step-04-predict.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>house</th>\n",
       "      <th>id</th>\n",
       "      <th>lat</th>\n",
       "      <th>lng</th>\n",
       "      <th>locality</th>\n",
       "      <th>province</th>\n",
       "      <th>atm_group</th>\n",
       "      <th>OSM:importance</th>\n",
       "      <th>square</th>\n",
       "      <th>density_province</th>\n",
       "      <th>...</th>\n",
       "      <th>province_Тульская_область</th>\n",
       "      <th>province_Тюменская_область</th>\n",
       "      <th>province_Удмуртская_Республика</th>\n",
       "      <th>province_Ульяновская_область</th>\n",
       "      <th>province_Хабаровский_край</th>\n",
       "      <th>province_Ханты-Мансийский_автономный_округ</th>\n",
       "      <th>province_Челябинская_область</th>\n",
       "      <th>province_Чувашская_Республика</th>\n",
       "      <th>province_Ярославская_область</th>\n",
       "      <th>target</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>34</td>\n",
       "      <td>8526</td>\n",
       "      <td>46.940995</td>\n",
       "      <td>142.738319</td>\n",
       "      <td>8553.5</td>\n",
       "      <td>7163.0</td>\n",
       "      <td>32.0</td>\n",
       "      <td>0.000</td>\n",
       "      <td>7.508988</td>\n",
       "      <td>563</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>0.011500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>259</td>\n",
       "      <td>8532</td>\n",
       "      <td>46.937341</td>\n",
       "      <td>142.753339</td>\n",
       "      <td>8553.5</td>\n",
       "      <td>7163.0</td>\n",
       "      <td>32.0</td>\n",
       "      <td>0.711</td>\n",
       "      <td>7.308826</td>\n",
       "      <td>563</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>0.029710</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>32</td>\n",
       "      <td>8533</td>\n",
       "      <td>46.960268</td>\n",
       "      <td>142.739011</td>\n",
       "      <td>8553.5</td>\n",
       "      <td>7163.0</td>\n",
       "      <td>32.0</td>\n",
       "      <td>0.711</td>\n",
       "      <td>9.547201</td>\n",
       "      <td>563</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>0.009540</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>8</td>\n",
       "      <td>470</td>\n",
       "      <td>46.966405</td>\n",
       "      <td>142.731815</td>\n",
       "      <td>8553.5</td>\n",
       "      <td>7163.0</td>\n",
       "      <td>1942.0</td>\n",
       "      <td>0.711</td>\n",
       "      <td>7.477486</td>\n",
       "      <td>563</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>0.041914</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>1183</td>\n",
       "      <td>46.900367</td>\n",
       "      <td>142.742227</td>\n",
       "      <td>8553.5</td>\n",
       "      <td>7163.0</td>\n",
       "      <td>1942.0</td>\n",
       "      <td>0.900</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>563</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>0.065990</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 1160 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   house    id        lat         lng  locality  province  atm_group  \\\n",
       "0     34  8526  46.940995  142.738319    8553.5    7163.0       32.0   \n",
       "1    259  8532  46.937341  142.753339    8553.5    7163.0       32.0   \n",
       "2     32  8533  46.960268  142.739011    8553.5    7163.0       32.0   \n",
       "3      8   470  46.966405  142.731815    8553.5    7163.0     1942.0   \n",
       "4      1  1183  46.900367  142.742227    8553.5    7163.0     1942.0   \n",
       "\n",
       "   OSM:importance    square  density_province    ...     \\\n",
       "0           0.000  7.508988               563    ...      \n",
       "1           0.711  7.308826               563    ...      \n",
       "2           0.711  9.547201               563    ...      \n",
       "3           0.711  7.477486               563    ...      \n",
       "4           0.900  0.000000               563    ...      \n",
       "\n",
       "   province_Тульская_область  province_Тюменская_область  \\\n",
       "0                          0                           0   \n",
       "1                          0                           0   \n",
       "2                          0                           0   \n",
       "3                          0                           0   \n",
       "4                          0                           0   \n",
       "\n",
       "   province_Удмуртская_Республика  province_Ульяновская_область  \\\n",
       "0                               0                             0   \n",
       "1                               0                             0   \n",
       "2                               0                             0   \n",
       "3                               0                             0   \n",
       "4                               0                             0   \n",
       "\n",
       "   province_Хабаровский_край  province_Ханты-Мансийский_автономный_округ  \\\n",
       "0                          0                                           0   \n",
       "1                          0                                           0   \n",
       "2                          0                                           0   \n",
       "3                          0                                           0   \n",
       "4                          0                                           0   \n",
       "\n",
       "   province_Челябинская_область  province_Чувашская_Республика  \\\n",
       "0                             0                              0   \n",
       "1                             0                              0   \n",
       "2                             0                              0   \n",
       "3                             0                              0   \n",
       "4                             0                              0   \n",
       "\n",
       "   province_Ярославская_область    target  \n",
       "0                             0  0.011500  \n",
       "1                             0  0.029710  \n",
       "2                             0  0.009540  \n",
       "3                             0  0.041914  \n",
       "4                             0  0.065990  \n",
       "\n",
       "[5 rows x 1160 columns]"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_dataset = pd.read_csv('dataset_features.csv', sep=',')\n",
    "df_dataset.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(6261, 1160)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>house</th>\n",
       "      <th>id</th>\n",
       "      <th>lat</th>\n",
       "      <th>lng</th>\n",
       "      <th>locality</th>\n",
       "      <th>province</th>\n",
       "      <th>atm_group</th>\n",
       "      <th>OSM:importance</th>\n",
       "      <th>square</th>\n",
       "      <th>density_province</th>\n",
       "      <th>...</th>\n",
       "      <th>province_Тульская_область</th>\n",
       "      <th>province_Тюменская_область</th>\n",
       "      <th>province_Удмуртская_Республика</th>\n",
       "      <th>province_Ульяновская_область</th>\n",
       "      <th>province_Хабаровский_край</th>\n",
       "      <th>province_Ханты-Мансийский_автономный_округ</th>\n",
       "      <th>province_Челябинская_область</th>\n",
       "      <th>province_Чувашская_Республика</th>\n",
       "      <th>province_Ярославская_область</th>\n",
       "      <th>target</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>34</td>\n",
       "      <td>8526</td>\n",
       "      <td>46.940995</td>\n",
       "      <td>142.738319</td>\n",
       "      <td>8553.5</td>\n",
       "      <td>7163.0</td>\n",
       "      <td>32.0</td>\n",
       "      <td>0.000</td>\n",
       "      <td>7.508988</td>\n",
       "      <td>563</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>0.011500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>259</td>\n",
       "      <td>8532</td>\n",
       "      <td>46.937341</td>\n",
       "      <td>142.753339</td>\n",
       "      <td>8553.5</td>\n",
       "      <td>7163.0</td>\n",
       "      <td>32.0</td>\n",
       "      <td>0.711</td>\n",
       "      <td>7.308826</td>\n",
       "      <td>563</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>0.029710</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>32</td>\n",
       "      <td>8533</td>\n",
       "      <td>46.960268</td>\n",
       "      <td>142.739011</td>\n",
       "      <td>8553.5</td>\n",
       "      <td>7163.0</td>\n",
       "      <td>32.0</td>\n",
       "      <td>0.711</td>\n",
       "      <td>9.547201</td>\n",
       "      <td>563</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>0.009540</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>8</td>\n",
       "      <td>470</td>\n",
       "      <td>46.966405</td>\n",
       "      <td>142.731815</td>\n",
       "      <td>8553.5</td>\n",
       "      <td>7163.0</td>\n",
       "      <td>1942.0</td>\n",
       "      <td>0.711</td>\n",
       "      <td>7.477486</td>\n",
       "      <td>563</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>0.041914</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>1183</td>\n",
       "      <td>46.900367</td>\n",
       "      <td>142.742227</td>\n",
       "      <td>8553.5</td>\n",
       "      <td>7163.0</td>\n",
       "      <td>1942.0</td>\n",
       "      <td>0.900</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>563</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>0.065990</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 1160 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   house    id        lat         lng  locality  province  atm_group  \\\n",
       "0     34  8526  46.940995  142.738319    8553.5    7163.0       32.0   \n",
       "1    259  8532  46.937341  142.753339    8553.5    7163.0       32.0   \n",
       "2     32  8533  46.960268  142.739011    8553.5    7163.0       32.0   \n",
       "3      8   470  46.966405  142.731815    8553.5    7163.0     1942.0   \n",
       "4      1  1183  46.900367  142.742227    8553.5    7163.0     1942.0   \n",
       "\n",
       "   OSM:importance    square  density_province    ...     \\\n",
       "0           0.000  7.508988               563    ...      \n",
       "1           0.711  7.308826               563    ...      \n",
       "2           0.711  9.547201               563    ...      \n",
       "3           0.711  7.477486               563    ...      \n",
       "4           0.900  0.000000               563    ...      \n",
       "\n",
       "   province_Тульская_область  province_Тюменская_область  \\\n",
       "0                          0                           0   \n",
       "1                          0                           0   \n",
       "2                          0                           0   \n",
       "3                          0                           0   \n",
       "4                          0                           0   \n",
       "\n",
       "   province_Удмуртская_Республика  province_Ульяновская_область  \\\n",
       "0                               0                             0   \n",
       "1                               0                             0   \n",
       "2                               0                             0   \n",
       "3                               0                             0   \n",
       "4                               0                             0   \n",
       "\n",
       "   province_Хабаровский_край  province_Ханты-Мансийский_автономный_округ  \\\n",
       "0                          0                                           0   \n",
       "1                          0                                           0   \n",
       "2                          0                                           0   \n",
       "3                          0                                           0   \n",
       "4                          0                                           0   \n",
       "\n",
       "   province_Челябинская_область  province_Чувашская_Республика  \\\n",
       "0                             0                              0   \n",
       "1                             0                              0   \n",
       "2                             0                              0   \n",
       "3                             0                              0   \n",
       "4                             0                              0   \n",
       "\n",
       "   province_Ярославская_область    target  \n",
       "0                             0  0.011500  \n",
       "1                             0  0.029710  \n",
       "2                             0  0.009540  \n",
       "3                             0  0.041914  \n",
       "4                             0  0.065990  \n",
       "\n",
       "[5 rows x 1160 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mask = df_dataset['target'].notna()\n",
    "df_train = df_dataset.loc[mask]\n",
    "print(df_train.shape)\n",
    "df_train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(2504, 1160)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>house</th>\n",
       "      <th>id</th>\n",
       "      <th>lat</th>\n",
       "      <th>lng</th>\n",
       "      <th>locality</th>\n",
       "      <th>province</th>\n",
       "      <th>atm_group</th>\n",
       "      <th>OSM:importance</th>\n",
       "      <th>square</th>\n",
       "      <th>density_province</th>\n",
       "      <th>...</th>\n",
       "      <th>province_Тульская_область</th>\n",
       "      <th>province_Тюменская_область</th>\n",
       "      <th>province_Удмуртская_Республика</th>\n",
       "      <th>province_Ульяновская_область</th>\n",
       "      <th>province_Хабаровский_край</th>\n",
       "      <th>province_Ханты-Мансийский_автономный_округ</th>\n",
       "      <th>province_Челябинская_область</th>\n",
       "      <th>province_Чувашская_Республика</th>\n",
       "      <th>province_Ярославская_область</th>\n",
       "      <th>target</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>36</th>\n",
       "      <td>2</td>\n",
       "      <td>3385</td>\n",
       "      <td>46.943832</td>\n",
       "      <td>142.752998</td>\n",
       "      <td>8553.5</td>\n",
       "      <td>7163.0</td>\n",
       "      <td>8083.0</td>\n",
       "      <td>0.000</td>\n",
       "      <td>10.016173</td>\n",
       "      <td>563</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>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>37</th>\n",
       "      <td>234</td>\n",
       "      <td>6077</td>\n",
       "      <td>46.955710</td>\n",
       "      <td>142.729767</td>\n",
       "      <td>8553.5</td>\n",
       "      <td>7163.0</td>\n",
       "      <td>5478.0</td>\n",
       "      <td>0.711</td>\n",
       "      <td>7.431945</td>\n",
       "      <td>563</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>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>38</th>\n",
       "      <td>234</td>\n",
       "      <td>6078</td>\n",
       "      <td>46.955710</td>\n",
       "      <td>142.729767</td>\n",
       "      <td>8553.5</td>\n",
       "      <td>7163.0</td>\n",
       "      <td>5478.0</td>\n",
       "      <td>0.711</td>\n",
       "      <td>7.431945</td>\n",
       "      <td>563</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>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39</th>\n",
       "      <td>234</td>\n",
       "      <td>7868</td>\n",
       "      <td>46.955710</td>\n",
       "      <td>142.729767</td>\n",
       "      <td>8553.5</td>\n",
       "      <td>7163.0</td>\n",
       "      <td>5478.0</td>\n",
       "      <td>0.711</td>\n",
       "      <td>7.431945</td>\n",
       "      <td>563</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>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>40</th>\n",
       "      <td>234</td>\n",
       "      <td>8098</td>\n",
       "      <td>46.955710</td>\n",
       "      <td>142.729767</td>\n",
       "      <td>8553.5</td>\n",
       "      <td>7163.0</td>\n",
       "      <td>5478.0</td>\n",
       "      <td>0.711</td>\n",
       "      <td>7.431945</td>\n",
       "      <td>563</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>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 1160 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "    house    id        lat         lng  locality  province  atm_group  \\\n",
       "36      2  3385  46.943832  142.752998    8553.5    7163.0     8083.0   \n",
       "37    234  6077  46.955710  142.729767    8553.5    7163.0     5478.0   \n",
       "38    234  6078  46.955710  142.729767    8553.5    7163.0     5478.0   \n",
       "39    234  7868  46.955710  142.729767    8553.5    7163.0     5478.0   \n",
       "40    234  8098  46.955710  142.729767    8553.5    7163.0     5478.0   \n",
       "\n",
       "    OSM:importance     square  density_province   ...    \\\n",
       "36           0.000  10.016173               563   ...     \n",
       "37           0.711   7.431945               563   ...     \n",
       "38           0.711   7.431945               563   ...     \n",
       "39           0.711   7.431945               563   ...     \n",
       "40           0.711   7.431945               563   ...     \n",
       "\n",
       "    province_Тульская_область  province_Тюменская_область  \\\n",
       "36                          0                           0   \n",
       "37                          0                           0   \n",
       "38                          0                           0   \n",
       "39                          0                           0   \n",
       "40                          0                           0   \n",
       "\n",
       "    province_Удмуртская_Республика  province_Ульяновская_область  \\\n",
       "36                               0                             0   \n",
       "37                               0                             0   \n",
       "38                               0                             0   \n",
       "39                               0                             0   \n",
       "40                               0                             0   \n",
       "\n",
       "    province_Хабаровский_край  province_Ханты-Мансийский_автономный_округ  \\\n",
       "36                          0                                           0   \n",
       "37                          0                                           0   \n",
       "38                          0                                           0   \n",
       "39                          0                                           0   \n",
       "40                          0                                           0   \n",
       "\n",
       "    province_Челябинская_область  province_Чувашская_Республика  \\\n",
       "36                             0                              0   \n",
       "37                             0                              0   \n",
       "38                             0                              0   \n",
       "39                             0                              0   \n",
       "40                             0                              0   \n",
       "\n",
       "    province_Ярославская_область  target  \n",
       "36                             0     NaN  \n",
       "37                             0     NaN  \n",
       "38                             0     NaN  \n",
       "39                             0     NaN  \n",
       "40                             0     NaN  \n",
       "\n",
       "[5 rows x 1160 columns]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mask = np.logical_not(mask)\n",
    "df_test = df_dataset.loc[mask]\n",
    "print(df_test.shape)\n",
    "df_test.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "features = np.asarray(df_dataset.columns.tolist()[:-1])\n",
    "target = df_dataset.columns.tolist()[-1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import KFold\n",
    "from sklearn.metrics import mean_squared_error"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "cl_params = {}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## XGBoost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import xgboost as xgb"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from xgboost import XGBRegressor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%time\n",
    "\n",
    "params = {\n",
    "    'max_depth': 5,\n",
    "    'learning_rate': 0.01,\n",
    "    'silent': 0,\n",
    "    'objective': 'reg:linear',\n",
    "    'seed': 8888,\n",
    "}\n",
    "\n",
    "X_train = xgb.DMatrix(df_train.loc[:, features], label=df_train.loc[:, target])\n",
    "\n",
    "trees = 2000\n",
    "cv = xgb.cv(params, X_train, metrics=('rmse'), show_stdv=False,\n",
    "            num_boost_round=trees, nfold=5, early_stopping_rounds=15)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "cv[100:].plot(figsize=(6, 6), y=['test-rmse-mean', 'train-rmse-mean'])\n",
    "print(cv.loc[cv['test-rmse-mean'].values.argmin()])\n",
    "trees = cv['test-rmse-mean'].values.argmin()\n",
    "trees"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "params['n_estimators'] = trees\n",
    "cl_params['xgb'] = params"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## LightGBM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "import lightgbm as lgbm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "from lightgbm import LGBMRegressor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/vladimir/anaconda3/lib/python3.6/site-packages/lightgbm/basic.py:657: UserWarning: silent keyword has been found in `params` and will be ignored. Please use silent argument of the Dataset constructor to pass this parameter.\n",
      "  'Please use {0} argument of the Dataset constructor to pass this parameter.'.format(key))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 4min 20s, sys: 6.21 s, total: 4min 26s\n",
      "Wall time: 1min 14s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "params = {\n",
    "    'max_depth': 5,\n",
    "    'learning_rate': 0.005,\n",
    "    'colsample_bytre': 0.8,\n",
    "    'subsample': 0.8,\n",
    "    'silent': 0,\n",
    "    'objective': 'regression',\n",
    "    'seed': 8888,\n",
    "}\n",
    "\n",
    "X_train = lgbm.Dataset(df_train.loc[:, features], label=df_train.loc[:, target])\n",
    "\n",
    "trees = 12000\n",
    "cv = lgbm.cv(params, X_train, metrics=('rmse'), show_stdv=False, stratified=False,\n",
    "             num_boost_round=trees, nfold=5, early_stopping_rounds=30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "cv = pd.DataFrame(cv)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "rmse-mean    0.042246\n",
      "rmse-stdv    0.000644\n",
      "Name: 4478, dtype: float64\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "4478"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "cv[100:].plot(figsize=(6, 6), y=['rmse-mean'])\n",
    "print(cv.loc[cv['rmse-mean'].values.argmin()])\n",
    "trees = cv['rmse-mean'].values.argmin()\n",
    "trees"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "params['n_estimators'] = trees\n",
    "cl_params['lgb'] = params"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'max_depth': 5,\n",
       " 'learning_rate': 0.005,\n",
       " 'colsample_bytre': 0.8,\n",
       " 'subsample': 0.8,\n",
       " 'silent': 0,\n",
       " 'objective': 'regression',\n",
       " 'seed': 8888,\n",
       " 'metric': 'rmse',\n",
       " 'verbose': 1,\n",
       " 'n_estimators': 4478}"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cl_params['lgb']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.0414299078268\n",
      "0.0413306940608\n",
      "0.0404711952204\n",
      "0.0423081578134\n",
      "0.0436478741512\n",
      "==============================\n",
      "MEAN: 0.0418375658145\n",
      "CPU times: user 7min 24s, sys: 27 s, total: 7min 51s\n",
      "Wall time: 4min 17s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "scores = []\n",
    "feature_importances = []\n",
    "\n",
    "kf = KFold(n_splits=5, shuffle=True, random_state=8888)\n",
    "\n",
    "for train_index, valid_index in kf.split(df_train):\n",
    "    cl = LGBMRegressor(**cl_params['lgb'], random_state=8888)\n",
    "    \n",
    "    X_train = df_train.iloc[train_index].loc[:, features].values\n",
    "    Y_train = df_train.iloc[train_index].loc[:, target].values\n",
    "    \n",
    "    X_valid = df_train.iloc[valid_index].loc[:, features].values\n",
    "    Y_valid = df_train.iloc[valid_index].loc[:, target].values\n",
    "    \n",
    "    cl.fit(X_train, Y_train)\n",
    "    Y_pred = cl.predict(X_valid)\n",
    "    \n",
    "    feature_importances.append(cl.feature_importances_)\n",
    "    scores.append(np.sqrt(mean_squared_error(Y_valid, Y_pred)))\n",
    "    print(scores[-1])\n",
    "    \n",
    "print(\"=\" * 30)\n",
    "print(\"MEAN:\", np.mean(scores))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Feature ranking:\n",
      "[001] id                                                 0.0487299842916659\n",
      "[002] distance_locality                                  0.03192268767733964\n",
      "[003] atm_group_8083.0_closest_distance                  0.029554019905501643\n",
      "[004] atm_group_496.5_closest_distance                   0.02819428663674012\n",
      "[005] house                                              0.026884223989668513\n",
      "[006] 2ГИС:norgs                                         0.02621367059685461\n",
      "[007] atm_group                                          0.02595910866069378\n",
      "[008] atm_group_3185.5_closest_distance                  0.025654876102843022\n",
      "[009] atm_group_32.0_closest_distance                    0.02477322256784697\n",
      "[010] atm_group_closest_distance                         0.02469561222145647\n",
      "[011] atm_group_1942.0_closest_distance                  0.022280378241784173\n",
      "[012] atm_group_1022.0_closest_distance                  0.02032770192659924\n",
      "[013] distance_province                                  0.020110392956705844\n",
      "[014] atm_group_5478.0_closest_distance                  0.02002346936874849\n",
      "[015] mul_square_density_province_div_count              0.020011051713326008\n",
      "[016] distance_1                                         0.01913250259218557\n",
      "[017] distance_2                                         0.018108046019830997\n",
      "[018] mul_square_density_province_div_count_mul_salary   0.016968726134818486\n",
      "[019] distance_4                                         0.016139847635367965\n",
      "[020] lat                                                0.015857345974506554\n",
      "[021] distance_5                                         0.015813884180527874\n",
      "[022] distance_3                                         0.015739378247993\n",
      "[023] OSM:importance                                     0.014503821533456269\n",
      "[024] square                                             0.014010219730412702\n",
      "[025] mul_square_density_province                        0.013892252003899146\n",
      "[026] mul_square_density_province_mul_salary             0.012582189356827539\n",
      "[027] metro                                              0.012582189356827539\n",
      "[028] mean                                               0.012007872793537853\n",
      "[029] salary                                             0.011998559551970994\n",
      "[030] lng                                                0.011967515413414793\n",
      "[031] atm_group_32.0_closest_index                       0.011169681052520475\n",
      "[032] atm_group_3185.5_closest_index                     0.011160367810953614\n",
      "[033] id-rosbank                                         0.010654348352487568\n",
      "[034] atm_group_496.5_closest_index                      0.010511545315129052\n",
      "[035] atm_group_1942.0_closest_index                     0.010409099657893594\n",
      "[036] province                                           0.01018247744643334\n",
      "[037] atm_group_5478.0_closest_index                     0.009931019924128126\n",
      "[038] 2ГИС:floors                                        0.008922085421051652\n",
      "[039] density_province                                   0.008515407205965443\n",
      "[040] atm_group_5478.0_1.2km                             0.008487467481264863\n",
      "[041] locality                                           0.008294993822216427\n",
      "[042] atm_group_8083.0_closest_index                     0.008040431886055594\n",
      "[043] atm_group_1022.0_closest_index                     0.007897628848697078\n",
      "[044] atm_group_1942.0_1.2km                             0.006832814896219445\n",
      "[045] atm_group_1942.0                                   0.006438554336555716\n",
      "[046] lat_locality                                       0.005920117222667188\n",
      "[047] atm_group_closest_index                            0.005783523013019913\n",
      "[048] atm_group_8083.0_1.2km                             0.005783523013019913\n",
      "[049] indexes_2                                          0.005771105357597433\n",
      "[050] count                                              0.005733852391329993\n",
      "[051] lat_province                                       0.005687286183495694\n",
      "[052] indexes_3                                          0.005575527284693378\n",
      "[053] indexes_4                                          0.005532065490714698\n",
      "[054] indexes_1                                          0.00510676079249477\n",
      "[055] indexes_5                                          0.005060194584660471\n",
      "[056] atm_group_3185.5                                   0.004932913616580054\n",
      "[057] access24h                                          0.004268569051477391\n",
      "[058] atm_group_8083.0_0.5km                             0.004156810152675074\n",
      "[059] lng_locality                                       0.004131974841830114\n",
      "[060] atm_group_5478.0_0.5km                             0.003799802559278783\n",
      "[061] atm_group_3185.5_1.2km                             0.0037470275237332443\n",
      "[062] comment_term_363                                   0.0034303773104600124\n",
      "[063] comment_term_493                                   0.003101309441764301\n",
      "[064] lng_province                                       0.002899522541149006\n",
      "[065] address_term_11                                    0.002610812052576353\n",
      "[066] comment_term_505                                   0.002589081155587014\n",
      "[067] address_term_18                                    0.0024835310844959365\n",
      "[068] 2ГИС:type_Сеть гипермаркетов                       0.00235625011641552\n",
      "[069] comment_term_173                                   0.0023438324609930403\n",
      "[070] 2ГИС:type_Жилой дом с административными помещениями 0.0023376236332818003\n",
      "[071] comment_term_615                                   0.0023158927362924605\n",
      "[072] comment_term_806                                   0.002272430942313782\n",
      "[073] comment_term_613                                   0.0021606720435114646\n",
      "[074] comment_term_106                                   0.002107897007965926\n",
      "[075] comment_term_313                                   0.0020830616971209666\n",
      "[076] comment_term_497                                   0.002067539627842867\n",
      "[077] comment_term_85                                    0.0020364954892866678\n",
      "[078] comment_term_364                                   0.002033391075431048\n",
      "[079] address_term_12                                    0.0019650939706074098\n",
      "[080] comment_term_747                                   0.0019278410043399709\n",
      "[081] comment_term_829                                   0.0018967968657837716\n",
      "[082] 2ГИС:type_rare                                     0.0018626483133719525\n",
      "[083] comment_term_641                                   0.0018533350718050927\n",
      "[084] 2ГИС:type_Административное здание                  0.0018502306579494727\n",
      "[085] OSM:type_shop:convenience                          0.0018347085886713731\n",
      "[086] OSM:type_amenity:bank                              0.001800560036259554\n",
      "[087] comment_term_742                                   0.0017881423808370743\n",
      "[088] atm_group_1022.0                                   0.0017664114838477347\n",
      "[089] OSM:type_highway:residential                       0.0017291585175802959\n",
      "[090] atm_group_3185.5_0.5km                             0.0016825923097459972\n",
      "[091] comment_term_228                                   0.0016763834820347572\n",
      "[092] province_Белгородская_область                      0.0016205040326335988\n",
      "[093] locality_Томск                                     0.0015428936862431006\n",
      "[094] OSM:type_highway:trunk                             0.0015397892723874808\n",
      "[095] province_Ивановская_область                        0.0014590745121413626\n",
      "[096] comment_term_579                                   0.001440448029007643\n",
      "[097] atm_group_496.5                                    0.0014031950627402042\n",
      "[098] comment_term_654                                   0.0013938818211733444\n",
      "[099] atm_group_496.5_1.2km                              0.001365942096472765\n",
      "[100] comment_term_86                                    0.001359733268761525\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "features_ = np.array(features)\n",
    "\n",
    "importances = np.mean(feature_importances, axis=0)\n",
    "importances = importances / importances.sum()\n",
    "indices = np.argsort(importances)[::-1]\n",
    "\n",
    "print(\"Feature ranking:\")\n",
    "\n",
    "n_features = len(features_)\n",
    "n_cut = 100\n",
    "\n",
    "for f in range(n_cut):\n",
    "    print(\"[{:03}] {:50} {}\".format(f + 1, features_[indices[f]], importances[indices[f]]))\n",
    "    \n",
    "plt.figure(figsize=(16, 6))\n",
    "plt.title(\"Feature importances\")\n",
    "plt.bar(range(n_cut), importances[indices][:n_cut], color=\"r\", align=\"center\")\n",
    "plt.xticks(range(n_cut), features_[indices][:n_cut], rotation='vertical')\n",
    "plt.xlim([-1, n_cut])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "tmp = pd.read_csv('../test.csv', sep=',', index_col=0)\n",
    "tmp = pd.DataFrame(tmp['id'].astype(int))\n",
    "tmp = pd.merge(tmp, df_test, on='id', how='left')\n",
    "\n",
    "df_test = tmp\n",
    "\n",
    "X_test = df_test.loc[:, features].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "seed=1885; rmse=0.04128884426247743\n",
      "seed=4565; rmse=0.042645893378066686\n",
      "seed=5060; rmse=0.03976941132047354\n",
      "seed=6588; rmse=0.04385921253850875\n",
      "seed=4867; rmse=0.03991377773791004\n",
      "seed=5651; rmse=0.0432534456373616\n",
      "seed=5021; rmse=0.043157056409710515\n",
      "seed=4498; rmse=0.04250425655694063\n",
      "seed=5840; rmse=0.042757398778707836\n",
      "seed=7810; rmse=0.0435923844515415\n",
      "==============================\n",
      "MEAN: 0.0422741681072\n"
     ]
    }
   ],
   "source": [
    "scores, Y_pred_test = [], []\n",
    "\n",
    "for seed in np.random.randint(100, 10000, 10):    \n",
    "    X_train, X_valid, Y_train, Y_valid = train_test_split(\n",
    "        df_train.loc[:, features].values,\n",
    "        df_train.loc[:, target].values,\n",
    "        test_size=0.2, random_state=seed)\n",
    "    \n",
    "    cl = LGBMRegressor(**cl_params['lgb'])\n",
    "    cl.fit(X_train, Y_train)\n",
    "    Y_pred = cl.predict(X_valid)\n",
    "    scores.append(np.sqrt(mean_squared_error(Y_valid, Y_pred)))\n",
    "    print(\"seed={:4d}; rmse={}\".format(seed, scores[-1]))\n",
    "    \n",
    "    Y_pred = cl.predict(X_test)\n",
    "    Y_pred_test.append(Y_pred)\n",
    "    \n",
    "print(\"=\" * 30)\n",
    "print(\"MEAN:\", np.mean(scores))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "Y_pred = np.mean(Y_pred_test, axis=0)\n",
    "assert len(Y_pred) == df_test.shape[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_pred = pd.DataFrame(Y_pred, columns=['target'])\n",
    "df_pred.to_csv('submission_11.csv', index=True, sep=',')"
   ]
  },
  {
   "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
}