.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>159.000000</td>\n",
       "      <td>65.000000</td>\n",
       "      <td>120.000000</td>\n",
       "      <td>80.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>19703.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>165.000000</td>\n",
       "      <td>72.000000</td>\n",
       "      <td>120.000000</td>\n",
       "      <td>80.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>21327.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>170.000000</td>\n",
       "      <td>82.000000</td>\n",
       "      <td>140.000000</td>\n",
       "      <td>90.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>23713.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>250.000000</td>\n",
       "      <td>200.000000</td>\n",
       "      <td>16020.000000</td>\n",
       "      <td>11000.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                age        gender        height        weight         ap_hi  \\\n",
       "count  70000.000000  70000.000000  70000.000000  70000.000000  70000.000000   \n",
       "mean   19468.865814      1.349571    164.359229     74.205690    128.817286   \n",
       "std     2467.251667      0.476838      8.210126     14.395757    154.011419   \n",
       "min    10798.000000      1.000000     55.000000     10.000000   -150.000000   \n",
       "25%    17664.000000      1.000000    159.000000     65.000000    120.000000   \n",
       "50%    19703.000000      1.000000    165.000000     72.000000    120.000000   \n",
       "75%    21327.000000      2.000000    170.000000     82.000000    140.000000   \n",
       "max    23713.000000      2.000000    250.000000    200.000000  16020.000000   \n",
       "\n",
       "              ap_lo   cholesterol          gluc         smoke          alco  \\\n",
       "count  70000.000000  70000.000000  70000.000000  70000.000000  70000.000000   \n",
       "mean      96.630414      1.366871      1.226457      0.088129      0.053771   \n",
       "std      188.472530      0.680250      0.572270      0.283484      0.225568   \n",
       "min      -70.000000      1.000000      1.000000      0.000000      0.000000   \n",
       "25%       80.000000      1.000000      1.000000      0.000000      0.000000   \n",
       "50%       80.000000      1.000000      1.000000      0.000000      0.000000   \n",
       "75%       90.000000      2.000000      1.000000      0.000000      0.000000   \n",
       "max    11000.000000      3.000000      3.000000      1.000000      1.000000   \n",
       "\n",
       "             active  \n",
       "count  70000.000000  \n",
       "mean       0.803729  \n",
       "std        0.397179  \n",
       "min        0.000000  \n",
       "25%        1.000000  \n",
       "50%        1.000000  \n",
       "75%        1.000000  \n",
       "max        1.000000  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Seems that height, weight, ap_hi and ap_lo data have some nonsense values... "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>gender</th>\n",
       "      <th>height</th>\n",
       "      <th>weight</th>\n",
       "      <th>ap_hi</th>\n",
       "      <th>ap_lo</th>\n",
       "      <th>cholesterol</th>\n",
       "      <th>gluc</th>\n",
       "      <th>smoke</th>\n",
       "      <th>alco</th>\n",
       "      <th>active</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>id</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>5</th>\n",
       "      <td>18888</td>\n",
       "      <td>1</td>\n",
       "      <td>154</td>\n",
       "      <td>85.0</td>\n",
       "      <td>130</td>\n",
       "      <td>80</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>19042</td>\n",
       "      <td>2</td>\n",
       "      <td>170</td>\n",
       "      <td>69.0</td>\n",
       "      <td>130</td>\n",
       "      <td>90</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>None</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>20432</td>\n",
       "      <td>1</td>\n",
       "      <td>160</td>\n",
       "      <td>70.0</td>\n",
       "      <td>120</td>\n",
       "      <td>75</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>18133</td>\n",
       "      <td>2</td>\n",
       "      <td>185</td>\n",
       "      <td>94.0</td>\n",
       "      <td>130</td>\n",
       "      <td>90</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>None</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>16093</td>\n",
       "      <td>2</td>\n",
       "      <td>170</td>\n",
       "      <td>76.0</td>\n",
       "      <td>120</td>\n",
       "      <td>80</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      age  gender  height  weight  ap_hi  ap_lo  cholesterol  gluc smoke alco  \\\n",
       "id                                                                              \n",
       "5   18888       1     154    85.0    130     80            1     1     0    0   \n",
       "6   19042       2     170    69.0    130     90            1     1  None    0   \n",
       "7   20432       1     160    70.0    120     75            1     1     0    0   \n",
       "10  18133       2     185    94.0    130     90            1     1  None    0   \n",
       "11  16093       2     170    76.0    120     80            1     1     0    0   \n",
       "\n",
       "   active  \n",
       "id         \n",
       "5       1  \n",
       "6       1  \n",
       "7       0  \n",
       "10      1  \n",
       "11      1  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_t.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "age              int64\n",
       "gender           int64\n",
       "height           int64\n",
       "weight         float64\n",
       "ap_hi            int64\n",
       "ap_lo            int64\n",
       "cholesterol      int64\n",
       "gluc             int64\n",
       "smoke           object\n",
       "alco            object\n",
       "active          object\n",
       "dtype: object"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_t.dtypes"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We have None that was not recognized as np.NaN so lets fix that"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "X_t=X_t.replace('None',np.nan)\n",
    "cols=['smoke','active','alco']\n",
    "for x in ['smoke','active','alco']:\n",
    "    X_t[x]=pd.to_numeric(X_t[x])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "age              int64\n",
       "gender           int64\n",
       "height           int64\n",
       "weight         float64\n",
       "ap_hi            int64\n",
       "ap_lo            int64\n",
       "cholesterol      int64\n",
       "gluc             int64\n",
       "smoke          float64\n",
       "alco           float64\n",
       "active         float64\n",
       "dtype: object"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_t.dtypes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>gender</th>\n",
       "      <th>height</th>\n",
       "      <th>weight</th>\n",
       "      <th>ap_hi</th>\n",
       "      <th>ap_lo</th>\n",
       "      <th>cholesterol</th>\n",
       "      <th>gluc</th>\n",
       "      <th>smoke</th>\n",
       "      <th>alco</th>\n",
       "      <th>active</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>30000.000000</td>\n",
       "      <td>30000.000000</td>\n",
       "      <td>30000.000000</td>\n",
       "      <td>30000.000000</td>\n",
       "      <td>30000.000000</td>\n",
       "      <td>30000.000000</td>\n",
       "      <td>30000.000000</td>\n",
       "      <td>30000.000000</td>\n",
       "      <td>26970.000000</td>\n",
       "      <td>26969.000000</td>\n",
       "      <td>27103.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>19455.528400</td>\n",
       "      <td>1.353733</td>\n",
       "      <td>164.322267</td>\n",
       "      <td>74.120615</td>\n",
       "      <td>130.271300</td>\n",
       "      <td>95.689000</td>\n",
       "      <td>1.368033</td>\n",
       "      <td>1.222700</td>\n",
       "      <td>0.087838</td>\n",
       "      <td>0.054470</td>\n",
       "      <td>0.805926</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>2477.840922</td>\n",
       "      <td>0.478136</td>\n",
       "      <td>8.213740</td>\n",
       "      <td>14.341540</td>\n",
       "      <td>212.501152</td>\n",
       "      <td>163.852598</td>\n",
       "      <td>0.679166</td>\n",
       "      <td>0.566494</td>\n",
       "      <td>0.283065</td>\n",
       "      <td>0.226947</td>\n",
       "      <td>0.395494</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>8865.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>50.000000</td>\n",
       "      <td>10.000000</td>\n",
       "      <td>-130.000000</td>\n",
       "      <td>-90.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>17658.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>159.000000</td>\n",
       "      <td>65.000000</td>\n",
       "      <td>120.000000</td>\n",
       "      <td>80.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>19694.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>165.000000</td>\n",
       "      <td>72.000000</td>\n",
       "      <td>120.000000</td>\n",
       "      <td>80.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>21320.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>170.000000</td>\n",
       "      <td>82.000000</td>\n",
       "      <td>140.000000</td>\n",
       "      <td>90.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>23705.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>198.000000</td>\n",
       "      <td>183.000000</td>\n",
       "      <td>16020.000000</td>\n",
       "      <td>9100.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                age        gender        height        weight         ap_hi  \\\n",
       "count  30000.000000  30000.000000  30000.000000  30000.000000  30000.000000   \n",
       "mean   19455.528400      1.353733    164.322267     74.120615    130.271300   \n",
       "std     2477.840922      0.478136      8.213740     14.341540    212.501152   \n",
       "min     8865.000000      1.000000     50.000000     10.000000   -130.000000   \n",
       "25%    17658.000000      1.000000    159.000000     65.000000    120.000000   \n",
       "50%    19694.000000      1.000000    165.000000     72.000000    120.000000   \n",
       "75%    21320.000000      2.000000    170.000000     82.000000    140.000000   \n",
       "max    23705.000000      2.000000    198.000000    183.000000  16020.000000   \n",
       "\n",
       "              ap_lo   cholesterol          gluc         smoke          alco  \\\n",
       "count  30000.000000  30000.000000  30000.000000  26970.000000  26969.000000   \n",
       "mean      95.689000      1.368033      1.222700      0.087838      0.054470   \n",
       "std      163.852598      0.679166      0.566494      0.283065      0.226947   \n",
       "min      -90.000000      1.000000      1.000000      0.000000      0.000000   \n",
       "25%       80.000000      1.000000      1.000000      0.000000      0.000000   \n",
       "50%       80.000000      1.000000      1.000000      0.000000      0.000000   \n",
       "75%       90.000000      2.000000      1.000000      0.000000      0.000000   \n",
       "max     9100.000000      3.000000      3.000000      1.000000      1.000000   \n",
       "\n",
       "             active  \n",
       "count  27103.000000  \n",
       "mean       0.805926  \n",
       "std        0.395494  \n",
       "min        0.000000  \n",
       "25%        1.000000  \n",
       "50%        1.000000  \n",
       "75%        1.000000  \n",
       "max        1.000000  "
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_t.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "age               0\n",
       "gender            0\n",
       "height            0\n",
       "weight            0\n",
       "ap_hi             0\n",
       "ap_lo             0\n",
       "cholesterol       0\n",
       "gluc              0\n",
       "smoke          3030\n",
       "alco           3031\n",
       "active         2897\n",
       "dtype: int64"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_t.isnull().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Test data have missing values in the last three columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "1d distribution plots"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZYAAAESCAYAAADe2fNYAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\npZUmFqWUUmn1/wMOcWRN72eSBAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f69c85b0710>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXYAAAESCAYAAADpO/4pAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f699096e510>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYAAAAESCAYAAAD0aQL3AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nlAJARCSi/j9HC8VvdWM7eAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f69909786d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAESCAYAAAASQMmzAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nMMYYM+//B9A02OU0l9gUAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f69905dd790>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYsAAAESCAYAAAAMifkAAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f6990431910>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYsAAAESCAYAAAAMifkAAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nmKz2AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f6990635650>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXYAAAESCAYAAADpO/4pAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRDyjYBcR8cz/B0mwhdhkQk8DAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f69905c1110>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXYAAAESCAYAAADpO/4pAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f699072e9d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXYAAAESCAYAAADpO/4pAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\npISx2CUpYSx2SUoYi12SEsZil6SE+f+rK/2aWv26XwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f6992a0ded0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXYAAAESCAYAAADpO/4pAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nS1LKWOySlDIWuySljMUuSSnz/wFMOtLuV8IPFwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f69901f7690>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXYAAAESCAYAAADpO/4pAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nrkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f6990316cd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "columns=X.columns\n",
    "for x in columns:\n",
    "    sns.distplot(X[x].dropna())\n",
    "    sns.plt.show()\n",
    "    \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAESCAYAAADnvkIDAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f69907321d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYcAAAESCAYAAAAWtRmOAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAExxJREFUeJzt3X2QXXV9x/H3wgYkDxCLW0LjiBQ7X0FkLNBCI09JCJGp\nxfI0zkAdgeJYRUopfUCnjJRaysCAiAp1UASZsaVYQNMKA4GooSLR0hmwg18VDR0JltUg2UASNntv\n/zhn6WZ/ZHdz703u3eX9msnk3nN/9+z3e+/d89nzcM/pazabSJI01m7dLkCS1HsMB0lSwXCQJBUM\nB0lSwXCQJBUMB0lSoX8qgyLiauAYYHfgKuC7wO1U4fIs8L7MHI6Is4GLgBHg5sy8JSL6gVuBA4Ct\nwLmZuTYiDgNuAhrA45l5QUc7kyS1bNI1h4g4ATgkMxcBJwPXA1cAn8nM44GngPMiYjZwGbAEWAxc\nHBHzgbOA5zPzWOBKqnChns+F9fT5EbG8o51Jklo2lc1K3wTOrG//CpgDHA98rZ62AlgGHAWsycyN\nmbkZeJhqbWMpcHc9diWwKCJmAQdm5mNj5nFim71Ikjpk0nDIzGZmbqrv/jHw78CczByupz0H7A/s\nBwyOeerg+OmZ2QSawAJg/Zixo/OQJPWAKe1zAIiI9wDnAScBPx7zUN92njLR9Oa4x7c3VpLUBVPd\nIb0c+CiwPDOHImIoIvbMzC3AQuAZYB3b/vW/EHiknr4AeKLeOd1HtRN733Fj101UQ7PZbPb1mSGS\ntINaWnBOGg4RsTdwNbA0M1+oJ68ETge+XP9/H7AG+Hw9vgEsojpyaR+qfRYPAKcAqzJzJCKejIhF\nmflt4DTghonq6OvrY3BwqIUWp4eBgXn2N43Z3/Q1k3uDqr9WTGXN4b1Uf+X/S0SMbhJ6P/CFiPgg\n8DRwW73AvxS4nyocLq/XMu4AlkXEamAzcE4934uBz9XzfDQzH2qpA0lSx/VNo1N2N2d6utvf9GV/\n09dM7g1gYGBeS5uV/Ia0JKlgOEiSClM+lFU7T6PR4N4Hv8PmLe0djTV7zz4OP+zgDlUl6bXMcOgR\nG1/eHWa1dlTBqJe2vDD5IEmaAjcrSZIKhoMkqWA4SJIKhoMkqWA4SJIKhoMkqWA4SJIKhoMkqWA4\nSJIKhoMkqWA4SJIKhoMkqWA4SJIKhoMkqWA4SJIKhoMkqWA4SJIKhoMkqWA4SJIKhoMkqWA4SJIK\nhoMkqWA4SJIKhoMkqWA4SJIKhoMkqWA4SJIKhoMkqWA4SJIKhoMkqWA4SJIKhoMkqWA4SJIKhoMk\nqWA4SJIKhoMkqWA4SJIKhoMkqWA4SJIKhoMkqWA4SJIKhoMkqWA4SJIK/VMZFBGHAvcA12XmjRHx\nReAI4Bf1kGsy896IOBu4CBgBbs7MWyKiH7gVOADYCpybmWsj4jDgJqABPJ6ZF3SyMUlS6yZdc4iI\n2cANwMpxD12amUvqf/fW4y4DlgCLgYsjYj5wFvB8Zh4LXAlcVT//euDCevr8iFjemZYkSe2aymal\nzcDJwLOTjDsKWJOZGzNzM/AwcAywFLi7HrMSWBQRs4ADM/OxevoK4MQdLV6StHNMulkpMxvAlogY\n/9BHIuIS4H+BC4EFwOCYxweB/YH9RqdnZjMimvXY9WPGPlePlST1gCntc3gVXwJ+mZmPR8RfAZcD\n3x43pm87z+0DmuMe397YbQwMzNvBMqeHRqMBwLy5r2trPrNGXu7p16iXa+sE+5u+ZnJvrWopHDJz\n1Zi7K4AbgTuBPxgzfSHwCLCOak3hiXrndB/VJqp9x41dN9nPHRwcaqXcnjcaDkMbN7c1n/7hTT37\nGg0MzOvZ2jrB/qavmdwbtB58LR3KGhFfiYgD67snAN8H1gBHRsTeETEXWASsBh4AzqzHngKsyswR\n4MmIWFRPPw24r6UOJEkdN+maQ0QcDlxLdSjqcEScAXwauCMiXgQ2Uh2eujkiLgXupzo89fLMHIqI\nO4BlEbGaauf2OfWsLwY+FxF9wKOZ+VCHe5Mktaiv2Wx2u4apas7UVb9Go8G3Hvsh9Le33bN/+AWO\n+Z1DOlRVZ70WVt3tb3qayb0BDAzMm9I+3fH8hrQkqWA4SJIKhoMkqWA4SJIKhoMkqWA4SJIKhoMk\nqWA4SJIKhoMkqWA4SJIKhoMkqWA4SJIKhoMkqWA4SJIKhoMkqWA4SJIKhoMkqWA4SJIKhoMkqWA4\nSJIKhoMkqWA4SJIKhoMkqWA4SJIKhoMkqWA4SJIKhoMkqWA4SJIKhoMkqWA4SJIKhoMkqWA4SJIK\nhoMkqWA4SJIKhoMkqWA4SJIKhoMkqWA4SJIKhoMkqWA4SJIKhoMkqWA4SJIKhoMkqdDf7QIkSaVm\ns8nQ0Ia25zMwMK+l5xkOktSDhoY28MCjP2av2XNanseml17koIPe2NJzDQdJ6lF7zZ7D7Dmt/eXf\nLvc5SJIKU1pziIhDgXuA6zLzxoh4I3A7Vbg8C7wvM4cj4mzgImAEuDkzb4mIfuBW4ABgK3BuZq6N\niMOAm4AG8HhmXtDh3iRJLZp0zSEiZgM3ACvHTL4C+HRmHg88BZxXj7sMWAIsBi6OiPnAWcDzmXks\ncCVwVT2P64EL6+nzI2J5h3qSJLVpKpuVNgMnU60hjDoBWFHfXgEsA44C1mTmxszcDDwMHAMsBe6u\nx64EFkXELODAzHxszDxObKMPSVIHTRoOmdnIzC3jJs/JzOH69nPA/sB+wOCYMYPjp2dmE2gCC4D1\nY8aOzkOS1AM6cbRSXwvTm+Me397YbbR6vG6vazQaAMyb+7q25jNr5OWefo16ubZOsL/pqxd722OP\nBnPnrGdOG8uF3Xi55ee2Gg5DEbFnvUaxEHgGWMe2f/0vBB6ppy8Anqh3TvdRbaLad9zYdZP90MHB\noRbL7W2j4TC0cXNb8+kf3tSzr9HAwLyera0T7G/66tXeNmwYYuOLW2jQ+nLhpRfHb/SZulYPZV0J\nnF7fPh24D1gDHBkRe0fEXGARsBp4ADizHnsKsCozR4AnI2JRPf20eh6SpB4w6ZpDRBwOXEt1KOpw\nRJwBnA3cFhEfBJ4GbsvMkYi4FLif6vDUyzNzKCLuAJZFxGqqndvn1LO+GPhcRPQBj2bmQx3uTZLU\noknDoT6iaPGrPHTSq4y9C7hr3LQGcN6rjH0SOG7KlUqSdhm/IS1JKhgOkqSC4SBJKhgOkqSC4SBJ\nKhgOkqSC4SBJKhgOkqSC4SBJKhgOkqSC4SBJKhgOkqSC4SBJKhgOkqSC4SBJKhgOkqSC4SBJKhgO\nkqSC4SBJKhgOkqSC4SBJKhgOkqSC4SBJKhgOkqSC4SBJKhgOkqSC4SBJKhgOkqSC4SBJKhgOkqSC\n4SBJKhgOkqSC4SBJKhgOkqSC4SBJKhgOkqSC4SBJKhgOkqSC4SBJKhgOkqSC4SBJKhgOkqSC4SBJ\nKhgOkqSC4SBJKhgOkqSC4SBJKvS38qSIOB64E/g+0Ac8DlwD3E4VOM8C78vM4Yg4G7gIGAFuzsxb\nIqIfuBU4ANgKnJuZa9trRZLUKe2sOXwjM5dk5uLMvAi4Avh0Zh4PPAWcFxGzgcuAJcBi4OKImA+c\nBTyfmccCVwJXtdWFJKmj2gmHvnH3TwBW1LdXAMuAo4A1mbkxMzcDDwPHAEuBu+uxK4F3tlGHJKnD\n2gmHQyLinoj4VkScCMzOzOH6seeA/YH9gMExzxkcPz0zm0Cj3tQkSeoBrS6QfwRcnpl3RsRvAqvG\nzWv8WsVk06cUUgMD86Ze4TTSaDQAmDf3dW3NZ9bIyz39GvVybZ1gf9NXL/a2xx4N5s5Zz5w2lgu7\n8XLLz20pHDJzHdUOaTLzJxHxc+DIiNgzM7cAC4FngHVUawqjFgKP1NMXAE+MrjFk5tbJfu7g4FAr\n5fa80XAY2ri5rfn0D2/q2ddoYGBez9bWCfY3ffVqbxs2DLHxxS00aH258NKLW1p+bkublSLirIi4\npL69gGoz0ReBM+ohpwP3AWuoQmPviJgLLAJWAw8AZ9ZjT6Fa85Ak9YhW9zl8DTg+Ir5FtWP5g8Df\nAO+PiG8Crwduq3dCXwrcX/+7PDOHgDuA/ohYDXwI+Gh7bUiSOqnVzUobqf7iH++kVxl7F3DXuGkN\n4LxWfrYkaefzG9KSpILhIEkqGA6SpILhIEkqGA6SpILhIEkqGA6SpILhIEkqGA6SpILhIEkqGA6S\npILhIEkqGA6SpILhIEkqGA6SpILhIEkqGA6SpILhIEkqGA6SpILhIEkqGA6SpILhIEkqGA6SpILh\nIEkqGA6SpILhIEkqGA6SpILhIEkqGA6SpILhIEkqGA6SpILhIEkqGA6SpILhIEkqGA6SpILhIEkq\nGA6SpILhIEkqGA6SpILhIEkqGA6SpILhIEkqGA6SpILhIEkqGA6SpILhIEkq9Hfzh0fEdcDRQAP4\ns8z8XjfrkSRVurbmEBHHAW/JzEXA+cAN3apFkrStbm5WWgrcA5CZPwDmR8TcLtYjSap1MxwWAINj\n7v+iniZJ6rKu7nMYp6/bBXTT8OYNDI9samsec/eADRte6FBFnbXHHg02bBjqdhk7jf1NX73a29DQ\nBja99GJb82jn+d0Mh3Vsu6bwG8CzE4zvGxiYt3Mr6qKzTl3a7RJ2un322afbJexU9jd99Wpv73jH\nIV372d3crHQ/cAZARBwOPJOZ7cWkJKkj+prNZtd+eERcCRwPjAAXZOYTXStGkvSKroaDJKk3+Q1p\nSVLBcJAkFQwHSVKhl77n8IqJzrkUERcAZwNbge9l5p93p8rWRcShVN8Ovy4zbxz32InA31P1d29m\nfqILJbZlkv4WA1dS9ZeZeX4XSmzZRL2NGfMPwNGZuXiXFtcBk7x3bwT+CZgFPJaZH+5CiW2ZpL+Z\nsGy5GjgG2B24KjPvHvPYDi1bem7NYaJzLkXEPOAvgHdm5nHA2yLid7tTaWsiYjZVTyu3M+RTwKlU\nb/BJEfHWXVVbJ0yhv38ETsvMY4G9I+Jdu6y4Nk2hNyLiYOBYYNod6TGF/q4FrsnMo4GROiymjYn6\nmyHLlhOAQ+pl58nA9eOG7NCypefCgYnPufQysIVqodIP7AWs70qVrdtM9cYVX/iLiAOBX2bmusxs\nAl+nej2mk+32VzsiM0cfGwT23SVVdcZkvUG1AP3Yrimn4yb6bPZRLVRWAGTmhZn5s11bXtsmev9m\nwrLlm8CZ9e1fAbPr962lZUsvhsN2z7mUmVuAK4CfAD8FHs3MH+/yCtuQmY26j1czvvfngP13flWd\nM0l/ZOZGgIjYH1hG9SGdFibrLSLeD6wCnt51VXXOJP0NABuB6yNidf0dpWllov5myLKlmZmj5+A5\nH/h6HQTQwrKlF8NhvFfOuVSv+n0MeAtwIHB0RLy9W4XtAjPyfFMR8evA14APZebz3a6nEyLi9cC5\nwHVU79tMe+/6gIXAJ6m+uPrbEXFyd0vqnJm0bImI91B9Fj8ywbBJP5+9GA4TnXPpYOCpzHw+M7cC\nq4EjdnF9O9M6tk3zhfW0GaP+Jfw68LHMfLDb9XTQEuANVJ/Ju6gWntd2t6SO+gWwNjPXZmYDeBB4\nW5dr6qQZsWyJiOXAR4F3ZebYswnu8LKlF8NhonMurQUOjog96/tHAj/a5RV2zjbpnZlPA/Mi4k31\nds93U70e09Wr/XVyHdWRIg/s6mI6bPx796+ZeWi9M/BUqqN5LulOaR0xvr8R4CcRcVA96Qggd3lV\nnTP+s7mWab5siYi9gauBd2fmNqdnbmXZ0pOnzxh/ziXgcOBXmfnViPgAcB4wDHw7My/tXqU7rg68\na4EDqHp4hmoTy0/r/o6heoObwFcy85NdK7YFE/VH9WFcDzxC9cvZBL6cmZ/vTrU7ZrL3bsy4A4Av\nZuaSrhTaoil8Ng8CbqV6757IzA91q9ZWTKG/6b5s+QDwceCH/P/v10NU79UOL1t6MhwkSd3Vi5uV\nJEldZjhIkgqGgySpYDhIkgqGgySpYDhIkgqGg7QTRMRBEfHTbtchtcpwkHYev0SkaasnL/Yj7WoR\n8VngKKrzeP2M6lxCD1J94xSqb81+IDOfrtcIPkV1+uc3A3+SmasiYhFwE9UZLx8bM+/5VNexeAOw\nD3BtZv5zRHyc6iRvbwIuycz/2umNSlPkmoNe8yJiKXBkZh4JvJfqPPezqRbop9ZXdPsM1akXRr2U\nmcuprqz1p/W0a4C/zMxlwM/HjP0E1ZW3TqQ6LczfRcTodSzenJlLDAb1GtccJHgH1Vk4ycyXIuI+\n4K1UZwe+q75gym5Ul60d9Y36/6eBX6tvvx34j/r2Q8CF9e3FwJERcU59fwvVGgPAdzrZiNQphoNU\nLvhHqBbg/zPByfO2jrndN+b/0fnsPubxLcCHM/OxMdOIiN+nugKZ1HPcrCTBD4Cj4ZXrDC8HngLe\nEBFvq6cfFxHnTzKf/wZ+r769bMz0h6k2VxERe0XEZyPC3z31ND+gUnXxoZ9FxHeB26k2DQ0BfwR8\nISJWAX9LdY1e2P5RSH8NfCoi/g2YO2b65cBvRcRqqs1R/1lfMEfqWZ6yW6959UVS/jAzv1Tf/yrV\ndSbu6G5lUve4z0Gq1hLeGREXAZuornB2Z3dLkrrLNQdJUsF9DpKkguEgSSoYDpKkguEgSSoYDpKk\nguEgSSr8H/yi6BqTgHX5AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f6990749590>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYMAAAESCAYAAAAfXrn0AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nJAwDSRLw/wGWtlPExWR8jAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f6990836e50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYMAAAESCAYAAAAfXrn0AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nDANJEvD/ARe8qFgiJjP9AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f699071b450>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAESCAYAAADqoDJEAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nkooYHJKkIgaHJKnI/wO7HE9PQuVkDwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f6990894390>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAESCAYAAADqoDJEAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nihgckqQiBockqYjBIUkq8v8ByukzU+tNlm8AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f6992adef50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYcAAAESCAYAAAAWtRmOAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nPy1kfr2ToPJNAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f6990012050>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYcAAAESCAYAAAAWtRmOAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f6990990e90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYcAAAESCAYAAAAWtRmOAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f6990214c50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYcAAAESCAYAAAAWtRmOAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f69901f6e10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYcAAAESCAYAAAAWtRmOAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nQZJUMBwkSYX/Bx6wgEFyxdaGAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f6990405750>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "for x in columns:\n",
    "    sns.distplot(X_t[x].dropna(),kde=False)\n",
    "    sns.plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<seaborn.axisgrid.PairGrid at 0x7f6990386550>"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXYAAAFgCAYAAACv0VLjAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nkZhRYBcRiZn/H5p3OcFBF6Q6AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f69907dc9d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.pairplot(X[['ap_hi','ap_lo']])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "We confirm outliers in blod pressure..."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "checking correlations for train data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f699023c410>"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAElCAYAAAD+wXUWAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nImEggiAIgor8P5k5rjUqInlDAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f6990543690>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.heatmap(data.corr(), annot=True, fmt=\".2f\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# fixing data, filling missing values and nonsense outliers of blood pressure values using XGBoost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_total=X.append(X_t)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python2.7/dist-packages/sklearn/cross_validation.py:44: DeprecationWarning: This module was deprecated in version 0.18 in favor of the model_selection module into which all the refactored classes and functions are moved. Also note that the interface of the new CV iterators are different from that of this module. This module will be removed in 0.20.\n",
      "  \"This module will be removed in 0.20.\", DeprecationWarning)\n"
     ]
    }
   ],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.model_selection import RandomizedSearchCV,GridSearchCV\n",
    "import xgboost as xgb\n",
    "from xgboost import XGBClassifier\n",
    "\n",
    "from itertools import combinations"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "I first fill missing values of test data as prediction based on the other data, loop over all possible combination and do cross validation =3  grid search"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "iters=['smoke', 'alco', 'active']\n",
    "combs=[map(str,x) for x in combinations(iters,3)]+[map(str,x) for x in combinations(iters,2)]+[map(str,x) for x in combinations(iters,1)]\n",
    "fixed_data=X_total.copy()\n",
    "to_train=X_total.dropna()\n",
    "the_rest= X_total[X_total.isnull().any(axis=1)]\n",
    "for i in range(len(combs)):\n",
    "    to_train_X=to_train.drop(combs[i],axis=1)\n",
    "    to_train_y=to_train[combs[i]]\n",
    "    for l in range(to_train_y.shape[1]):\n",
    "        xgbc=xgb.XGBClassifier()\n",
    "        params={'n_estimators':[100],'max_depth':[3,5,8],'learning_rate':[0.1]}\n",
    "        clf=GridSearchCV(xgbc,params,cv=3)\n",
    "        col=combs[i][l]\n",
    "        to_fill_X=the_rest.drop(combs[i],axis=1)\n",
    "        to_fill_X=to_fill_X.dropna()\n",
    "        indx=to_fill_X.index\n",
    "        to_train_X_col=to_train_X\n",
    "        to_train_y_col=to_train[col]\n",
    "        clf.fit(to_train_X_col,list(to_train_y_col))\n",
    "        to_fill_y=clf.predict(to_fill_X)\n",
    "        fixed_data.loc[indx,col]=to_fill_y"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now lets identify outliers of blood pressure that dont make sense and try to predict them the same way we did for missing value"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "X_tot=fixed_data\n",
    "indx=X_tot[ (X_tot['ap_lo']>200) | (X_tot['ap_hi']>=500) | (X_tot['ap_lo']<20) | (X_tot['ap_hi']<50) |(X_tot['ap_hi']<=X_tot['ap_lo']) | (X_tot['ap_hi']<0) | (X_tot ['ap_lo']<0)].index\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_tr_all=X_tot.drop(indx,axis=0)\n",
    "y_tr_all=X_tr_all[['ap_lo','ap_hi']]\n",
    "X_tr_all.drop(['ap_lo','ap_hi'],axis=1,inplace=True)\n",
    "   \n",
    "xgbc1=xgb.XGBRegressor(n_estimators=10)\n",
    "xgbc2=xgb.XGBRegressor(n_estimators=10)\n",
    " \n",
    "xgbc1.fit(X_tr_all,list(y_tr_all['ap_lo']))\n",
    "xgbc2.fit(X_tr_all,list(y_tr_all['ap_hi']))\n",
    "helpr=(X_tot.loc[indx]).drop(['ap_lo','ap_hi'],axis=1)\n",
    "X_tot.loc[indx,'ap_lo']=xgbc1.predict(helpr)\n",
    "X_tot.loc[indx,'ap_hi']=xgbc2.predict(helpr)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>gender</th>\n",
       "      <th>height</th>\n",
       "      <th>weight</th>\n",
       "      <th>ap_hi</th>\n",
       "      <th>ap_lo</th>\n",
       "      <th>cholesterol</th>\n",
       "      <th>gluc</th>\n",
       "      <th>smoke</th>\n",
       "      <th>alco</th>\n",
       "      <th>active</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>100000.000000</td>\n",
       "      <td>100000.000000</td>\n",
       "      <td>100000.000000</td>\n",
       "      <td>100000.000000</td>\n",
       "      <td>100000.000000</td>\n",
       "      <td>100000.000000</td>\n",
       "      <td>100000.000000</td>\n",
       "      <td>100000.000000</td>\n",
       "      <td>100000.000000</td>\n",
       "      <td>100000.000000</td>\n",
       "      <td>100000.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>19464.864590</td>\n",
       "      <td>1.350820</td>\n",
       "      <td>164.348140</td>\n",
       "      <td>74.180167</td>\n",
       "      <td>125.889511</td>\n",
       "      <td>80.788696</td>\n",
       "      <td>1.367220</td>\n",
       "      <td>1.225330</td>\n",
       "      <td>0.082610</td>\n",
       "      <td>0.049450</td>\n",
       "      <td>0.820010</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>2470.428376</td>\n",
       "      <td>0.477229</td>\n",
       "      <td>8.211187</td>\n",
       "      <td>14.379494</td>\n",
       "      <td>17.555061</td>\n",
       "      <td>10.109677</td>\n",
       "      <td>0.679922</td>\n",
       "      <td>0.570543</td>\n",
       "      <td>0.275293</td>\n",
       "      <td>0.216807</td>\n",
       "      <td>0.384181</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>8865.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>50.000000</td>\n",
       "      <td>10.000000</td>\n",
       "      <td>60.000000</td>\n",
       "      <td>20.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>17662.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>159.000000</td>\n",
       "      <td>65.000000</td>\n",
       "      <td>120.000000</td>\n",
       "      <td>80.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>19700.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>165.000000</td>\n",
       "      <td>72.000000</td>\n",
       "      <td>120.000000</td>\n",
       "      <td>80.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>21324.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>170.000000</td>\n",
       "      <td>82.000000</td>\n",
       "      <td>140.000000</td>\n",
       "      <td>90.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>23713.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>250.000000</td>\n",
       "      <td>200.000000</td>\n",
       "      <td>401.000000</td>\n",
       "      <td>182.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 age         gender         height         weight  \\\n",
       "count  100000.000000  100000.000000  100000.000000  100000.000000   \n",
       "mean    19464.864590       1.350820     164.348140      74.180167   \n",
       "std      2470.428376       0.477229       8.211187      14.379494   \n",
       "min      8865.000000       1.000000      50.000000      10.000000   \n",
       "25%     17662.000000       1.000000     159.000000      65.000000   \n",
       "50%     19700.000000       1.000000     165.000000      72.000000   \n",
       "75%     21324.000000       2.000000     170.000000      82.000000   \n",
       "max     23713.000000       2.000000     250.000000     200.000000   \n",
       "\n",
       "               ap_hi          ap_lo    cholesterol           gluc  \\\n",
       "count  100000.000000  100000.000000  100000.000000  100000.000000   \n",
       "mean      125.889511      80.788696       1.367220       1.225330   \n",
       "std        17.555061      10.109677       0.679922       0.570543   \n",
       "min        60.000000      20.000000       1.000000       1.000000   \n",
       "25%       120.000000      80.000000       1.000000       1.000000   \n",
       "50%       120.000000      80.000000       1.000000       1.000000   \n",
       "75%       140.000000      90.000000       2.000000       1.000000   \n",
       "max       401.000000     182.000000       3.000000       3.000000   \n",
       "\n",
       "               smoke           alco         active  \n",
       "count  100000.000000  100000.000000  100000.000000  \n",
       "mean        0.082610       0.049450       0.820010  \n",
       "std         0.275293       0.216807       0.384181  \n",
       "min         0.000000       0.000000       0.000000  \n",
       "25%         0.000000       0.000000       1.000000  \n",
       "50%         0.000000       0.000000       1.000000  \n",
       "75%         0.000000       0.000000       1.000000  \n",
       "max         1.000000       1.000000       1.000000  "
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_tot.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Since real test data are not labeled, lets divide the train data so we know how well we perform.. Again, I use XGBoostfor classification and I will also try to tune the hyperparameters based on https://www.analyticsvidhya.com/blog/2016/03/complete-guide-parameter-tuning-xgboost-with-codes-python/"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "X_train=X_tot.loc[X.index]\n",
    "y_train=y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "X,X_test,y,y_test=train_test_split(X_train,y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "param_test1 = {\n",
    "         'max_depth':[4,5,6],\n",
    "         'min_child_weight':[1,2]\n",
    "     }"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "xgbc=XGBClassifier( learning_rate =0.1, n_estimators=140, max_depth=5,\n",
    "    min_child_weight=1, gamma=0, subsample=0.8, colsample_bytree=0.8,\n",
    "    objective= 'binary:logistic', nthread=4, scale_pos_weight=1, seed=27) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=5, error_score='raise',\n",
       "       estimator=XGBClassifier(base_score=0.5, colsample_bylevel=1, colsample_bytree=0.8,\n",
       "       gamma=0, learning_rate=0.1, max_delta_step=0, max_depth=5,\n",
       "       min_child_weight=1, missing=None, n_estimators=140, nthread=4,\n",
       "       objective='binary:logistic', reg_alpha=0, reg_lambda=1,\n",
       "       scale_pos_weight=1, seed=27, silent=True, subsample=0.8),\n",
       "       fit_params={}, iid=True, n_jobs=1,\n",
       "       param_grid={'max_depth': [4, 5, 6], 'min_child_weight': [1, 2]},\n",
       "       pre_dispatch='2*n_jobs', refit=True, return_train_score=True,\n",
       "       scoring='neg_log_loss', verbose=0)"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gsearch1=GridSearchCV(xgbc,param_test1,cv=5,scoring='neg_log_loss')\n",
    "gsearch1.fit(X,y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python2.7/dist-packages/sklearn/model_selection/_search.py:667: DeprecationWarning: The grid_scores_ attribute was deprecated in version 0.18 in favor of the more elaborate cv_results_ attribute. The grid_scores_ attribute will not be available from 0.20\n",
      "  DeprecationWarning)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "([mean: -0.54338, std: 0.00624, params: {'max_depth': 4, 'min_child_weight': 1},\n",
       "  mean: -0.54329, std: 0.00623, params: {'max_depth': 4, 'min_child_weight': 2},\n",
       "  mean: -0.54417, std: 0.00639, params: {'max_depth': 5, 'min_child_weight': 1},\n",
       "  mean: -0.54429, std: 0.00639, params: {'max_depth': 5, 'min_child_weight': 2},\n",
       "  mean: -0.54580, std: 0.00665, params: {'max_depth': 6, 'min_child_weight': 1},\n",
       "  mean: -0.54540, std: 0.00627, params: {'max_depth': 6, 'min_child_weight': 2}],\n",
       " {'max_depth': 4, 'min_child_weight': 2},\n",
       " -0.54328906323217918)"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gsearch1.grid_scores_, gsearch1.best_params_, gsearch1.best_score_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Note that when the best value is on the border of our search grid, new search should be performed since the best value might be outside the given boundaries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "param_test2={'gamma':[i/10.0 for i in range(0,5)]}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=5, error_score='raise',\n",
       "       estimator=XGBClassifier(base_score=0.5, colsample_bylevel=1, colsample_bytree=0.8,\n",
       "       gamma=0, learning_rate=0.1, max_delta_step=0, max_depth=4,\n",
       "       min_child_weight=2, missing=None, n_estimators=140, nthread=4,\n",
       "       objective='binary:logistic', reg_alpha=0, reg_lambda=1,\n",
       "       scale_pos_weight=1, seed=27, silent=True, subsample=0.8),\n",
       "       fit_params={}, iid=True, n_jobs=1,\n",
       "       param_grid={'gamma': [0.0, 0.1, 0.2, 0.3, 0.4]},\n",
       "       pre_dispatch='2*n_jobs', refit=True, return_train_score=True,\n",
       "       scoring='neg_log_loss', verbose=0)"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "xgbc=XGBClassifier( learning_rate =0.1, n_estimators=140, max_depth=4,\n",
    "    min_child_weight=2, gamma=0, subsample=0.8, colsample_bytree=0.8,\n",
    "    objective= 'binary:logistic', nthread=4, scale_pos_weight=1, seed=27) \n",
    "gsearch2=GridSearchCV(xgbc,param_test2,cv=5,scoring='neg_log_loss')\n",
    "gsearch2.fit(X,y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python2.7/dist-packages/sklearn/model_selection/_search.py:667: DeprecationWarning: The grid_scores_ attribute was deprecated in version 0.18 in favor of the more elaborate cv_results_ attribute. The grid_scores_ attribute will not be available from 0.20\n",
      "  DeprecationWarning)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "([mean: -0.54329, std: 0.00623, params: {'gamma': 0.0},\n",
       "  mean: -0.54338, std: 0.00624, params: {'gamma': 0.1},\n",
       "  mean: -0.54340, std: 0.00625, params: {'gamma': 0.2},\n",
       "  mean: -0.54340, std: 0.00625, params: {'gamma': 0.3},\n",
       "  mean: -0.54339, std: 0.00624, params: {'gamma': 0.4}],\n",
       " {'gamma': 0.0},\n",
       " -0.54328906323217918)"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gsearch2.grid_scores_, gsearch2.best_params_, gsearch2.best_score_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "param_test3 = {\n",
    "     'subsample':[i/10.0 for i in range(6,10)],\n",
    "     'colsample_bytree':[i/10.0 for i in range(6,10)]\n",
    "     }"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=5, error_score='raise',\n",
       "       estimator=XGBClassifier(base_score=0.5, colsample_bylevel=1, colsample_bytree=0.8,\n",
       "       gamma=0, learning_rate=0.1, max_delta_step=0, max_depth=4,\n",
       "       min_child_weight=1, missing=None, n_estimators=140, nthread=4,\n",
       "       objective='binary:logistic', reg_alpha=0, reg_lambda=1,\n",
       "       scale_pos_weight=1, seed=27, silent=True, subsample=0.8),\n",
       "       fit_params={}, iid=True, n_jobs=1,\n",
       "       param_grid={'subsample': [0.6, 0.7, 0.8, 0.9], 'colsample_bytree': [0.6, 0.7, 0.8, 0.9]},\n",
       "       pre_dispatch='2*n_jobs', refit=True, return_train_score=True,\n",
       "       scoring='neg_log_loss', verbose=0)"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "xgbc=XGBClassifier( learning_rate =0.1, n_estimators=140, max_depth=4,\n",
    "    min_child_weight=1, gamma=0., subsample=0.8, colsample_bytree=0.8,\n",
    "    objective= 'binary:logistic', nthread=4, scale_pos_weight=1, seed=27) \n",
    "gsearch3=GridSearchCV(xgbc,param_test3,cv=5,scoring='neg_log_loss')\n",
    "gsearch3.fit(X,y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python2.7/dist-packages/sklearn/model_selection/_search.py:667: DeprecationWarning: The grid_scores_ attribute was deprecated in version 0.18 in favor of the more elaborate cv_results_ attribute. The grid_scores_ attribute will not be available from 0.20\n",
      "  DeprecationWarning)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "([mean: -0.54364, std: 0.00646, params: {'subsample': 0.6, 'colsample_bytree': 0.6},\n",
       "  mean: -0.54363, std: 0.00631, params: {'subsample': 0.7, 'colsample_bytree': 0.6},\n",
       "  mean: -0.54347, std: 0.00635, params: {'subsample': 0.8, 'colsample_bytree': 0.6},\n",
       "  mean: -0.54339, std: 0.00622, params: {'subsample': 0.9, 'colsample_bytree': 0.6},\n",
       "  mean: -0.54371, std: 0.00646, params: {'subsample': 0.6, 'colsample_bytree': 0.7},\n",
       "  mean: -0.54343, std: 0.00656, params: {'subsample': 0.7, 'colsample_bytree': 0.7},\n",
       "  mean: -0.54334, std: 0.00610, params: {'subsample': 0.8, 'colsample_bytree': 0.7},\n",
       "  mean: -0.54350, std: 0.00645, params: {'subsample': 0.9, 'colsample_bytree': 0.7},\n",
       "  mean: -0.54362, std: 0.00652, params: {'subsample': 0.6, 'colsample_bytree': 0.8},\n",
       "  mean: -0.54344, std: 0.00653, params: {'subsample': 0.7, 'colsample_bytree': 0.8},\n",
       "  mean: -0.54338, std: 0.00624, params: {'subsample': 0.8, 'colsample_bytree': 0.8},\n",
       "  mean: -0.54324, std: 0.00633, params: {'subsample': 0.9, 'colsample_bytree': 0.8},\n",
       "  mean: -0.54313, std: 0.00624, params: {'subsample': 0.6, 'colsample_bytree': 0.9},\n",
       "  mean: -0.54339, std: 0.00632, params: {'subsample': 0.7, 'colsample_bytree': 0.9},\n",
       "  mean: -0.54303, std: 0.00621, params: {'subsample': 0.8, 'colsample_bytree': 0.9},\n",
       "  mean: -0.54335, std: 0.00593, params: {'subsample': 0.9, 'colsample_bytree': 0.9}],\n",
       " {'colsample_bytree': 0.9, 'subsample': 0.8},\n",
       " -0.54302809248877071)"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gsearch3.grid_scores_, gsearch3.best_params_, gsearch3.best_score_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "param_test4 = {\n",
    "     'reg_alpha':[1e-5, 1e-2, 0.1, 1, 100]\n",
    "     }"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=5, error_score='raise',\n",
       "       estimator=XGBClassifier(base_score=0.5, colsample_bylevel=1, colsample_bytree=0.9,\n",
       "       gamma=0.1, learning_rate=0.1, max_delta_step=0, max_depth=4,\n",
       "       min_child_weight=1, missing=None, n_estimators=140, nthread=4,\n",
       "       objective='binary:logistic', reg_alpha=0, reg_lambda=1,\n",
       "       scale_pos_weight=1, seed=27, silent=True, subsample=0.8),\n",
       "       fit_params={}, iid=True, n_jobs=1,\n",
       "       param_grid={'reg_alpha': [1e-05, 0.01, 0.1, 1, 100]},\n",
       "       pre_dispatch='2*n_jobs', refit=True, return_train_score=True,\n",
       "       scoring='neg_log_loss', verbose=0)"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "xgbc=XGBClassifier( learning_rate =0.1, n_estimators=140, max_depth=4,\n",
    "    min_child_weight=1, gamma=0.1, subsample=0.8, colsample_bytree=0.9,\n",
    "    objective= 'binary:logistic', nthread=4, scale_pos_weight=1, seed=27) \n",
    "gsearch4=GridSearchCV(xgbc,param_test4,cv=5,scoring='neg_log_loss')\n",
    "gsearch4.fit(X,y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python2.7/dist-packages/sklearn/model_selection/_search.py:667: DeprecationWarning: The grid_scores_ attribute was deprecated in version 0.18 in favor of the more elaborate cv_results_ attribute. The grid_scores_ attribute will not be available from 0.20\n",
      "  DeprecationWarning)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "([mean: -0.54311, std: 0.00623, params: {'reg_alpha': 1e-05},\n",
       "  mean: -0.54318, std: 0.00606, params: {'reg_alpha': 0.01},\n",
       "  mean: -0.54320, std: 0.00614, params: {'reg_alpha': 0.1},\n",
       "  mean: -0.54297, std: 0.00623, params: {'reg_alpha': 1},\n",
       "  mean: -0.54684, std: 0.00584, params: {'reg_alpha': 100}],\n",
       " {'reg_alpha': 1},\n",
       " -0.54296788868804768)"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gsearch4.grid_scores_, gsearch4.best_params_, gsearch4.best_score_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "param_test5 = {\n",
    "     'learning_rate':[1e-3, 1e-2, 0.1, 1]\n",
    "     }"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=5, error_score='raise',\n",
       "       estimator=XGBClassifier(base_score=0.5, colsample_bylevel=1, colsample_bytree=0.9,\n",
       "       gamma=0.1, learning_rate=0.1, max_delta_step=0, max_depth=4,\n",
       "       min_child_weight=1, missing=None, n_estimators=140, nthread=4,\n",
       "       objective='binary:logistic', reg_alpha=1, reg_lambda=1,\n",
       "       scale_pos_weight=1, seed=27, silent=True, subsample=0.8),\n",
       "       fit_params={}, iid=True, n_jobs=1,\n",
       "       param_grid={'learning_rate': [0.001, 0.01, 0.1, 1]},\n",
       "       pre_dispatch='2*n_jobs', refit=True, return_train_score=True,\n",
       "       scoring='neg_log_loss', verbose=0)"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "xgbc=XGBClassifier( learning_rate =0.1, n_estimators=140, max_depth=4,\n",
    "    min_child_weight=1, gamma=0.1, subsample=0.8, colsample_bytree=0.9,\n",
    "    objective= 'binary:logistic', nthread=4, scale_pos_weight=1, seed=27,reg_alpha=1) \n",
    "gsearch5=GridSearchCV(xgbc,param_test5,cv=5,scoring='neg_log_loss')\n",
    "gsearch5.fit(X,y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python2.7/dist-packages/sklearn/model_selection/_search.py:667: DeprecationWarning: The grid_scores_ attribute was deprecated in version 0.18 in favor of the more elaborate cv_results_ attribute. The grid_scores_ attribute will not be available from 0.20\n",
      "  DeprecationWarning)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "([mean: -0.66434, std: 0.00049, params: {'learning_rate': 0.001},\n",
       "  mean: -0.56585, std: 0.00365, params: {'learning_rate': 0.01},\n",
       "  mean: -0.54297, std: 0.00623, params: {'learning_rate': 0.1},\n",
       "  mean: -0.58825, std: 0.00647, params: {'learning_rate': 1}],\n",
       " {'learning_rate': 0.1},\n",
       " -0.54296788868804768)"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gsearch5.grid_scores_, gsearch5.best_params_, gsearch5.best_score_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "lets see the final performance on train and test set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "XGBClassifier(base_score=0.5, colsample_bylevel=1, colsample_bytree=0.9,\n",
       "       gamma=0.1, learning_rate=0.1, max_delta_step=0, max_depth=4,\n",
       "       min_child_weight=1, missing=None, n_estimators=1000, nthread=4,\n",
       "       objective='binary:logistic', reg_alpha=1.0, reg_lambda=1,\n",
       "       scale_pos_weight=1, seed=27, silent=True, subsample=0.9)"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "xgbc=XGBClassifier( learning_rate =0.1, n_estimators=1000, max_depth=4,\n",
    "    min_child_weight=1, gamma=0.1, subsample=0.9, colsample_bytree=0.9,\n",
    "    objective= 'binary:logistic', nthread=4, scale_pos_weight=1, seed=27,reg_alpha=1.) \n",
    "xgbc.fit(X,y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.48982864123603242"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import log_loss\n",
    "y_pred_train=xgbc.predict_proba(X)\n",
    "log_loss(y,y_pred_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.49943714523001442"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import log_loss\n",
    "y_pred_test=xgbc.predict_proba(X_test)\n",
    "log_loss(y_test,y_pred_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Train on full dataset and predict unlabeled test data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "XGBClassifier(base_score=0.5, colsample_bylevel=1, colsample_bytree=0.9,\n",
       "       gamma=0.1, learning_rate=0.1, max_delta_step=0, max_depth=4,\n",
       "       min_child_weight=1, missing=None, n_estimators=1000, nthread=4,\n",
       "       objective='binary:logistic', reg_alpha=0.01, reg_lambda=1,\n",
       "       scale_pos_weight=1, seed=27, silent=True, subsample=0.9)"
      ]
     },
     "execution_count": 68,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "xgbc=XGBClassifier( learning_rate =0.1, n_estimators=1000, max_depth=4,\n",
    "    min_child_weight=1, gamma=0.1, subsample=0.9, colsample_bytree=0.9,\n",
    "    objective= 'binary:logistic', nthread=4, scale_pos_weight=1, seed=27,reg_alpha=0.01) \n",
    "xgbc.fit(X_train,y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.49867226887597105"
      ]
     },
     "execution_count": 69,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_tr=xgbc.predict_proba(X_train)\n",
    "log_loss(y_train,y_tr)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}