determining-market-value-of-a-car .ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Deskripsi Proyek\n",
    "Rusty Bargain adalah perusahaan jual beli mobil bekas yang sedang mengembangkan sebuah aplikasi untuk menarik pembeli baru. Dalam aplikasi tersebut, kamu bisa dengan cepat mengetahui nilai pasar mobilmu. Kamu memiliki akses ke data historis, spesifikasi teknis kendaraan, versi model kendaraan, dan harga kendaraan. Tugasmu adalah membuat suatu model yang mampu menentukan nilai pasar mobil.\n",
    "Rusty Bargain tertarik pada:\n",
    "\n",
    "- kualitas prediksi;\n",
    "- kecepatan model dalam memprediksi;\n",
    "- waktu yang diperlukan untuk melatih model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Persiapan Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Import Library untuk mengolah data\n",
    "import pandas as pd, numpy as np\n",
    "\n",
    "# Import Library untuk data viz\n",
    "import matplotlib.pyplot as plt, seaborn as sns\n",
    "\n",
    "\n",
    "# Import Library untuk Machine Learning\n",
    "from sklearn.metrics import mean_squared_error as mse\n",
    "from sklearn.model_selection import train_test_split\n",
    "import lightgbm as lgb\n",
    "from xgboost import XGBRegressor\n",
    "from lightgbm import LGBMRegressor\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.linear_model import LinearRegression\n",
    "from sklearn.ensemble import RandomForestRegressor, GradientBoostingRegressor\n",
    "import time"
   ]
  },
  {
   "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>DateCrawled</th>\n",
       "      <th>Price</th>\n",
       "      <th>VehicleType</th>\n",
       "      <th>RegistrationYear</th>\n",
       "      <th>Gearbox</th>\n",
       "      <th>Power</th>\n",
       "      <th>Model</th>\n",
       "      <th>Mileage</th>\n",
       "      <th>RegistrationMonth</th>\n",
       "      <th>FuelType</th>\n",
       "      <th>Brand</th>\n",
       "      <th>NotRepaired</th>\n",
       "      <th>DateCreated</th>\n",
       "      <th>NumberOfPictures</th>\n",
       "      <th>PostalCode</th>\n",
       "      <th>LastSeen</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>24/03/2016 11:52</td>\n",
       "      <td>480</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1993</td>\n",
       "      <td>manual</td>\n",
       "      <td>0</td>\n",
       "      <td>golf</td>\n",
       "      <td>150000</td>\n",
       "      <td>0</td>\n",
       "      <td>petrol</td>\n",
       "      <td>volkswagen</td>\n",
       "      <td>NaN</td>\n",
       "      <td>24/03/2016 00:00</td>\n",
       "      <td>0</td>\n",
       "      <td>70435</td>\n",
       "      <td>07/04/2016 03:16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>24/03/2016 10:58</td>\n",
       "      <td>18300</td>\n",
       "      <td>coupe</td>\n",
       "      <td>2011</td>\n",
       "      <td>manual</td>\n",
       "      <td>190</td>\n",
       "      <td>NaN</td>\n",
       "      <td>125000</td>\n",
       "      <td>5</td>\n",
       "      <td>gasoline</td>\n",
       "      <td>audi</td>\n",
       "      <td>yes</td>\n",
       "      <td>24/03/2016 00:00</td>\n",
       "      <td>0</td>\n",
       "      <td>66954</td>\n",
       "      <td>07/04/2016 01:46</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>14/03/2016 12:52</td>\n",
       "      <td>9800</td>\n",
       "      <td>suv</td>\n",
       "      <td>2004</td>\n",
       "      <td>auto</td>\n",
       "      <td>163</td>\n",
       "      <td>grand</td>\n",
       "      <td>125000</td>\n",
       "      <td>8</td>\n",
       "      <td>gasoline</td>\n",
       "      <td>jeep</td>\n",
       "      <td>NaN</td>\n",
       "      <td>14/03/2016 00:00</td>\n",
       "      <td>0</td>\n",
       "      <td>90480</td>\n",
       "      <td>05/04/2016 12:47</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>17/03/2016 16:54</td>\n",
       "      <td>1500</td>\n",
       "      <td>small</td>\n",
       "      <td>2001</td>\n",
       "      <td>manual</td>\n",
       "      <td>75</td>\n",
       "      <td>golf</td>\n",
       "      <td>150000</td>\n",
       "      <td>6</td>\n",
       "      <td>petrol</td>\n",
       "      <td>volkswagen</td>\n",
       "      <td>no</td>\n",
       "      <td>17/03/2016 00:00</td>\n",
       "      <td>0</td>\n",
       "      <td>91074</td>\n",
       "      <td>17/03/2016 17:40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>31/03/2016 17:25</td>\n",
       "      <td>3600</td>\n",
       "      <td>small</td>\n",
       "      <td>2008</td>\n",
       "      <td>manual</td>\n",
       "      <td>69</td>\n",
       "      <td>fabia</td>\n",
       "      <td>90000</td>\n",
       "      <td>7</td>\n",
       "      <td>gasoline</td>\n",
       "      <td>skoda</td>\n",
       "      <td>no</td>\n",
       "      <td>31/03/2016 00:00</td>\n",
       "      <td>0</td>\n",
       "      <td>60437</td>\n",
       "      <td>06/04/2016 10:17</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        DateCrawled  Price VehicleType  RegistrationYear Gearbox  Power  \\\n",
       "0  24/03/2016 11:52    480         NaN              1993  manual      0   \n",
       "1  24/03/2016 10:58  18300       coupe              2011  manual    190   \n",
       "2  14/03/2016 12:52   9800         suv              2004    auto    163   \n",
       "3  17/03/2016 16:54   1500       small              2001  manual     75   \n",
       "4  31/03/2016 17:25   3600       small              2008  manual     69   \n",
       "\n",
       "   Model  Mileage  RegistrationMonth  FuelType       Brand NotRepaired  \\\n",
       "0   golf   150000                  0    petrol  volkswagen         NaN   \n",
       "1    NaN   125000                  5  gasoline        audi         yes   \n",
       "2  grand   125000                  8  gasoline        jeep         NaN   \n",
       "3   golf   150000                  6    petrol  volkswagen          no   \n",
       "4  fabia    90000                  7  gasoline       skoda          no   \n",
       "\n",
       "        DateCreated  NumberOfPictures  PostalCode          LastSeen  \n",
       "0  24/03/2016 00:00                 0       70435  07/04/2016 03:16  \n",
       "1  24/03/2016 00:00                 0       66954  07/04/2016 01:46  \n",
       "2  14/03/2016 00:00                 0       90480  05/04/2016 12:47  \n",
       "3  17/03/2016 00:00                 0       91074  17/03/2016 17:40  \n",
       "4  31/03/2016 00:00                 0       60437  06/04/2016 10:17  "
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Load dataset\n",
    "path = 'https://practicum-content.s3.us-west-1.amazonaws.com/datasets/car_data.csv'\n",
    "df = pd.read_csv(path)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 354369 entries, 0 to 354368\n",
      "Data columns (total 16 columns):\n",
      " #   Column             Non-Null Count   Dtype \n",
      "---  ------             --------------   ----- \n",
      " 0   DateCrawled        354369 non-null  object\n",
      " 1   Price              354369 non-null  int64 \n",
      " 2   VehicleType        316879 non-null  object\n",
      " 3   RegistrationYear   354369 non-null  int64 \n",
      " 4   Gearbox            334536 non-null  object\n",
      " 5   Power              354369 non-null  int64 \n",
      " 6   Model              334664 non-null  object\n",
      " 7   Mileage            354369 non-null  int64 \n",
      " 8   RegistrationMonth  354369 non-null  int64 \n",
      " 9   FuelType           321474 non-null  object\n",
      " 10  Brand              354369 non-null  object\n",
      " 11  NotRepaired        283215 non-null  object\n",
      " 12  DateCreated        354369 non-null  object\n",
      " 13  NumberOfPictures   354369 non-null  int64 \n",
      " 14  PostalCode         354369 non-null  int64 \n",
      " 15  LastSeen           354369 non-null  object\n",
      "dtypes: int64(7), object(9)\n",
      "memory usage: 43.3+ MB\n"
     ]
    }
   ],
   "source": [
    "# Mendapatkan informasi data\n",
    "df.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert-info\">\n",
    "    \n",
    "\n",
    "<b>Note</b>\n",
    "\n",
    "<b>DateCrawled</b> — tanggal ketika profil diunduh dari database - data is OK <br>\n",
    "<b>VehicleType</b> — jenis bodi kendaraan - Terdapat data yang hilang, Perlu dicek lebih lanjut <br>\n",
    "<b>RegistrationYear</b> — tahun registrasi kendaraan - Data is OK! <br>\n",
    "<b>Gearbox</b> — jenis transmisi - Terdapat data yang hilang, Perlu di cek lebih lanjut <br>\n",
    "<b>Power</b> — daya (hp) - Data is OK<br>\n",
    "<b>Model</b> — model kendaraan - Terdapat data yang hilang, perlu dicek lebih lanjut <br>\n",
    "<b>Mileage</b> — jarak tempuh (diukur dalam satuan km berdasarkan dataset regional tertentu) - Data is OK! <br>\n",
    "<b>RegistrationMonth</b> — bulan registrasi kendaraan - Data is OK!<br>\n",
    "<b>FuelType</b> — jenis bahan bakar - Terdapat data yang hilang, Perlu di cek lebih lanjut. <br>\n",
    "<b>Brand</b> — merek kendaraan - Data is OK!<br>\n",
    "<b>NotRepaired</b> — apakah kendaraan pernah direparasi sebelumnya - Terdapat data yang hilang, Perlu dicek lebih lanjut <br>\n",
    "<b>DateCreated</b> — tanggal pembuatan profil - Data is OK! <br>\n",
    "<b>NumberOfPictures</b> — jumlah gambar kendaraan - Data is OK!<br>\n",
    "<b>PostalCode</b> — kode pos pemilik profil (pengguna) - Data is OK! <br>\n",
    "<b>LastSeen</b> — tanggal aktivitas terakhir pengguna - Data is OK! <br>\n",
    "\n",
    "\n",
    "<b>Price</b> — harga (dalam satuan Euro) - Data is OK! <br>\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Data yang hilang"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### VehicleType"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "sedan          91457\n",
      "small          79831\n",
      "wagon          65166\n",
      "bus            28775\n",
      "convertible    20203\n",
      "coupe          16163\n",
      "suv            11996\n",
      "other           3288\n",
      "Name: VehicleType, dtype: int64\n",
      "data yang hilang pada kolom VehicleType    : 37490\n"
     ]
    }
   ],
   "source": [
    "# Melihat data Vehicle Type.\n",
    "print(df.VehicleType.value_counts())\n",
    "print(f'data yang hilang pada kolom VehicleType    : {df.VehicleType.isna().sum()}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert-info\">\n",
    "    \n",
    "\n",
    "<b>Insight</b>\n",
    "\n",
    "- Data Tipe Kendaraan yang hilang sebanyak 37,484 data.\n",
    "- 3 Tipe data terbanyak yaitu Sedan, Small dan Wagon."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Mengubah data yang hilang pada kolom Vehicle Type menjadi Other\n",
    "df.VehicleType = df.VehicleType.fillna('other')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "sedan          91457\n",
      "small          79831\n",
      "wagon          65166\n",
      "other          40778\n",
      "bus            28775\n",
      "convertible    20203\n",
      "coupe          16163\n",
      "suv            11996\n",
      "Name: VehicleType, dtype: int64\n",
      "data yang hilang pada kolom VehicleType    : 0\n"
     ]
    }
   ],
   "source": [
    "# Melakukan pengecekan data kembali\n",
    "print(df.VehicleType.value_counts())\n",
    "print(f'data yang hilang pada kolom VehicleType    : {df.VehicleType.isna().sum()}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert-info\">\n",
    "    \n",
    "\n",
    "<b>Note</b>\n",
    "\n",
    "- Data yang hilang sudah berhasil diisi"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Gearbox"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "manual    268251\n",
      "auto       66285\n",
      "Name: Gearbox, dtype: int64\n",
      "data yang hilang pada kolom Gearbox    : 19833\n"
     ]
    }
   ],
   "source": [
    "# Melihat data Jenis Transmisi.\n",
    "print(df.Gearbox.value_counts())\n",
    "print(f'data yang hilang pada kolom Gearbox    : {df.Gearbox.isna().sum()}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert-info\">\n",
    "    \n",
    "\n",
    "<b>Insight</b>\n",
    "\n",
    "- Rata-rata jenis transmisi kendaraan yaitu Manual\n",
    "- Data yang hilang sebanyak 19,830 data dari 354,369 data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Mengubah data yang hilang pada kolom Gearbox menjadi nilai rata-ratanya yaitu Manual\n",
    "df.Gearbox = df.Gearbox.fillna('manual')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "manual    288084\n",
      "auto       66285\n",
      "Name: Gearbox, dtype: int64\n",
      "data yang hilang pada kolom Gearbox    : 0\n"
     ]
    }
   ],
   "source": [
    "# Melakukan pengecekan data kembali.\n",
    "print(df.Gearbox.value_counts())\n",
    "print(f'data yang hilang pada kolom Gearbox    : {df.Gearbox.isna().sum()}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert-info\">\n",
    "    \n",
    "\n",
    "<b>Note</b>\n",
    "\n",
    "- Data yang hilang sudah berhasil diisi"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "golf                  29232\n",
       "other                 24421\n",
       "3er                   19761\n",
       "polo                  13066\n",
       "corsa                 12570\n",
       "                      ...  \n",
       "i3                        8\n",
       "serie_3                   4\n",
       "rangerover                4\n",
       "range_rover_evoque        2\n",
       "serie_1                   2\n",
       "Name: Model, Length: 250, dtype: int64"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Melihat data Model.\n",
    "df.Model.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Data yang hilang pada kolom model    : 19705\n"
     ]
    }
   ],
   "source": [
    "# Melakukan pengecekan banyaknya data yang hilang pada kolom Model\n",
    "print(f'Data yang hilang pada kolom model    : {df.Model.isna().sum()}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert-info\">\n",
    "    \n",
    "\n",
    "<b>Insight</b>\n",
    "\n",
    "- Data mencatat, kendaraan memiliki lebih dari 200 Model\n",
    "- Data model kendaraan yang hilang sebanyak 19,701 Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Mengubah data yang hilang pada kolom Model menjadi other\n",
    "df.Model = df.Model.fillna('other')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "other                 44126\n",
      "golf                  29232\n",
      "3er                   19761\n",
      "polo                  13066\n",
      "corsa                 12570\n",
      "                      ...  \n",
      "i3                        8\n",
      "serie_3                   4\n",
      "rangerover                4\n",
      "range_rover_evoque        2\n",
      "serie_1                   2\n",
      "Name: Model, Length: 250, dtype: int64\n",
      "Data yang hilang pada kolom model    : 0\n"
     ]
    }
   ],
   "source": [
    "# Melakukan pengecekan kembali\n",
    "print(df.Model.value_counts())\n",
    "print(f'Data yang hilang pada kolom model    : {df.Model.isna().sum()}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert-info\">\n",
    "    \n",
    "\n",
    "<b>Note</b>\n",
    "\n",
    "- Data yang hilang sudah berhasil diisi"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### FuelType"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "petrol      216352\n",
      "gasoline     98720\n",
      "lpg           5310\n",
      "cng            565\n",
      "hybrid         233\n",
      "other          204\n",
      "electric        90\n",
      "Name: FuelType, dtype: int64\n",
      "Data yang hilang pada kolom Fueltype    : 32895\n"
     ]
    }
   ],
   "source": [
    "# Melihat data Jenis Bahan Bakar .\n",
    "print(df.FuelType.value_counts())\n",
    "print(f'Data yang hilang pada kolom Fueltype    : {df.FuelType.isna().sum()}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert-info\">\n",
    "    \n",
    "\n",
    "<b>Insight</b>\n",
    "\n",
    "- Rata-rata jenis bahan bakar kendaraan, yaitu Petrol\n",
    "- data yang hilang sebanyak 32,889 data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Mengubah data yang hilang pada kolom FuelType menjadi other\n",
    "df.FuelType = df.FuelType.fillna('other')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "petrol      216352\n",
      "gasoline     98720\n",
      "other        33099\n",
      "lpg           5310\n",
      "cng            565\n",
      "hybrid         233\n",
      "electric        90\n",
      "Name: FuelType, dtype: int64\n",
      "Data yang hilang pada kolom Fueltype    : 0\n"
     ]
    }
   ],
   "source": [
    "# Melakukan pengecekan kembali.\n",
    "print(df.FuelType.value_counts())\n",
    "print(f'Data yang hilang pada kolom Fueltype    : {df.FuelType.isna().sum()}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert-info\">\n",
    "    \n",
    "\n",
    "<b>Note</b>\n",
    "\n",
    "- Data yang hilang sudah berhasil diisi"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### NotRepaired"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "no     247161\n",
      "yes     36054\n",
      "Name: NotRepaired, dtype: int64\n",
      "Data yang hilang pada kolom Fueltype    : 71154\n"
     ]
    }
   ],
   "source": [
    "# Melihat data reparasi .\n",
    "print(df.NotRepaired.value_counts())\n",
    "print(f'Data yang hilang pada kolom Fueltype    : {df.NotRepaired.isna().sum()}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert-info\">\n",
    "    \n",
    "\n",
    "<b>Insight</b>\n",
    "\n",
    "- Rata-rata kendaraan belum pernah direparasi\n",
    "- data yang hilang sebanyak 71,145 data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Mengubah data yang hilang pada kolom NotRepaired menjadi nilai rata-ratanya yaitu : no\n",
    "df.NotRepaired = df.NotRepaired.fillna('no')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "no     318315\n",
      "yes     36054\n",
      "Name: NotRepaired, dtype: int64\n",
      "Data yang hilang pada kolom Fueltype    : 0\n"
     ]
    }
   ],
   "source": [
    "# Melakukan pengecekan data kembali\n",
    "print(df.NotRepaired.value_counts())\n",
    "print(f'Data yang hilang pada kolom Fueltype    : {df.NotRepaired.isna().sum()}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert-info\">\n",
    "    \n",
    "\n",
    "<b>Note</b>\n",
    "\n",
    "- Data yang hilang sudah berhasil diisi"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>DateCrawled</th>\n",
       "      <th>Price</th>\n",
       "      <th>VehicleType</th>\n",
       "      <th>RegistrationYear</th>\n",
       "      <th>Gearbox</th>\n",
       "      <th>Power</th>\n",
       "      <th>Model</th>\n",
       "      <th>Mileage</th>\n",
       "      <th>RegistrationMonth</th>\n",
       "      <th>FuelType</th>\n",
       "      <th>Brand</th>\n",
       "      <th>NotRepaired</th>\n",
       "      <th>DateCreated</th>\n",
       "      <th>NumberOfPictures</th>\n",
       "      <th>PostalCode</th>\n",
       "      <th>LastSeen</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>354364</th>\n",
       "      <td>21/03/2016 09:50</td>\n",
       "      <td>0</td>\n",
       "      <td>other</td>\n",
       "      <td>2005</td>\n",
       "      <td>manual</td>\n",
       "      <td>0</td>\n",
       "      <td>colt</td>\n",
       "      <td>150000</td>\n",
       "      <td>7</td>\n",
       "      <td>petrol</td>\n",
       "      <td>mitsubishi</td>\n",
       "      <td>yes</td>\n",
       "      <td>21/03/2016 00:00</td>\n",
       "      <td>0</td>\n",
       "      <td>2694</td>\n",
       "      <td>21/03/2016 10:42</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>354365</th>\n",
       "      <td>14/03/2016 17:48</td>\n",
       "      <td>2200</td>\n",
       "      <td>other</td>\n",
       "      <td>2005</td>\n",
       "      <td>manual</td>\n",
       "      <td>0</td>\n",
       "      <td>other</td>\n",
       "      <td>20000</td>\n",
       "      <td>1</td>\n",
       "      <td>other</td>\n",
       "      <td>sonstige_autos</td>\n",
       "      <td>no</td>\n",
       "      <td>14/03/2016 00:00</td>\n",
       "      <td>0</td>\n",
       "      <td>39576</td>\n",
       "      <td>06/04/2016 00:46</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>354366</th>\n",
       "      <td>05/03/2016 19:56</td>\n",
       "      <td>1199</td>\n",
       "      <td>convertible</td>\n",
       "      <td>2000</td>\n",
       "      <td>auto</td>\n",
       "      <td>101</td>\n",
       "      <td>fortwo</td>\n",
       "      <td>125000</td>\n",
       "      <td>3</td>\n",
       "      <td>petrol</td>\n",
       "      <td>smart</td>\n",
       "      <td>no</td>\n",
       "      <td>05/03/2016 00:00</td>\n",
       "      <td>0</td>\n",
       "      <td>26135</td>\n",
       "      <td>11/03/2016 18:17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>354367</th>\n",
       "      <td>19/03/2016 18:57</td>\n",
       "      <td>9200</td>\n",
       "      <td>bus</td>\n",
       "      <td>1996</td>\n",
       "      <td>manual</td>\n",
       "      <td>102</td>\n",
       "      <td>transporter</td>\n",
       "      <td>150000</td>\n",
       "      <td>3</td>\n",
       "      <td>gasoline</td>\n",
       "      <td>volkswagen</td>\n",
       "      <td>no</td>\n",
       "      <td>19/03/2016 00:00</td>\n",
       "      <td>0</td>\n",
       "      <td>87439</td>\n",
       "      <td>07/04/2016 07:15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>354368</th>\n",
       "      <td>20/03/2016 19:41</td>\n",
       "      <td>3400</td>\n",
       "      <td>wagon</td>\n",
       "      <td>2002</td>\n",
       "      <td>manual</td>\n",
       "      <td>100</td>\n",
       "      <td>golf</td>\n",
       "      <td>150000</td>\n",
       "      <td>6</td>\n",
       "      <td>gasoline</td>\n",
       "      <td>volkswagen</td>\n",
       "      <td>no</td>\n",
       "      <td>20/03/2016 00:00</td>\n",
       "      <td>0</td>\n",
       "      <td>40764</td>\n",
       "      <td>24/03/2016 12:45</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             DateCrawled  Price  VehicleType  RegistrationYear Gearbox  Power  \\\n",
       "354364  21/03/2016 09:50      0        other              2005  manual      0   \n",
       "354365  14/03/2016 17:48   2200        other              2005  manual      0   \n",
       "354366  05/03/2016 19:56   1199  convertible              2000    auto    101   \n",
       "354367  19/03/2016 18:57   9200          bus              1996  manual    102   \n",
       "354368  20/03/2016 19:41   3400        wagon              2002  manual    100   \n",
       "\n",
       "              Model  Mileage  RegistrationMonth  FuelType           Brand  \\\n",
       "354364         colt   150000                  7    petrol      mitsubishi   \n",
       "354365        other    20000                  1     other  sonstige_autos   \n",
       "354366       fortwo   125000                  3    petrol           smart   \n",
       "354367  transporter   150000                  3  gasoline      volkswagen   \n",
       "354368         golf   150000                  6  gasoline      volkswagen   \n",
       "\n",
       "       NotRepaired       DateCreated  NumberOfPictures  PostalCode  \\\n",
       "354364         yes  21/03/2016 00:00                 0        2694   \n",
       "354365          no  14/03/2016 00:00                 0       39576   \n",
       "354366          no  05/03/2016 00:00                 0       26135   \n",
       "354367          no  19/03/2016 00:00                 0       87439   \n",
       "354368          no  20/03/2016 00:00                 0       40764   \n",
       "\n",
       "                LastSeen  \n",
       "354364  21/03/2016 10:42  \n",
       "354365  06/04/2016 00:46  \n",
       "354366  11/03/2016 18:17  \n",
       "354367  07/04/2016 07:15  \n",
       "354368  24/03/2016 12:45  "
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.tail(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Data Duplikat"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "294"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Melakukan pengecekan data duplikat\n",
    "df.duplicated().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Menghapus data duplikat agar tidak terjadi informasi redundant\n",
    "df = df.drop_duplicates()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Melakukan pengecekan kembali\n",
    "df.duplicated().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Data Preprocessing\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1080x1080 with 5 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Melihat persebaran data kategorikal\n",
    "plt.subplots(figsize=(15, 15))\n",
    " \n",
    "# melihat persebaran data categorikal\n",
    "categorical = ['VehicleType', 'Gearbox', 'Brand', 'FuelType', 'NotRepaired']\n",
    " \n",
    "for i, col in enumerate(categorical):\n",
    "    plt.subplot(3, 3, i + 1)\n",
    "    sns.countplot(data=df, x=col)\n",
    "     \n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert-info\">\n",
    "    \n",
    "\n",
    "<b>Note</b>\n",
    "\n",
    "- Varian produk Brand terlalu banyak, sehingga nantinya tidak digunakan sebagai fitur dalam model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/conda/lib/python3.9/site-packages/seaborn/distributions.py:2557: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n",
      "/opt/conda/lib/python3.9/site-packages/seaborn/distributions.py:2557: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n",
      "/opt/conda/lib/python3.9/site-packages/seaborn/distributions.py:2557: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n",
      "/opt/conda/lib/python3.9/site-packages/seaborn/distributions.py:2557: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n",
      "/opt/conda/lib/python3.9/site-packages/seaborn/distributions.py:2557: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n",
      "/opt/conda/lib/python3.9/site-packages/seaborn/distributions.py:306: UserWarning: Dataset has 0 variance; skipping density estimate.\n",
      "  warnings.warn(msg, UserWarning)\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1080x1080 with 5 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Melihat persebaran data kategorikal\n",
    "plt.subplots(figsize=(15, 15))\n",
    " \n",
    "# melihat persebaran data categorikal\n",
    "numerical = ['RegistrationYear', 'Power', 'Mileage', 'RegistrationMonth','NumberOfPictures']\n",
    " \n",
    "for i, col in enumerate(numerical):\n",
    "    plt.subplot(3, 3, i + 1)\n",
    "    sns.distplot(df[col])\n",
    "     \n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Price</th>\n",
       "      <th>RegistrationYear</th>\n",
       "      <th>Power</th>\n",
       "      <th>Mileage</th>\n",
       "      <th>RegistrationMonth</th>\n",
       "      <th>NumberOfPictures</th>\n",
       "      <th>PostalCode</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>354075.000000</td>\n",
       "      <td>354075.000000</td>\n",
       "      <td>354075.000000</td>\n",
       "      <td>354075.000000</td>\n",
       "      <td>354075.000000</td>\n",
       "      <td>354075.0</td>\n",
       "      <td>354075.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>4416.407818</td>\n",
       "      <td>2004.235693</td>\n",
       "      <td>110.088216</td>\n",
       "      <td>128211.282920</td>\n",
       "      <td>5.714086</td>\n",
       "      <td>0.0</td>\n",
       "      <td>50506.857875</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>4514.359273</td>\n",
       "      <td>90.265205</td>\n",
       "      <td>189.922740</td>\n",
       "      <td>37907.128915</td>\n",
       "      <td>3.726688</td>\n",
       "      <td>0.0</td>\n",
       "      <td>25784.403198</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>1000.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>5000.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1067.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>1050.000000</td>\n",
       "      <td>1999.000000</td>\n",
       "      <td>69.000000</td>\n",
       "      <td>125000.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>30165.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>2700.000000</td>\n",
       "      <td>2003.000000</td>\n",
       "      <td>105.000000</td>\n",
       "      <td>150000.000000</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>49406.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>6400.000000</td>\n",
       "      <td>2008.000000</td>\n",
       "      <td>143.000000</td>\n",
       "      <td>150000.000000</td>\n",
       "      <td>9.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>71083.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>20000.000000</td>\n",
       "      <td>9999.000000</td>\n",
       "      <td>20000.000000</td>\n",
       "      <td>150000.000000</td>\n",
       "      <td>12.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>99998.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               Price  RegistrationYear          Power        Mileage  \\\n",
       "count  354075.000000     354075.000000  354075.000000  354075.000000   \n",
       "mean     4416.407818       2004.235693     110.088216  128211.282920   \n",
       "std      4514.359273         90.265205     189.922740   37907.128915   \n",
       "min         0.000000       1000.000000       0.000000    5000.000000   \n",
       "25%      1050.000000       1999.000000      69.000000  125000.000000   \n",
       "50%      2700.000000       2003.000000     105.000000  150000.000000   \n",
       "75%      6400.000000       2008.000000     143.000000  150000.000000   \n",
       "max     20000.000000       9999.000000   20000.000000  150000.000000   \n",
       "\n",
       "       RegistrationMonth  NumberOfPictures     PostalCode  \n",
       "count      354075.000000          354075.0  354075.000000  \n",
       "mean            5.714086               0.0   50506.857875  \n",
       "std             3.726688               0.0   25784.403198  \n",
       "min             0.000000               0.0    1067.000000  \n",
       "25%             3.000000               0.0   30165.000000  \n",
       "50%             6.000000               0.0   49406.000000  \n",
       "75%             9.000000               0.0   71083.000000  \n",
       "max            12.000000               0.0   99998.000000  "
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    354075\n",
       "Name: NumberOfPictures, dtype: int64"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['NumberOfPictures'].value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert-info\">\n",
    "    \n",
    "\n",
    "<b>Note</b>\n",
    "\n",
    "- Seluruh datanya yaitu 0, sehingga nantinya tidak dapat digunakan dalam proses pelatihan model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Memotong outlier pada kolom Registration Year\n",
    "\n",
    "df.drop(df[(df['RegistrationYear']<1924) & (df['RegistrationYear']>2024)].index, inplace=True)\n",
    "df.reset_index(drop=True, inplace=True)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['age_of_vehicle'] = 2024 - df['RegistrationYear']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0         31\n",
       "1         13\n",
       "2         20\n",
       "3         23\n",
       "4         16\n",
       "          ..\n",
       "354070    19\n",
       "354071    19\n",
       "354072    24\n",
       "354073    28\n",
       "354074    22\n",
       "Name: age_of_vehicle, Length: 354075, dtype: int64"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['age_of_vehicle']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>DateCrawled</th>\n",
       "      <th>Price</th>\n",
       "      <th>VehicleType</th>\n",
       "      <th>RegistrationYear</th>\n",
       "      <th>Gearbox</th>\n",
       "      <th>Power</th>\n",
       "      <th>Model</th>\n",
       "      <th>Mileage</th>\n",
       "      <th>RegistrationMonth</th>\n",
       "      <th>FuelType</th>\n",
       "      <th>Brand</th>\n",
       "      <th>NotRepaired</th>\n",
       "      <th>DateCreated</th>\n",
       "      <th>NumberOfPictures</th>\n",
       "      <th>PostalCode</th>\n",
       "      <th>LastSeen</th>\n",
       "      <th>age_of_vehicle</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>24/03/2016 11:52</td>\n",
       "      <td>480</td>\n",
       "      <td>other</td>\n",
       "      <td>1993</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>golf</td>\n",
       "      <td>150000</td>\n",
       "      <td>0</td>\n",
       "      <td>petrol</td>\n",
       "      <td>volkswagen</td>\n",
       "      <td>0</td>\n",
       "      <td>24/03/2016 00:00</td>\n",
       "      <td>0</td>\n",
       "      <td>70435</td>\n",
       "      <td>07/04/2016 03:16</td>\n",
       "      <td>31</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>24/03/2016 10:58</td>\n",
       "      <td>18300</td>\n",
       "      <td>coupe</td>\n",
       "      <td>2011</td>\n",
       "      <td>0</td>\n",
       "      <td>190</td>\n",
       "      <td>other</td>\n",
       "      <td>125000</td>\n",
       "      <td>5</td>\n",
       "      <td>gasoline</td>\n",
       "      <td>audi</td>\n",
       "      <td>1</td>\n",
       "      <td>24/03/2016 00:00</td>\n",
       "      <td>0</td>\n",
       "      <td>66954</td>\n",
       "      <td>07/04/2016 01:46</td>\n",
       "      <td>13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>14/03/2016 12:52</td>\n",
       "      <td>9800</td>\n",
       "      <td>suv</td>\n",
       "      <td>2004</td>\n",
       "      <td>1</td>\n",
       "      <td>163</td>\n",
       "      <td>grand</td>\n",
       "      <td>125000</td>\n",
       "      <td>8</td>\n",
       "      <td>gasoline</td>\n",
       "      <td>jeep</td>\n",
       "      <td>0</td>\n",
       "      <td>14/03/2016 00:00</td>\n",
       "      <td>0</td>\n",
       "      <td>90480</td>\n",
       "      <td>05/04/2016 12:47</td>\n",
       "      <td>20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>17/03/2016 16:54</td>\n",
       "      <td>1500</td>\n",
       "      <td>small</td>\n",
       "      <td>2001</td>\n",
       "      <td>0</td>\n",
       "      <td>75</td>\n",
       "      <td>golf</td>\n",
       "      <td>150000</td>\n",
       "      <td>6</td>\n",
       "      <td>petrol</td>\n",
       "      <td>volkswagen</td>\n",
       "      <td>0</td>\n",
       "      <td>17/03/2016 00:00</td>\n",
       "      <td>0</td>\n",
       "      <td>91074</td>\n",
       "      <td>17/03/2016 17:40</td>\n",
       "      <td>23</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>31/03/2016 17:25</td>\n",
       "      <td>3600</td>\n",
       "      <td>small</td>\n",
       "      <td>2008</td>\n",
       "      <td>0</td>\n",
       "      <td>69</td>\n",
       "      <td>fabia</td>\n",
       "      <td>90000</td>\n",
       "      <td>7</td>\n",
       "      <td>gasoline</td>\n",
       "      <td>skoda</td>\n",
       "      <td>0</td>\n",
       "      <td>31/03/2016 00:00</td>\n",
       "      <td>0</td>\n",
       "      <td>60437</td>\n",
       "      <td>06/04/2016 10:17</td>\n",
       "      <td>16</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        DateCrawled  Price VehicleType  RegistrationYear  Gearbox  Power  \\\n",
       "0  24/03/2016 11:52    480       other              1993        0      0   \n",
       "1  24/03/2016 10:58  18300       coupe              2011        0    190   \n",
       "2  14/03/2016 12:52   9800         suv              2004        1    163   \n",
       "3  17/03/2016 16:54   1500       small              2001        0     75   \n",
       "4  31/03/2016 17:25   3600       small              2008        0     69   \n",
       "\n",
       "   Model  Mileage  RegistrationMonth  FuelType       Brand  NotRepaired  \\\n",
       "0   golf   150000                  0    petrol  volkswagen            0   \n",
       "1  other   125000                  5  gasoline        audi            1   \n",
       "2  grand   125000                  8  gasoline        jeep            0   \n",
       "3   golf   150000                  6    petrol  volkswagen            0   \n",
       "4  fabia    90000                  7  gasoline       skoda            0   \n",
       "\n",
       "        DateCreated  NumberOfPictures  PostalCode          LastSeen  \\\n",
       "0  24/03/2016 00:00                 0       70435  07/04/2016 03:16   \n",
       "1  24/03/2016 00:00                 0       66954  07/04/2016 01:46   \n",
       "2  14/03/2016 00:00                 0       90480  05/04/2016 12:47   \n",
       "3  17/03/2016 00:00                 0       91074  17/03/2016 17:40   \n",
       "4  31/03/2016 00:00                 0       60437  06/04/2016 10:17   \n",
       "\n",
       "   age_of_vehicle  \n",
       "0              31  \n",
       "1              13  \n",
       "2              20  \n",
       "3              23  \n",
       "4              16  "
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Melakukan mapping pada data kategorikal\n",
    " \n",
    "# Mapping kolom \"Gearbox\" ('Manual' adalah 0, 'Auto' adalah 1)\n",
    "df['Gearbox'] = df['Gearbox'].map({'manual': 0, 'auto': 1})\n",
    " \n",
    "# Mapping kolom \"NotRepaired\" ('No' adalah 0, 'Yes' adalah 1)\n",
    "df['NotRepaired'] = df['NotRepaired'].map({'no': 0, 'yes': 1})\n",
    " \n",
    "# Display the DataFrame's first few rows to show the transformations\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 354075 entries, 0 to 354074\n",
      "Data columns (total 17 columns):\n",
      " #   Column             Non-Null Count   Dtype \n",
      "---  ------             --------------   ----- \n",
      " 0   DateCrawled        354075 non-null  object\n",
      " 1   Price              354075 non-null  int64 \n",
      " 2   VehicleType        354075 non-null  object\n",
      " 3   RegistrationYear   354075 non-null  int64 \n",
      " 4   Gearbox            354075 non-null  int64 \n",
      " 5   Power              354075 non-null  int64 \n",
      " 6   Model              354075 non-null  object\n",
      " 7   Mileage            354075 non-null  int64 \n",
      " 8   RegistrationMonth  354075 non-null  int64 \n",
      " 9   FuelType           354075 non-null  object\n",
      " 10  Brand              354075 non-null  object\n",
      " 11  NotRepaired        354075 non-null  int64 \n",
      " 12  DateCreated        354075 non-null  object\n",
      " 13  NumberOfPictures   354075 non-null  int64 \n",
      " 14  PostalCode         354075 non-null  int64 \n",
      " 15  LastSeen           354075 non-null  object\n",
      " 16  age_of_vehicle     354075 non-null  int64 \n",
      "dtypes: int64(10), object(7)\n",
      "memory usage: 45.9+ MB\n"
     ]
    }
   ],
   "source": [
    "df.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Pelatihan Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Menghilangkan type data kategoric yang memiliki cukup banyak varian dan tidak dapat dijadikan feature.\n",
    "df = df.drop(['DateCrawled' , 'RegistrationYear' , 'DateCreated', 'RegistrationMonth', 'Model', 'LastSeen','NumberOfPictures', 'PostalCode'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "# One hot encoding ke data yang memiliki tipe data kategorikal nominal\n",
    "data = pd.get_dummies(data=df, drop_first=True).astype('int')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Membagi X train dan y train\n",
    "X = data.drop('Price', axis=1)\n",
    "y = data['Price']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Gearbox</th>\n",
       "      <th>Power</th>\n",
       "      <th>Mileage</th>\n",
       "      <th>NotRepaired</th>\n",
       "      <th>age_of_vehicle</th>\n",
       "      <th>VehicleType_convertible</th>\n",
       "      <th>VehicleType_coupe</th>\n",
       "      <th>VehicleType_other</th>\n",
       "      <th>VehicleType_sedan</th>\n",
       "      <th>VehicleType_small</th>\n",
       "      <th>...</th>\n",
       "      <th>Brand_seat</th>\n",
       "      <th>Brand_skoda</th>\n",
       "      <th>Brand_smart</th>\n",
       "      <th>Brand_sonstige_autos</th>\n",
       "      <th>Brand_subaru</th>\n",
       "      <th>Brand_suzuki</th>\n",
       "      <th>Brand_toyota</th>\n",
       "      <th>Brand_trabant</th>\n",
       "      <th>Brand_volkswagen</th>\n",
       "      <th>Brand_volvo</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>150000</td>\n",
       "      <td>0</td>\n",
       "      <td>31</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>190</td>\n",
       "      <td>125000</td>\n",
       "      <td>1</td>\n",
       "      <td>13</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>163</td>\n",
       "      <td>125000</td>\n",
       "      <td>0</td>\n",
       "      <td>20</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>75</td>\n",
       "      <td>150000</td>\n",
       "      <td>0</td>\n",
       "      <td>23</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</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>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>69</td>\n",
       "      <td>90000</td>\n",
       "      <td>0</td>\n",
       "      <td>16</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>354070</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>150000</td>\n",
       "      <td>1</td>\n",
       "      <td>19</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>354071</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>20000</td>\n",
       "      <td>0</td>\n",
       "      <td>19</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>354072</th>\n",
       "      <td>1</td>\n",
       "      <td>101</td>\n",
       "      <td>125000</td>\n",
       "      <td>0</td>\n",
       "      <td>24</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>354073</th>\n",
       "      <td>0</td>\n",
       "      <td>102</td>\n",
       "      <td>150000</td>\n",
       "      <td>0</td>\n",
       "      <td>28</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>354074</th>\n",
       "      <td>0</td>\n",
       "      <td>100</td>\n",
       "      <td>150000</td>\n",
       "      <td>0</td>\n",
       "      <td>22</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>354075 rows × 57 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        Gearbox  Power  Mileage  NotRepaired  age_of_vehicle  \\\n",
       "0             0      0   150000            0              31   \n",
       "1             0    190   125000            1              13   \n",
       "2             1    163   125000            0              20   \n",
       "3             0     75   150000            0              23   \n",
       "4             0     69    90000            0              16   \n",
       "...         ...    ...      ...          ...             ...   \n",
       "354070        0      0   150000            1              19   \n",
       "354071        0      0    20000            0              19   \n",
       "354072        1    101   125000            0              24   \n",
       "354073        0    102   150000            0              28   \n",
       "354074        0    100   150000            0              22   \n",
       "\n",
       "        VehicleType_convertible  VehicleType_coupe  VehicleType_other  \\\n",
       "0                             0                  0                  1   \n",
       "1                             0                  1                  0   \n",
       "2                             0                  0                  0   \n",
       "3                             0                  0                  0   \n",
       "4                             0                  0                  0   \n",
       "...                         ...                ...                ...   \n",
       "354070                        0                  0                  1   \n",
       "354071                        0                  0                  1   \n",
       "354072                        1                  0                  0   \n",
       "354073                        0                  0                  0   \n",
       "354074                        0                  0                  0   \n",
       "\n",
       "        VehicleType_sedan  VehicleType_small  ...  Brand_seat  Brand_skoda  \\\n",
       "0                       0                  0  ...           0            0   \n",
       "1                       0                  0  ...           0            0   \n",
       "2                       0                  0  ...           0            0   \n",
       "3                       0                  1  ...           0            0   \n",
       "4                       0                  1  ...           0            1   \n",
       "...                   ...                ...  ...         ...          ...   \n",
       "354070                  0                  0  ...           0            0   \n",
       "354071                  0                  0  ...           0            0   \n",
       "354072                  0                  0  ...           0            0   \n",
       "354073                  0                  0  ...           0            0   \n",
       "354074                  0                  0  ...           0            0   \n",
       "\n",
       "        Brand_smart  Brand_sonstige_autos  Brand_subaru  Brand_suzuki  \\\n",
       "0                 0                     0             0             0   \n",
       "1                 0                     0             0             0   \n",
       "2                 0                     0             0             0   \n",
       "3                 0                     0             0             0   \n",
       "4                 0                     0             0             0   \n",
       "...             ...                   ...           ...           ...   \n",
       "354070            0                     0             0             0   \n",
       "354071            0                     1             0             0   \n",
       "354072            1                     0             0             0   \n",
       "354073            0                     0             0             0   \n",
       "354074            0                     0             0             0   \n",
       "\n",
       "        Brand_toyota  Brand_trabant  Brand_volkswagen  Brand_volvo  \n",
       "0                  0              0                 1            0  \n",
       "1                  0              0                 0            0  \n",
       "2                  0              0                 0            0  \n",
       "3                  0              0                 1            0  \n",
       "4                  0              0                 0            0  \n",
       "...              ...            ...               ...          ...  \n",
       "354070             0              0                 0            0  \n",
       "354071             0              0                 0            0  \n",
       "354072             0              0                 0            0  \n",
       "354073             0              0                 1            0  \n",
       "354074             0              0                 1            0  \n",
       "\n",
       "[354075 rows x 57 columns]"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 354075 entries, 0 to 354074\n",
      "Data columns (total 57 columns):\n",
      " #   Column                   Non-Null Count   Dtype\n",
      "---  ------                   --------------   -----\n",
      " 0   Gearbox                  354075 non-null  int64\n",
      " 1   Power                    354075 non-null  int64\n",
      " 2   Mileage                  354075 non-null  int64\n",
      " 3   NotRepaired              354075 non-null  int64\n",
      " 4   age_of_vehicle           354075 non-null  int64\n",
      " 5   VehicleType_convertible  354075 non-null  int64\n",
      " 6   VehicleType_coupe        354075 non-null  int64\n",
      " 7   VehicleType_other        354075 non-null  int64\n",
      " 8   VehicleType_sedan        354075 non-null  int64\n",
      " 9   VehicleType_small        354075 non-null  int64\n",
      " 10  VehicleType_suv          354075 non-null  int64\n",
      " 11  VehicleType_wagon        354075 non-null  int64\n",
      " 12  FuelType_electric        354075 non-null  int64\n",
      " 13  FuelType_gasoline        354075 non-null  int64\n",
      " 14  FuelType_hybrid          354075 non-null  int64\n",
      " 15  FuelType_lpg             354075 non-null  int64\n",
      " 16  FuelType_other           354075 non-null  int64\n",
      " 17  FuelType_petrol          354075 non-null  int64\n",
      " 18  Brand_audi               354075 non-null  int64\n",
      " 19  Brand_bmw                354075 non-null  int64\n",
      " 20  Brand_chevrolet          354075 non-null  int64\n",
      " 21  Brand_chrysler           354075 non-null  int64\n",
      " 22  Brand_citroen            354075 non-null  int64\n",
      " 23  Brand_dacia              354075 non-null  int64\n",
      " 24  Brand_daewoo             354075 non-null  int64\n",
      " 25  Brand_daihatsu           354075 non-null  int64\n",
      " 26  Brand_fiat               354075 non-null  int64\n",
      " 27  Brand_ford               354075 non-null  int64\n",
      " 28  Brand_honda              354075 non-null  int64\n",
      " 29  Brand_hyundai            354075 non-null  int64\n",
      " 30  Brand_jaguar             354075 non-null  int64\n",
      " 31  Brand_jeep               354075 non-null  int64\n",
      " 32  Brand_kia                354075 non-null  int64\n",
      " 33  Brand_lada               354075 non-null  int64\n",
      " 34  Brand_lancia             354075 non-null  int64\n",
      " 35  Brand_land_rover         354075 non-null  int64\n",
      " 36  Brand_mazda              354075 non-null  int64\n",
      " 37  Brand_mercedes_benz      354075 non-null  int64\n",
      " 38  Brand_mini               354075 non-null  int64\n",
      " 39  Brand_mitsubishi         354075 non-null  int64\n",
      " 40  Brand_nissan             354075 non-null  int64\n",
      " 41  Brand_opel               354075 non-null  int64\n",
      " 42  Brand_peugeot            354075 non-null  int64\n",
      " 43  Brand_porsche            354075 non-null  int64\n",
      " 44  Brand_renault            354075 non-null  int64\n",
      " 45  Brand_rover              354075 non-null  int64\n",
      " 46  Brand_saab               354075 non-null  int64\n",
      " 47  Brand_seat               354075 non-null  int64\n",
      " 48  Brand_skoda              354075 non-null  int64\n",
      " 49  Brand_smart              354075 non-null  int64\n",
      " 50  Brand_sonstige_autos     354075 non-null  int64\n",
      " 51  Brand_subaru             354075 non-null  int64\n",
      " 52  Brand_suzuki             354075 non-null  int64\n",
      " 53  Brand_toyota             354075 non-null  int64\n",
      " 54  Brand_trabant            354075 non-null  int64\n",
      " 55  Brand_volkswagen         354075 non-null  int64\n",
      " 56  Brand_volvo              354075 non-null  int64\n",
      "dtypes: int64(57)\n",
      "memory usage: 154.0 MB\n"
     ]
    }
   ],
   "source": [
    "X.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Train val split\n",
    "X_train, X_val, y_train, y_val = train_test_split(X, y, test_size=0.25, random_state=12)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((265556, 57), (88519, 57))"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.shape, X_val.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Linear Regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LinearRegression()"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Linear Regression\n",
    "lr = LinearRegression()\n",
    "lr.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "lr_pred = lr.predict(X_val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Root Mean Squared Error (RMSE): 3419.9005802904935\n"
     ]
    }
   ],
   "source": [
    "# Calculate RMSE\n",
    "rmse = np.sqrt(mse(y_val, lr_pred))\n",
    "\n",
    "print(f'Root Mean Squared Error (RMSE): {rmse}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(88519,)"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_val.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Random Forrest Regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Random Forest Regression\n",
    "rf_model = RandomForestRegressor(n_estimators=20, random_state=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestRegressor(n_estimators=20, random_state=0)"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Train the model\n",
    "rf_model.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Make predictions on the test set\n",
    "rf_pred = rf_model.predict(X_val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Root Mean Squared Error (RMSE): 1825.118347557872\n"
     ]
    }
   ],
   "source": [
    "# Calculate RMSE\n",
    "rmse = np.sqrt(mse(y_val, rf_pred))\n",
    "\n",
    "print(f'Root Mean Squared Error (RMSE): {rmse}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Xgb Regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "xgb = XGBRegressor(n_estimators=20).fit(X_train, y_train)\n",
    "xgb_pred = xgb.predict(X_val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Root Mean Squared Error (RMSE): 1960.59625879506\n"
     ]
    }
   ],
   "source": [
    "# Calculate RMSE\n",
    "rmse = np.sqrt(mse(y_val, xgb_pred))\n",
    "print(f'Root Mean Squared Error (RMSE): {rmse}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Light GBM Regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "gbr = GradientBoostingRegressor(n_estimators=20).fit(X_train, y_train)\n",
    "gbr_pred = gbr.predict(X_val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Root Mean Squared Error (RMSE): 2571.7029267247844\n"
     ]
    }
   ],
   "source": [
    "# Calculate RMSE\n",
    "rmse = np.sqrt(mse(y_val, gbr_pred))\n",
    "print(f'Root Mean Squared Error (RMSE): {rmse}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Standarisasi Fitur\n",
    "scaler = StandardScaler()\n",
    "scaler.fit(X_train)\n",
    "X_train = scaler.transform(X_train)\n",
    "X_val = scaler.transform(X_val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create a LightGBM dataset for training with features X_train and labels Y_train \n",
    "train_data = lgb.Dataset(X_train, label=y_train) \n",
    "  \n",
    "# Create a LightGBM dataset for testing with features X_val and labels Y_val, \n",
    "# and specify the reference dataset as train_data for consistent evaluation \n",
    "test_data = lgb.Dataset(X_val, label=y_val, reference=train_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Define a dictionary of parameters for configuring the LightGBM regression model. \n",
    "params = { \n",
    "    'objective': 'regression', \n",
    "    'metric': 'rmse', \n",
    "    'boosting_type': 'gbdt', \n",
    "    'num_leaves': 31, \n",
    "    'learning_rate': 0.05, \n",
    "    'feature_fraction': 0.9, \n",
    "} "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/conda/lib/python3.9/site-packages/lightgbm/engine.py:181: UserWarning: 'early_stopping_rounds' argument is deprecated and will be removed in a future release of LightGBM. Pass 'early_stopping()' callback via 'callbacks' argument instead.\n",
      "  _log_warning(\"'early_stopping_rounds' argument is deprecated and will be removed in a future release of LightGBM. \"\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.196818 seconds.\n",
      "You can set `force_row_wise=true` to remove the overhead.\n",
      "And if memory is not enough, you can set `force_col_wise=true`.\n",
      "[LightGBM] [Info] Total Bins 528\n",
      "[LightGBM] [Info] Number of data points in the train set: 265556, number of used features: 57\n",
      "[LightGBM] [Info] Start training from score 4411.099745\n",
      "[1]\tvalid_0's rmse: 4372.97\n",
      "Training until validation scores don't improve for 10 rounds\n",
      "[2]\tvalid_0's rmse: 4226.71\n",
      "[3]\tvalid_0's rmse: 4102\n",
      "[4]\tvalid_0's rmse: 3972.93\n",
      "[5]\tvalid_0's rmse: 3852.38\n",
      "[6]\tvalid_0's rmse: 3755.94\n",
      "[7]\tvalid_0's rmse: 3648\n",
      "[8]\tvalid_0's rmse: 3548.18\n",
      "[9]\tvalid_0's rmse: 3453.82\n",
      "[10]\tvalid_0's rmse: 3367.53\n",
      "[11]\tvalid_0's rmse: 3297.05\n",
      "[12]\tvalid_0's rmse: 3219.38\n",
      "[13]\tvalid_0's rmse: 3147.36\n",
      "[14]\tvalid_0's rmse: 3080.34\n",
      "[15]\tvalid_0's rmse: 3017.23\n",
      "[16]\tvalid_0's rmse: 2959.37\n",
      "[17]\tvalid_0's rmse: 2904.5\n",
      "[18]\tvalid_0's rmse: 2853.35\n",
      "[19]\tvalid_0's rmse: 2804.85\n",
      "[20]\tvalid_0's rmse: 2761.18\n",
      "[21]\tvalid_0's rmse: 2720.56\n",
      "[22]\tvalid_0's rmse: 2680.21\n",
      "[23]\tvalid_0's rmse: 2641.84\n",
      "[24]\tvalid_0's rmse: 2606.39\n",
      "[25]\tvalid_0's rmse: 2580.66\n",
      "[26]\tvalid_0's rmse: 2549.18\n",
      "[27]\tvalid_0's rmse: 2518.1\n",
      "[28]\tvalid_0's rmse: 2490.39\n",
      "[29]\tvalid_0's rmse: 2463.55\n",
      "[30]\tvalid_0's rmse: 2439.91\n",
      "[31]\tvalid_0's rmse: 2417.63\n",
      "[32]\tvalid_0's rmse: 2396.83\n",
      "[33]\tvalid_0's rmse: 2376.87\n",
      "[34]\tvalid_0's rmse: 2358.22\n",
      "[35]\tvalid_0's rmse: 2339.41\n",
      "[36]\tvalid_0's rmse: 2322.2\n",
      "[37]\tvalid_0's rmse: 2305.57\n",
      "[38]\tvalid_0's rmse: 2291.29\n",
      "[39]\tvalid_0's rmse: 2276.6\n",
      "[40]\tvalid_0's rmse: 2263.44\n",
      "[41]\tvalid_0's rmse: 2249.61\n",
      "[42]\tvalid_0's rmse: 2236.69\n",
      "[43]\tvalid_0's rmse: 2225.39\n",
      "[44]\tvalid_0's rmse: 2214.13\n",
      "[45]\tvalid_0's rmse: 2203.88\n",
      "[46]\tvalid_0's rmse: 2192.79\n",
      "[47]\tvalid_0's rmse: 2183.54\n",
      "[48]\tvalid_0's rmse: 2175.2\n",
      "[49]\tvalid_0's rmse: 2165.6\n",
      "[50]\tvalid_0's rmse: 2155.47\n",
      "[51]\tvalid_0's rmse: 2146.61\n",
      "[52]\tvalid_0's rmse: 2139.45\n",
      "[53]\tvalid_0's rmse: 2131.3\n",
      "[54]\tvalid_0's rmse: 2124.55\n",
      "[55]\tvalid_0's rmse: 2117.51\n",
      "[56]\tvalid_0's rmse: 2111.03\n",
      "[57]\tvalid_0's rmse: 2104.5\n",
      "[58]\tvalid_0's rmse: 2098.65\n",
      "[59]\tvalid_0's rmse: 2092.18\n",
      "[60]\tvalid_0's rmse: 2087.08\n",
      "[61]\tvalid_0's rmse: 2081.57\n",
      "[62]\tvalid_0's rmse: 2076.38\n",
      "[63]\tvalid_0's rmse: 2071.13\n",
      "[64]\tvalid_0's rmse: 2066.42\n",
      "[65]\tvalid_0's rmse: 2062.11\n",
      "[66]\tvalid_0's rmse: 2057.47\n",
      "[67]\tvalid_0's rmse: 2052.31\n",
      "[68]\tvalid_0's rmse: 2048\n",
      "[69]\tvalid_0's rmse: 2044.4\n",
      "[70]\tvalid_0's rmse: 2040.26\n",
      "[71]\tvalid_0's rmse: 2036.94\n",
      "[72]\tvalid_0's rmse: 2032.88\n",
      "[73]\tvalid_0's rmse: 2029.43\n",
      "[74]\tvalid_0's rmse: 2025.96\n",
      "[75]\tvalid_0's rmse: 2022.76\n",
      "[76]\tvalid_0's rmse: 2019.46\n",
      "[77]\tvalid_0's rmse: 2016.87\n",
      "[78]\tvalid_0's rmse: 2014.32\n",
      "[79]\tvalid_0's rmse: 2011.83\n",
      "[80]\tvalid_0's rmse: 2008.98\n",
      "[81]\tvalid_0's rmse: 2006.42\n",
      "[82]\tvalid_0's rmse: 2003.75\n",
      "[83]\tvalid_0's rmse: 2001\n",
      "[84]\tvalid_0's rmse: 1998.88\n",
      "[85]\tvalid_0's rmse: 1996.32\n",
      "[86]\tvalid_0's rmse: 1993.94\n",
      "[87]\tvalid_0's rmse: 1992.09\n",
      "[88]\tvalid_0's rmse: 1989.88\n",
      "[89]\tvalid_0's rmse: 1988.04\n",
      "[90]\tvalid_0's rmse: 1985.91\n",
      "[91]\tvalid_0's rmse: 1983.6\n",
      "[92]\tvalid_0's rmse: 1981.75\n",
      "[93]\tvalid_0's rmse: 1979.85\n",
      "[94]\tvalid_0's rmse: 1977.8\n",
      "[95]\tvalid_0's rmse: 1975.97\n",
      "[96]\tvalid_0's rmse: 1974.19\n",
      "[97]\tvalid_0's rmse: 1972.14\n",
      "[98]\tvalid_0's rmse: 1970.82\n",
      "[99]\tvalid_0's rmse: 1969.29\n",
      "[100]\tvalid_0's rmse: 1967.46\n",
      "[101]\tvalid_0's rmse: 1966.17\n",
      "[102]\tvalid_0's rmse: 1964.53\n",
      "[103]\tvalid_0's rmse: 1963.21\n",
      "[104]\tvalid_0's rmse: 1961.73\n",
      "[105]\tvalid_0's rmse: 1960.24\n",
      "[106]\tvalid_0's rmse: 1958.8\n",
      "[107]\tvalid_0's rmse: 1957.14\n",
      "[108]\tvalid_0's rmse: 1955.78\n",
      "[109]\tvalid_0's rmse: 1954.63\n",
      "[110]\tvalid_0's rmse: 1953.46\n",
      "[111]\tvalid_0's rmse: 1952.4\n",
      "[112]\tvalid_0's rmse: 1951.24\n",
      "[113]\tvalid_0's rmse: 1949.83\n",
      "[114]\tvalid_0's rmse: 1948.58\n",
      "[115]\tvalid_0's rmse: 1947.3\n",
      "[116]\tvalid_0's rmse: 1946.22\n",
      "[117]\tvalid_0's rmse: 1945.2\n",
      "[118]\tvalid_0's rmse: 1943.75\n",
      "[119]\tvalid_0's rmse: 1942.73\n",
      "[120]\tvalid_0's rmse: 1941.67\n",
      "[121]\tvalid_0's rmse: 1940.26\n",
      "[122]\tvalid_0's rmse: 1939.29\n",
      "[123]\tvalid_0's rmse: 1938.35\n",
      "[124]\tvalid_0's rmse: 1937.4\n",
      "[125]\tvalid_0's rmse: 1936.41\n",
      "[126]\tvalid_0's rmse: 1935.13\n",
      "[127]\tvalid_0's rmse: 1934.25\n",
      "[128]\tvalid_0's rmse: 1933.34\n",
      "[129]\tvalid_0's rmse: 1932.14\n",
      "[130]\tvalid_0's rmse: 1931.24\n",
      "[131]\tvalid_0's rmse: 1930.4\n",
      "[132]\tvalid_0's rmse: 1929.21\n",
      "[133]\tvalid_0's rmse: 1928.18\n",
      "[134]\tvalid_0's rmse: 1927.47\n",
      "[135]\tvalid_0's rmse: 1926.59\n",
      "[136]\tvalid_0's rmse: 1925.92\n",
      "[137]\tvalid_0's rmse: 1925.43\n",
      "[138]\tvalid_0's rmse: 1924.74\n",
      "[139]\tvalid_0's rmse: 1923.86\n",
      "[140]\tvalid_0's rmse: 1923.16\n",
      "[141]\tvalid_0's rmse: 1922.43\n",
      "[142]\tvalid_0's rmse: 1921.68\n",
      "[143]\tvalid_0's rmse: 1920.98\n",
      "[144]\tvalid_0's rmse: 1920.32\n",
      "[145]\tvalid_0's rmse: 1919.46\n",
      "[146]\tvalid_0's rmse: 1918.64\n",
      "[147]\tvalid_0's rmse: 1917.96\n",
      "[148]\tvalid_0's rmse: 1917.3\n",
      "[149]\tvalid_0's rmse: 1916.84\n",
      "[150]\tvalid_0's rmse: 1916.16\n",
      "[151]\tvalid_0's rmse: 1915.58\n",
      "[152]\tvalid_0's rmse: 1914.79\n",
      "[153]\tvalid_0's rmse: 1914.37\n",
      "[154]\tvalid_0's rmse: 1913.64\n",
      "[155]\tvalid_0's rmse: 1913.13\n",
      "[156]\tvalid_0's rmse: 1912.65\n",
      "[157]\tvalid_0's rmse: 1911.9\n",
      "[158]\tvalid_0's rmse: 1911.4\n",
      "[159]\tvalid_0's rmse: 1910.91\n",
      "[160]\tvalid_0's rmse: 1910.58\n",
      "[161]\tvalid_0's rmse: 1910.12\n",
      "[162]\tvalid_0's rmse: 1909.81\n",
      "[163]\tvalid_0's rmse: 1909.29\n",
      "[164]\tvalid_0's rmse: 1908.77\n",
      "[165]\tvalid_0's rmse: 1908.27\n",
      "[166]\tvalid_0's rmse: 1907.88\n",
      "[167]\tvalid_0's rmse: 1907.47\n",
      "[168]\tvalid_0's rmse: 1906.84\n",
      "[169]\tvalid_0's rmse: 1906.29\n",
      "[170]\tvalid_0's rmse: 1905.63\n",
      "[171]\tvalid_0's rmse: 1905\n",
      "[172]\tvalid_0's rmse: 1904.19\n",
      "[173]\tvalid_0's rmse: 1903.38\n",
      "[174]\tvalid_0's rmse: 1902.88\n",
      "[175]\tvalid_0's rmse: 1902.56\n",
      "[176]\tvalid_0's rmse: 1902\n",
      "[177]\tvalid_0's rmse: 1901.44\n",
      "[178]\tvalid_0's rmse: 1901.14\n",
      "[179]\tvalid_0's rmse: 1900.8\n",
      "[180]\tvalid_0's rmse: 1900.49\n",
      "[181]\tvalid_0's rmse: 1900.22\n",
      "[182]\tvalid_0's rmse: 1899.95\n",
      "[183]\tvalid_0's rmse: 1899.48\n",
      "[184]\tvalid_0's rmse: 1898.98\n",
      "[185]\tvalid_0's rmse: 1898.5\n",
      "[186]\tvalid_0's rmse: 1898.22\n",
      "[187]\tvalid_0's rmse: 1897.73\n",
      "[188]\tvalid_0's rmse: 1897.29\n",
      "[189]\tvalid_0's rmse: 1896.77\n",
      "[190]\tvalid_0's rmse: 1896.48\n",
      "[191]\tvalid_0's rmse: 1896.25\n",
      "[192]\tvalid_0's rmse: 1895.65\n",
      "[193]\tvalid_0's rmse: 1895.22\n",
      "[194]\tvalid_0's rmse: 1895.04\n",
      "[195]\tvalid_0's rmse: 1894.79\n",
      "[196]\tvalid_0's rmse: 1894.5\n",
      "[197]\tvalid_0's rmse: 1894.02\n",
      "[198]\tvalid_0's rmse: 1893.76\n",
      "[199]\tvalid_0's rmse: 1893.5\n",
      "[200]\tvalid_0's rmse: 1893.11\n",
      "[201]\tvalid_0's rmse: 1892.68\n",
      "[202]\tvalid_0's rmse: 1892.21\n",
      "[203]\tvalid_0's rmse: 1891.84\n",
      "[204]\tvalid_0's rmse: 1891.71\n",
      "[205]\tvalid_0's rmse: 1891.34\n",
      "[206]\tvalid_0's rmse: 1890.91\n",
      "[207]\tvalid_0's rmse: 1890.8\n",
      "[208]\tvalid_0's rmse: 1890.56\n",
      "[209]\tvalid_0's rmse: 1890.25\n",
      "[210]\tvalid_0's rmse: 1889.6\n",
      "[211]\tvalid_0's rmse: 1889.22\n",
      "[212]\tvalid_0's rmse: 1888.9\n",
      "[213]\tvalid_0's rmse: 1888.71\n",
      "[214]\tvalid_0's rmse: 1888.38\n",
      "[215]\tvalid_0's rmse: 1888.15\n",
      "[216]\tvalid_0's rmse: 1887.78\n",
      "[217]\tvalid_0's rmse: 1887.54\n",
      "[218]\tvalid_0's rmse: 1887.32\n",
      "[219]\tvalid_0's rmse: 1887\n",
      "[220]\tvalid_0's rmse: 1886.64\n",
      "[221]\tvalid_0's rmse: 1886.48\n",
      "[222]\tvalid_0's rmse: 1886.14\n",
      "[223]\tvalid_0's rmse: 1885.88\n",
      "[224]\tvalid_0's rmse: 1885.47\n",
      "[225]\tvalid_0's rmse: 1885.13\n",
      "[226]\tvalid_0's rmse: 1884.81\n",
      "[227]\tvalid_0's rmse: 1884.72\n",
      "[228]\tvalid_0's rmse: 1884.54\n",
      "[229]\tvalid_0's rmse: 1884.37\n",
      "[230]\tvalid_0's rmse: 1884.1\n",
      "[231]\tvalid_0's rmse: 1883.89\n",
      "[232]\tvalid_0's rmse: 1883.67\n",
      "[233]\tvalid_0's rmse: 1883.25\n",
      "[234]\tvalid_0's rmse: 1883.04\n",
      "[235]\tvalid_0's rmse: 1882.91\n",
      "[236]\tvalid_0's rmse: 1882.71\n",
      "[237]\tvalid_0's rmse: 1882.48\n",
      "[238]\tvalid_0's rmse: 1882.16\n",
      "[239]\tvalid_0's rmse: 1881.91\n",
      "[240]\tvalid_0's rmse: 1881.59\n",
      "[241]\tvalid_0's rmse: 1881.43\n",
      "[242]\tvalid_0's rmse: 1881.21\n",
      "[243]\tvalid_0's rmse: 1881\n",
      "[244]\tvalid_0's rmse: 1880.81\n",
      "[245]\tvalid_0's rmse: 1880.53\n",
      "[246]\tvalid_0's rmse: 1880.25\n",
      "[247]\tvalid_0's rmse: 1879.86\n",
      "[248]\tvalid_0's rmse: 1879.59\n",
      "[249]\tvalid_0's rmse: 1879.18\n",
      "[250]\tvalid_0's rmse: 1878.99\n",
      "[251]\tvalid_0's rmse: 1878.68\n",
      "[252]\tvalid_0's rmse: 1878.44\n",
      "[253]\tvalid_0's rmse: 1878.04\n",
      "[254]\tvalid_0's rmse: 1877.82\n",
      "[255]\tvalid_0's rmse: 1877.64\n",
      "[256]\tvalid_0's rmse: 1877.38\n",
      "[257]\tvalid_0's rmse: 1877.12\n",
      "[258]\tvalid_0's rmse: 1876.77\n",
      "[259]\tvalid_0's rmse: 1876.46\n",
      "[260]\tvalid_0's rmse: 1876.33\n",
      "[261]\tvalid_0's rmse: 1876.08\n",
      "[262]\tvalid_0's rmse: 1875.81\n",
      "[263]\tvalid_0's rmse: 1875.61\n",
      "[264]\tvalid_0's rmse: 1875.34\n",
      "[265]\tvalid_0's rmse: 1874.99\n",
      "[266]\tvalid_0's rmse: 1874.61\n",
      "[267]\tvalid_0's rmse: 1874.48\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[268]\tvalid_0's rmse: 1874.3\n",
      "[269]\tvalid_0's rmse: 1874.17\n",
      "[270]\tvalid_0's rmse: 1873.97\n",
      "[271]\tvalid_0's rmse: 1873.76\n",
      "[272]\tvalid_0's rmse: 1873.57\n",
      "[273]\tvalid_0's rmse: 1873.04\n",
      "[274]\tvalid_0's rmse: 1872.88\n",
      "[275]\tvalid_0's rmse: 1872.52\n",
      "[276]\tvalid_0's rmse: 1872.37\n",
      "[277]\tvalid_0's rmse: 1872.17\n",
      "[278]\tvalid_0's rmse: 1872.04\n",
      "[279]\tvalid_0's rmse: 1871.85\n",
      "[280]\tvalid_0's rmse: 1871.54\n",
      "[281]\tvalid_0's rmse: 1871.47\n",
      "[282]\tvalid_0's rmse: 1871.36\n",
      "[283]\tvalid_0's rmse: 1871.12\n",
      "[284]\tvalid_0's rmse: 1870.9\n",
      "[285]\tvalid_0's rmse: 1870.77\n",
      "[286]\tvalid_0's rmse: 1870.37\n",
      "[287]\tvalid_0's rmse: 1870.29\n",
      "[288]\tvalid_0's rmse: 1870.11\n",
      "[289]\tvalid_0's rmse: 1869.94\n",
      "[290]\tvalid_0's rmse: 1869.73\n",
      "[291]\tvalid_0's rmse: 1869.62\n",
      "[292]\tvalid_0's rmse: 1869.4\n",
      "[293]\tvalid_0's rmse: 1869.29\n",
      "[294]\tvalid_0's rmse: 1869.11\n",
      "[295]\tvalid_0's rmse: 1868.88\n",
      "[296]\tvalid_0's rmse: 1868.73\n",
      "[297]\tvalid_0's rmse: 1868.56\n",
      "[298]\tvalid_0's rmse: 1868.37\n",
      "[299]\tvalid_0's rmse: 1868.21\n",
      "[300]\tvalid_0's rmse: 1868.04\n",
      "[301]\tvalid_0's rmse: 1867.92\n",
      "[302]\tvalid_0's rmse: 1867.82\n",
      "[303]\tvalid_0's rmse: 1867.64\n",
      "[304]\tvalid_0's rmse: 1867.33\n",
      "[305]\tvalid_0's rmse: 1867.17\n",
      "[306]\tvalid_0's rmse: 1866.58\n",
      "[307]\tvalid_0's rmse: 1866.48\n",
      "[308]\tvalid_0's rmse: 1866.32\n",
      "[309]\tvalid_0's rmse: 1865.73\n",
      "[310]\tvalid_0's rmse: 1865.54\n",
      "[311]\tvalid_0's rmse: 1865.38\n",
      "[312]\tvalid_0's rmse: 1865.31\n",
      "[313]\tvalid_0's rmse: 1865.19\n",
      "[314]\tvalid_0's rmse: 1865.06\n",
      "[315]\tvalid_0's rmse: 1864.94\n",
      "[316]\tvalid_0's rmse: 1864.69\n",
      "[317]\tvalid_0's rmse: 1864.52\n",
      "[318]\tvalid_0's rmse: 1864.45\n",
      "[319]\tvalid_0's rmse: 1864.2\n",
      "[320]\tvalid_0's rmse: 1864.15\n",
      "[321]\tvalid_0's rmse: 1863.86\n",
      "[322]\tvalid_0's rmse: 1863.77\n",
      "[323]\tvalid_0's rmse: 1863.68\n",
      "[324]\tvalid_0's rmse: 1863.58\n",
      "[325]\tvalid_0's rmse: 1863.45\n",
      "[326]\tvalid_0's rmse: 1863.35\n",
      "[327]\tvalid_0's rmse: 1863.2\n",
      "[328]\tvalid_0's rmse: 1863.05\n",
      "[329]\tvalid_0's rmse: 1862.76\n",
      "[330]\tvalid_0's rmse: 1862.64\n",
      "[331]\tvalid_0's rmse: 1862.48\n",
      "[332]\tvalid_0's rmse: 1862.36\n",
      "[333]\tvalid_0's rmse: 1862.21\n",
      "[334]\tvalid_0's rmse: 1861.97\n",
      "[335]\tvalid_0's rmse: 1861.83\n",
      "[336]\tvalid_0's rmse: 1861.76\n",
      "[337]\tvalid_0's rmse: 1861.55\n",
      "[338]\tvalid_0's rmse: 1861.49\n",
      "[339]\tvalid_0's rmse: 1861.31\n",
      "[340]\tvalid_0's rmse: 1861.23\n",
      "[341]\tvalid_0's rmse: 1861.12\n",
      "[342]\tvalid_0's rmse: 1860.75\n",
      "[343]\tvalid_0's rmse: 1860.41\n",
      "[344]\tvalid_0's rmse: 1860.31\n",
      "[345]\tvalid_0's rmse: 1860.28\n",
      "[346]\tvalid_0's rmse: 1860.06\n",
      "[347]\tvalid_0's rmse: 1859.76\n",
      "[348]\tvalid_0's rmse: 1859.51\n",
      "[349]\tvalid_0's rmse: 1859.24\n",
      "[350]\tvalid_0's rmse: 1859.06\n",
      "[351]\tvalid_0's rmse: 1858.93\n",
      "[352]\tvalid_0's rmse: 1858.77\n",
      "[353]\tvalid_0's rmse: 1858.58\n",
      "[354]\tvalid_0's rmse: 1858.34\n",
      "[355]\tvalid_0's rmse: 1858.27\n",
      "[356]\tvalid_0's rmse: 1858.19\n",
      "[357]\tvalid_0's rmse: 1858.13\n",
      "[358]\tvalid_0's rmse: 1858.11\n",
      "[359]\tvalid_0's rmse: 1857.93\n",
      "[360]\tvalid_0's rmse: 1857.72\n",
      "[361]\tvalid_0's rmse: 1857.49\n",
      "[362]\tvalid_0's rmse: 1857.39\n",
      "[363]\tvalid_0's rmse: 1857.2\n",
      "[364]\tvalid_0's rmse: 1857.05\n",
      "[365]\tvalid_0's rmse: 1856.8\n",
      "[366]\tvalid_0's rmse: 1856.73\n",
      "[367]\tvalid_0's rmse: 1856.54\n",
      "[368]\tvalid_0's rmse: 1856.47\n",
      "[369]\tvalid_0's rmse: 1856.28\n",
      "[370]\tvalid_0's rmse: 1855.94\n",
      "[371]\tvalid_0's rmse: 1855.83\n",
      "[372]\tvalid_0's rmse: 1855.75\n",
      "[373]\tvalid_0's rmse: 1855.47\n",
      "[374]\tvalid_0's rmse: 1855.13\n",
      "[375]\tvalid_0's rmse: 1854.91\n",
      "[376]\tvalid_0's rmse: 1854.87\n",
      "[377]\tvalid_0's rmse: 1854.78\n",
      "[378]\tvalid_0's rmse: 1854.7\n",
      "[379]\tvalid_0's rmse: 1854.42\n",
      "[380]\tvalid_0's rmse: 1854.34\n",
      "[381]\tvalid_0's rmse: 1854.09\n",
      "[382]\tvalid_0's rmse: 1853.96\n",
      "[383]\tvalid_0's rmse: 1853.9\n",
      "[384]\tvalid_0's rmse: 1853.76\n",
      "[385]\tvalid_0's rmse: 1853.55\n",
      "[386]\tvalid_0's rmse: 1853.55\n",
      "[387]\tvalid_0's rmse: 1853.44\n",
      "[388]\tvalid_0's rmse: 1853.2\n",
      "[389]\tvalid_0's rmse: 1852.99\n",
      "[390]\tvalid_0's rmse: 1852.67\n",
      "[391]\tvalid_0's rmse: 1852.59\n",
      "[392]\tvalid_0's rmse: 1852.29\n",
      "[393]\tvalid_0's rmse: 1852.14\n",
      "[394]\tvalid_0's rmse: 1852.08\n",
      "[395]\tvalid_0's rmse: 1851.92\n",
      "[396]\tvalid_0's rmse: 1851.85\n",
      "[397]\tvalid_0's rmse: 1851.76\n",
      "[398]\tvalid_0's rmse: 1851.67\n",
      "[399]\tvalid_0's rmse: 1851.46\n",
      "[400]\tvalid_0's rmse: 1851.16\n",
      "[401]\tvalid_0's rmse: 1851\n",
      "[402]\tvalid_0's rmse: 1850.86\n",
      "[403]\tvalid_0's rmse: 1850.77\n",
      "[404]\tvalid_0's rmse: 1850.52\n",
      "[405]\tvalid_0's rmse: 1850.5\n",
      "[406]\tvalid_0's rmse: 1850.31\n",
      "[407]\tvalid_0's rmse: 1850.18\n",
      "[408]\tvalid_0's rmse: 1849.89\n",
      "[409]\tvalid_0's rmse: 1849.84\n",
      "[410]\tvalid_0's rmse: 1849.72\n",
      "[411]\tvalid_0's rmse: 1849.6\n",
      "[412]\tvalid_0's rmse: 1849.53\n",
      "[413]\tvalid_0's rmse: 1849.37\n",
      "[414]\tvalid_0's rmse: 1849.13\n",
      "[415]\tvalid_0's rmse: 1848.93\n",
      "[416]\tvalid_0's rmse: 1848.88\n",
      "[417]\tvalid_0's rmse: 1848.85\n",
      "[418]\tvalid_0's rmse: 1848.8\n",
      "[419]\tvalid_0's rmse: 1848.58\n",
      "[420]\tvalid_0's rmse: 1848.52\n",
      "[421]\tvalid_0's rmse: 1848.25\n",
      "[422]\tvalid_0's rmse: 1848.16\n",
      "[423]\tvalid_0's rmse: 1847.96\n",
      "[424]\tvalid_0's rmse: 1847.71\n",
      "[425]\tvalid_0's rmse: 1847.48\n",
      "[426]\tvalid_0's rmse: 1847.4\n",
      "[427]\tvalid_0's rmse: 1847.22\n",
      "[428]\tvalid_0's rmse: 1847.16\n",
      "[429]\tvalid_0's rmse: 1846.98\n",
      "[430]\tvalid_0's rmse: 1846.91\n",
      "[431]\tvalid_0's rmse: 1846.79\n",
      "[432]\tvalid_0's rmse: 1846.81\n",
      "[433]\tvalid_0's rmse: 1846.65\n",
      "[434]\tvalid_0's rmse: 1846.62\n",
      "[435]\tvalid_0's rmse: 1846.58\n",
      "[436]\tvalid_0's rmse: 1846.41\n",
      "[437]\tvalid_0's rmse: 1846.36\n",
      "[438]\tvalid_0's rmse: 1846.26\n",
      "[439]\tvalid_0's rmse: 1846.01\n",
      "[440]\tvalid_0's rmse: 1845.9\n",
      "[441]\tvalid_0's rmse: 1845.73\n",
      "[442]\tvalid_0's rmse: 1845.52\n",
      "[443]\tvalid_0's rmse: 1845.29\n",
      "[444]\tvalid_0's rmse: 1845.21\n",
      "[445]\tvalid_0's rmse: 1845.14\n",
      "[446]\tvalid_0's rmse: 1844.99\n",
      "[447]\tvalid_0's rmse: 1844.89\n",
      "[448]\tvalid_0's rmse: 1844.76\n",
      "[449]\tvalid_0's rmse: 1844.71\n",
      "[450]\tvalid_0's rmse: 1844.68\n",
      "[451]\tvalid_0's rmse: 1844.43\n",
      "[452]\tvalid_0's rmse: 1844.28\n",
      "[453]\tvalid_0's rmse: 1844.21\n",
      "[454]\tvalid_0's rmse: 1844.09\n",
      "[455]\tvalid_0's rmse: 1844.03\n",
      "[456]\tvalid_0's rmse: 1843.85\n",
      "[457]\tvalid_0's rmse: 1843.77\n",
      "[458]\tvalid_0's rmse: 1843.72\n",
      "[459]\tvalid_0's rmse: 1843.68\n",
      "[460]\tvalid_0's rmse: 1843.54\n",
      "[461]\tvalid_0's rmse: 1843.49\n",
      "[462]\tvalid_0's rmse: 1843.44\n",
      "[463]\tvalid_0's rmse: 1843.39\n",
      "[464]\tvalid_0's rmse: 1843.28\n",
      "[465]\tvalid_0's rmse: 1843.16\n",
      "[466]\tvalid_0's rmse: 1843\n",
      "[467]\tvalid_0's rmse: 1842.84\n",
      "[468]\tvalid_0's rmse: 1842.64\n",
      "[469]\tvalid_0's rmse: 1842.43\n",
      "[470]\tvalid_0's rmse: 1842.35\n",
      "[471]\tvalid_0's rmse: 1842.07\n",
      "[472]\tvalid_0's rmse: 1841.92\n",
      "[473]\tvalid_0's rmse: 1841.74\n",
      "[474]\tvalid_0's rmse: 1841.56\n",
      "[475]\tvalid_0's rmse: 1841.2\n",
      "[476]\tvalid_0's rmse: 1841.11\n",
      "[477]\tvalid_0's rmse: 1840.96\n",
      "[478]\tvalid_0's rmse: 1840.85\n",
      "[479]\tvalid_0's rmse: 1840.64\n",
      "[480]\tvalid_0's rmse: 1840.55\n",
      "[481]\tvalid_0's rmse: 1840.53\n",
      "[482]\tvalid_0's rmse: 1840.5\n",
      "[483]\tvalid_0's rmse: 1840.44\n",
      "[484]\tvalid_0's rmse: 1840.31\n",
      "[485]\tvalid_0's rmse: 1840.02\n",
      "[486]\tvalid_0's rmse: 1840.03\n",
      "[487]\tvalid_0's rmse: 1839.98\n",
      "[488]\tvalid_0's rmse: 1839.92\n",
      "[489]\tvalid_0's rmse: 1839.89\n",
      "[490]\tvalid_0's rmse: 1839.72\n",
      "[491]\tvalid_0's rmse: 1839.55\n",
      "[492]\tvalid_0's rmse: 1839.33\n",
      "[493]\tvalid_0's rmse: 1839.19\n",
      "[494]\tvalid_0's rmse: 1839.06\n",
      "[495]\tvalid_0's rmse: 1839.02\n",
      "[496]\tvalid_0's rmse: 1838.89\n",
      "[497]\tvalid_0's rmse: 1838.76\n",
      "[498]\tvalid_0's rmse: 1838.63\n",
      "[499]\tvalid_0's rmse: 1838.39\n",
      "[500]\tvalid_0's rmse: 1838.35\n",
      "[501]\tvalid_0's rmse: 1838.29\n",
      "[502]\tvalid_0's rmse: 1838.25\n",
      "[503]\tvalid_0's rmse: 1838.08\n",
      "[504]\tvalid_0's rmse: 1838.01\n",
      "[505]\tvalid_0's rmse: 1837.87\n",
      "[506]\tvalid_0's rmse: 1837.82\n",
      "[507]\tvalid_0's rmse: 1837.73\n",
      "[508]\tvalid_0's rmse: 1837.66\n",
      "[509]\tvalid_0's rmse: 1837.6\n",
      "[510]\tvalid_0's rmse: 1837.5\n",
      "[511]\tvalid_0's rmse: 1837.43\n",
      "[512]\tvalid_0's rmse: 1837.25\n",
      "[513]\tvalid_0's rmse: 1837.22\n",
      "[514]\tvalid_0's rmse: 1837.18\n",
      "[515]\tvalid_0's rmse: 1837.14\n",
      "[516]\tvalid_0's rmse: 1837.13\n",
      "[517]\tvalid_0's rmse: 1837.04\n",
      "[518]\tvalid_0's rmse: 1836.99\n",
      "[519]\tvalid_0's rmse: 1836.82\n",
      "[520]\tvalid_0's rmse: 1836.72\n",
      "[521]\tvalid_0's rmse: 1836.45\n",
      "[522]\tvalid_0's rmse: 1836.34\n",
      "[523]\tvalid_0's rmse: 1836.29\n",
      "[524]\tvalid_0's rmse: 1836.12\n",
      "[525]\tvalid_0's rmse: 1836.01\n",
      "[526]\tvalid_0's rmse: 1835.81\n",
      "[527]\tvalid_0's rmse: 1835.73\n",
      "[528]\tvalid_0's rmse: 1835.59\n",
      "[529]\tvalid_0's rmse: 1835.56\n",
      "[530]\tvalid_0's rmse: 1835.44\n",
      "[531]\tvalid_0's rmse: 1835.38\n",
      "[532]\tvalid_0's rmse: 1835.32\n",
      "[533]\tvalid_0's rmse: 1835.06\n",
      "[534]\tvalid_0's rmse: 1834.84\n",
      "[535]\tvalid_0's rmse: 1834.69\n",
      "[536]\tvalid_0's rmse: 1834.58\n",
      "[537]\tvalid_0's rmse: 1834.54\n",
      "[538]\tvalid_0's rmse: 1834.5\n",
      "[539]\tvalid_0's rmse: 1834.36\n",
      "[540]\tvalid_0's rmse: 1834.28\n",
      "[541]\tvalid_0's rmse: 1834.2\n",
      "[542]\tvalid_0's rmse: 1834.14\n",
      "[543]\tvalid_0's rmse: 1834.02\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[544]\tvalid_0's rmse: 1833.95\n",
      "[545]\tvalid_0's rmse: 1833.86\n",
      "[546]\tvalid_0's rmse: 1833.76\n",
      "[547]\tvalid_0's rmse: 1833.68\n",
      "[548]\tvalid_0's rmse: 1833.67\n",
      "[549]\tvalid_0's rmse: 1833.63\n",
      "[550]\tvalid_0's rmse: 1833.54\n",
      "[551]\tvalid_0's rmse: 1833.43\n",
      "[552]\tvalid_0's rmse: 1833.43\n",
      "[553]\tvalid_0's rmse: 1833.22\n",
      "[554]\tvalid_0's rmse: 1833.09\n",
      "[555]\tvalid_0's rmse: 1832.98\n",
      "[556]\tvalid_0's rmse: 1832.94\n",
      "[557]\tvalid_0's rmse: 1832.88\n",
      "[558]\tvalid_0's rmse: 1832.8\n",
      "[559]\tvalid_0's rmse: 1832.67\n",
      "[560]\tvalid_0's rmse: 1832.59\n",
      "[561]\tvalid_0's rmse: 1832.46\n",
      "[562]\tvalid_0's rmse: 1832.3\n",
      "[563]\tvalid_0's rmse: 1832.26\n",
      "[564]\tvalid_0's rmse: 1832.14\n",
      "[565]\tvalid_0's rmse: 1831.97\n",
      "[566]\tvalid_0's rmse: 1831.79\n",
      "[567]\tvalid_0's rmse: 1831.75\n",
      "[568]\tvalid_0's rmse: 1831.7\n",
      "[569]\tvalid_0's rmse: 1831.71\n",
      "[570]\tvalid_0's rmse: 1831.61\n",
      "[571]\tvalid_0's rmse: 1831.46\n",
      "[572]\tvalid_0's rmse: 1831.42\n",
      "[573]\tvalid_0's rmse: 1831.41\n",
      "[574]\tvalid_0's rmse: 1831.34\n",
      "[575]\tvalid_0's rmse: 1831.26\n",
      "[576]\tvalid_0's rmse: 1831.15\n",
      "[577]\tvalid_0's rmse: 1831.06\n",
      "[578]\tvalid_0's rmse: 1830.84\n",
      "[579]\tvalid_0's rmse: 1830.7\n",
      "[580]\tvalid_0's rmse: 1830.49\n",
      "[581]\tvalid_0's rmse: 1830.27\n",
      "[582]\tvalid_0's rmse: 1830.28\n",
      "[583]\tvalid_0's rmse: 1830.22\n",
      "[584]\tvalid_0's rmse: 1830.13\n",
      "[585]\tvalid_0's rmse: 1830.09\n",
      "[586]\tvalid_0's rmse: 1829.97\n",
      "[587]\tvalid_0's rmse: 1829.92\n",
      "[588]\tvalid_0's rmse: 1829.91\n",
      "[589]\tvalid_0's rmse: 1829.86\n",
      "[590]\tvalid_0's rmse: 1829.82\n",
      "[591]\tvalid_0's rmse: 1829.77\n",
      "[592]\tvalid_0's rmse: 1829.67\n",
      "[593]\tvalid_0's rmse: 1829.56\n",
      "[594]\tvalid_0's rmse: 1829.43\n",
      "[595]\tvalid_0's rmse: 1829.25\n",
      "[596]\tvalid_0's rmse: 1829.2\n",
      "[597]\tvalid_0's rmse: 1829.2\n",
      "[598]\tvalid_0's rmse: 1829.18\n",
      "[599]\tvalid_0's rmse: 1829.15\n",
      "[600]\tvalid_0's rmse: 1829.06\n",
      "[601]\tvalid_0's rmse: 1828.77\n",
      "[602]\tvalid_0's rmse: 1828.56\n",
      "[603]\tvalid_0's rmse: 1828.49\n",
      "[604]\tvalid_0's rmse: 1828.45\n",
      "[605]\tvalid_0's rmse: 1828.4\n",
      "[606]\tvalid_0's rmse: 1828.35\n",
      "[607]\tvalid_0's rmse: 1828.27\n",
      "[608]\tvalid_0's rmse: 1828.13\n",
      "[609]\tvalid_0's rmse: 1828.04\n",
      "[610]\tvalid_0's rmse: 1828.02\n",
      "[611]\tvalid_0's rmse: 1827.85\n",
      "[612]\tvalid_0's rmse: 1827.85\n",
      "[613]\tvalid_0's rmse: 1827.71\n",
      "[614]\tvalid_0's rmse: 1827.65\n",
      "[615]\tvalid_0's rmse: 1827.62\n",
      "[616]\tvalid_0's rmse: 1827.48\n",
      "[617]\tvalid_0's rmse: 1827.35\n",
      "[618]\tvalid_0's rmse: 1827.29\n",
      "[619]\tvalid_0's rmse: 1827.14\n",
      "[620]\tvalid_0's rmse: 1827.09\n",
      "[621]\tvalid_0's rmse: 1826.94\n",
      "[622]\tvalid_0's rmse: 1826.74\n",
      "[623]\tvalid_0's rmse: 1826.66\n",
      "[624]\tvalid_0's rmse: 1826.59\n",
      "[625]\tvalid_0's rmse: 1826.47\n",
      "[626]\tvalid_0's rmse: 1826.41\n",
      "[627]\tvalid_0's rmse: 1826.32\n",
      "[628]\tvalid_0's rmse: 1826.15\n",
      "[629]\tvalid_0's rmse: 1826.11\n",
      "[630]\tvalid_0's rmse: 1826.09\n",
      "[631]\tvalid_0's rmse: 1826.05\n",
      "[632]\tvalid_0's rmse: 1825.95\n",
      "[633]\tvalid_0's rmse: 1825.9\n",
      "[634]\tvalid_0's rmse: 1825.82\n",
      "[635]\tvalid_0's rmse: 1825.58\n",
      "[636]\tvalid_0's rmse: 1825.4\n",
      "[637]\tvalid_0's rmse: 1825.36\n",
      "[638]\tvalid_0's rmse: 1825.3\n",
      "[639]\tvalid_0's rmse: 1825.29\n",
      "[640]\tvalid_0's rmse: 1825.25\n",
      "[641]\tvalid_0's rmse: 1825.14\n",
      "[642]\tvalid_0's rmse: 1824.9\n",
      "[643]\tvalid_0's rmse: 1824.66\n",
      "[644]\tvalid_0's rmse: 1824.48\n",
      "[645]\tvalid_0's rmse: 1824.38\n",
      "[646]\tvalid_0's rmse: 1824.35\n",
      "[647]\tvalid_0's rmse: 1824.3\n",
      "[648]\tvalid_0's rmse: 1824.27\n",
      "[649]\tvalid_0's rmse: 1824.19\n",
      "[650]\tvalid_0's rmse: 1824.15\n",
      "[651]\tvalid_0's rmse: 1824.08\n",
      "[652]\tvalid_0's rmse: 1823.9\n",
      "[653]\tvalid_0's rmse: 1823.64\n",
      "[654]\tvalid_0's rmse: 1823.65\n",
      "[655]\tvalid_0's rmse: 1823.63\n",
      "[656]\tvalid_0's rmse: 1823.47\n",
      "[657]\tvalid_0's rmse: 1823.32\n",
      "[658]\tvalid_0's rmse: 1823.22\n",
      "[659]\tvalid_0's rmse: 1823.18\n",
      "[660]\tvalid_0's rmse: 1823.09\n",
      "[661]\tvalid_0's rmse: 1822.95\n",
      "[662]\tvalid_0's rmse: 1822.94\n",
      "[663]\tvalid_0's rmse: 1822.84\n",
      "[664]\tvalid_0's rmse: 1822.8\n",
      "[665]\tvalid_0's rmse: 1822.73\n",
      "[666]\tvalid_0's rmse: 1822.79\n",
      "[667]\tvalid_0's rmse: 1822.78\n",
      "[668]\tvalid_0's rmse: 1822.73\n",
      "[669]\tvalid_0's rmse: 1822.7\n",
      "[670]\tvalid_0's rmse: 1822.55\n",
      "[671]\tvalid_0's rmse: 1822.35\n",
      "[672]\tvalid_0's rmse: 1822.25\n",
      "[673]\tvalid_0's rmse: 1822.07\n",
      "[674]\tvalid_0's rmse: 1822.04\n",
      "[675]\tvalid_0's rmse: 1821.84\n",
      "[676]\tvalid_0's rmse: 1821.73\n",
      "[677]\tvalid_0's rmse: 1821.75\n",
      "[678]\tvalid_0's rmse: 1821.78\n",
      "[679]\tvalid_0's rmse: 1821.68\n",
      "[680]\tvalid_0's rmse: 1821.53\n",
      "[681]\tvalid_0's rmse: 1821.39\n",
      "[682]\tvalid_0's rmse: 1821.34\n",
      "[683]\tvalid_0's rmse: 1821.31\n",
      "[684]\tvalid_0's rmse: 1821.26\n",
      "[685]\tvalid_0's rmse: 1821.13\n",
      "[686]\tvalid_0's rmse: 1821.07\n",
      "[687]\tvalid_0's rmse: 1820.94\n",
      "[688]\tvalid_0's rmse: 1820.77\n",
      "[689]\tvalid_0's rmse: 1820.76\n",
      "[690]\tvalid_0's rmse: 1820.46\n",
      "[691]\tvalid_0's rmse: 1820.33\n",
      "[692]\tvalid_0's rmse: 1820.15\n",
      "[693]\tvalid_0's rmse: 1820.07\n",
      "[694]\tvalid_0's rmse: 1820.01\n",
      "[695]\tvalid_0's rmse: 1819.94\n",
      "[696]\tvalid_0's rmse: 1819.82\n",
      "[697]\tvalid_0's rmse: 1819.78\n",
      "[698]\tvalid_0's rmse: 1819.61\n",
      "[699]\tvalid_0's rmse: 1819.55\n",
      "[700]\tvalid_0's rmse: 1819.52\n",
      "[701]\tvalid_0's rmse: 1819.51\n",
      "[702]\tvalid_0's rmse: 1819.42\n",
      "[703]\tvalid_0's rmse: 1819.33\n",
      "[704]\tvalid_0's rmse: 1819.25\n",
      "[705]\tvalid_0's rmse: 1819.13\n",
      "[706]\tvalid_0's rmse: 1819.01\n",
      "[707]\tvalid_0's rmse: 1819.02\n",
      "[708]\tvalid_0's rmse: 1818.92\n",
      "[709]\tvalid_0's rmse: 1818.89\n",
      "[710]\tvalid_0's rmse: 1818.75\n",
      "[711]\tvalid_0's rmse: 1818.71\n",
      "[712]\tvalid_0's rmse: 1818.68\n",
      "[713]\tvalid_0's rmse: 1818.62\n",
      "[714]\tvalid_0's rmse: 1818.44\n",
      "[715]\tvalid_0's rmse: 1818.37\n",
      "[716]\tvalid_0's rmse: 1818.34\n",
      "[717]\tvalid_0's rmse: 1818.28\n",
      "[718]\tvalid_0's rmse: 1818.25\n",
      "[719]\tvalid_0's rmse: 1818.19\n",
      "[720]\tvalid_0's rmse: 1817.94\n",
      "[721]\tvalid_0's rmse: 1817.85\n",
      "[722]\tvalid_0's rmse: 1817.81\n",
      "[723]\tvalid_0's rmse: 1817.76\n",
      "[724]\tvalid_0's rmse: 1817.61\n",
      "[725]\tvalid_0's rmse: 1817.55\n",
      "[726]\tvalid_0's rmse: 1817.44\n",
      "[727]\tvalid_0's rmse: 1817.28\n",
      "[728]\tvalid_0's rmse: 1817.25\n",
      "[729]\tvalid_0's rmse: 1817.14\n",
      "[730]\tvalid_0's rmse: 1817.14\n",
      "[731]\tvalid_0's rmse: 1817.11\n",
      "[732]\tvalid_0's rmse: 1816.97\n",
      "[733]\tvalid_0's rmse: 1816.9\n",
      "[734]\tvalid_0's rmse: 1816.87\n",
      "[735]\tvalid_0's rmse: 1816.72\n",
      "[736]\tvalid_0's rmse: 1816.67\n",
      "[737]\tvalid_0's rmse: 1816.5\n",
      "[738]\tvalid_0's rmse: 1816.4\n",
      "[739]\tvalid_0's rmse: 1816.31\n",
      "[740]\tvalid_0's rmse: 1816.27\n",
      "[741]\tvalid_0's rmse: 1816.14\n",
      "[742]\tvalid_0's rmse: 1816.01\n",
      "[743]\tvalid_0's rmse: 1815.92\n",
      "[744]\tvalid_0's rmse: 1815.89\n",
      "[745]\tvalid_0's rmse: 1815.83\n",
      "[746]\tvalid_0's rmse: 1815.85\n",
      "[747]\tvalid_0's rmse: 1815.77\n",
      "[748]\tvalid_0's rmse: 1815.6\n",
      "[749]\tvalid_0's rmse: 1815.52\n",
      "[750]\tvalid_0's rmse: 1815.44\n",
      "[751]\tvalid_0's rmse: 1815.45\n",
      "[752]\tvalid_0's rmse: 1815.32\n",
      "[753]\tvalid_0's rmse: 1815.29\n",
      "[754]\tvalid_0's rmse: 1815.18\n",
      "[755]\tvalid_0's rmse: 1815.15\n",
      "[756]\tvalid_0's rmse: 1815.1\n",
      "[757]\tvalid_0's rmse: 1815\n",
      "[758]\tvalid_0's rmse: 1814.86\n",
      "[759]\tvalid_0's rmse: 1814.8\n",
      "[760]\tvalid_0's rmse: 1814.62\n",
      "[761]\tvalid_0's rmse: 1814.54\n",
      "[762]\tvalid_0's rmse: 1814.43\n",
      "[763]\tvalid_0's rmse: 1814.4\n",
      "[764]\tvalid_0's rmse: 1814.26\n",
      "[765]\tvalid_0's rmse: 1814.18\n",
      "[766]\tvalid_0's rmse: 1814.12\n",
      "[767]\tvalid_0's rmse: 1814.09\n",
      "[768]\tvalid_0's rmse: 1814.05\n",
      "[769]\tvalid_0's rmse: 1813.97\n",
      "[770]\tvalid_0's rmse: 1813.88\n",
      "[771]\tvalid_0's rmse: 1813.8\n",
      "[772]\tvalid_0's rmse: 1813.82\n",
      "[773]\tvalid_0's rmse: 1813.73\n",
      "[774]\tvalid_0's rmse: 1813.74\n",
      "[775]\tvalid_0's rmse: 1813.51\n",
      "[776]\tvalid_0's rmse: 1813.46\n",
      "[777]\tvalid_0's rmse: 1813.44\n",
      "[778]\tvalid_0's rmse: 1813.37\n",
      "[779]\tvalid_0's rmse: 1813.36\n",
      "[780]\tvalid_0's rmse: 1813.35\n",
      "[781]\tvalid_0's rmse: 1813.28\n",
      "[782]\tvalid_0's rmse: 1813.23\n",
      "[783]\tvalid_0's rmse: 1813.18\n",
      "[784]\tvalid_0's rmse: 1813.15\n",
      "[785]\tvalid_0's rmse: 1813.01\n",
      "[786]\tvalid_0's rmse: 1813\n",
      "[787]\tvalid_0's rmse: 1812.97\n",
      "[788]\tvalid_0's rmse: 1812.86\n",
      "[789]\tvalid_0's rmse: 1812.76\n",
      "[790]\tvalid_0's rmse: 1812.68\n",
      "[791]\tvalid_0's rmse: 1812.62\n",
      "[792]\tvalid_0's rmse: 1812.61\n",
      "[793]\tvalid_0's rmse: 1812.56\n",
      "[794]\tvalid_0's rmse: 1812.49\n",
      "[795]\tvalid_0's rmse: 1812.34\n",
      "[796]\tvalid_0's rmse: 1812.3\n",
      "[797]\tvalid_0's rmse: 1812.32\n",
      "[798]\tvalid_0's rmse: 1812.28\n",
      "[799]\tvalid_0's rmse: 1812.22\n",
      "[800]\tvalid_0's rmse: 1812.08\n",
      "[801]\tvalid_0's rmse: 1812.01\n",
      "[802]\tvalid_0's rmse: 1811.92\n",
      "[803]\tvalid_0's rmse: 1811.91\n",
      "[804]\tvalid_0's rmse: 1811.9\n",
      "[805]\tvalid_0's rmse: 1811.75\n",
      "[806]\tvalid_0's rmse: 1811.69\n",
      "[807]\tvalid_0's rmse: 1811.61\n",
      "[808]\tvalid_0's rmse: 1811.58\n",
      "[809]\tvalid_0's rmse: 1811.46\n",
      "[810]\tvalid_0's rmse: 1811.42\n",
      "[811]\tvalid_0's rmse: 1811.44\n",
      "[812]\tvalid_0's rmse: 1811.42\n",
      "[813]\tvalid_0's rmse: 1811.39\n",
      "[814]\tvalid_0's rmse: 1811.35\n",
      "[815]\tvalid_0's rmse: 1811.32\n",
      "[816]\tvalid_0's rmse: 1811.19\n",
      "[817]\tvalid_0's rmse: 1811.13\n",
      "[818]\tvalid_0's rmse: 1811.11\n",
      "[819]\tvalid_0's rmse: 1811.05\n",
      "[820]\tvalid_0's rmse: 1811.03\n",
      "[821]\tvalid_0's rmse: 1811\n",
      "[822]\tvalid_0's rmse: 1810.99\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[823]\tvalid_0's rmse: 1810.99\n",
      "[824]\tvalid_0's rmse: 1810.86\n",
      "[825]\tvalid_0's rmse: 1810.74\n",
      "[826]\tvalid_0's rmse: 1810.58\n",
      "[827]\tvalid_0's rmse: 1810.54\n",
      "[828]\tvalid_0's rmse: 1810.45\n",
      "[829]\tvalid_0's rmse: 1810.42\n",
      "[830]\tvalid_0's rmse: 1810.36\n",
      "[831]\tvalid_0's rmse: 1810.31\n",
      "[832]\tvalid_0's rmse: 1810.27\n",
      "[833]\tvalid_0's rmse: 1810.23\n",
      "[834]\tvalid_0's rmse: 1810.19\n",
      "[835]\tvalid_0's rmse: 1810.14\n",
      "[836]\tvalid_0's rmse: 1810.12\n",
      "[837]\tvalid_0's rmse: 1810.11\n",
      "[838]\tvalid_0's rmse: 1810.1\n",
      "[839]\tvalid_0's rmse: 1809.95\n",
      "[840]\tvalid_0's rmse: 1809.94\n",
      "[841]\tvalid_0's rmse: 1809.86\n",
      "[842]\tvalid_0's rmse: 1809.79\n",
      "[843]\tvalid_0's rmse: 1809.77\n",
      "[844]\tvalid_0's rmse: 1809.78\n",
      "[845]\tvalid_0's rmse: 1809.71\n",
      "[846]\tvalid_0's rmse: 1809.76\n",
      "[847]\tvalid_0's rmse: 1809.75\n",
      "[848]\tvalid_0's rmse: 1809.75\n",
      "[849]\tvalid_0's rmse: 1809.61\n",
      "[850]\tvalid_0's rmse: 1809.36\n",
      "[851]\tvalid_0's rmse: 1809.16\n",
      "[852]\tvalid_0's rmse: 1809.07\n",
      "[853]\tvalid_0's rmse: 1809.05\n",
      "[854]\tvalid_0's rmse: 1808.96\n",
      "[855]\tvalid_0's rmse: 1808.87\n",
      "[856]\tvalid_0's rmse: 1808.79\n",
      "[857]\tvalid_0's rmse: 1808.78\n",
      "[858]\tvalid_0's rmse: 1808.7\n",
      "[859]\tvalid_0's rmse: 1808.67\n",
      "[860]\tvalid_0's rmse: 1808.62\n",
      "[861]\tvalid_0's rmse: 1808.57\n",
      "[862]\tvalid_0's rmse: 1808.54\n",
      "[863]\tvalid_0's rmse: 1808.54\n",
      "[864]\tvalid_0's rmse: 1808.52\n",
      "[865]\tvalid_0's rmse: 1808.55\n",
      "[866]\tvalid_0's rmse: 1808.5\n",
      "[867]\tvalid_0's rmse: 1808.48\n",
      "[868]\tvalid_0's rmse: 1808.37\n",
      "[869]\tvalid_0's rmse: 1808.33\n",
      "[870]\tvalid_0's rmse: 1808.28\n",
      "[871]\tvalid_0's rmse: 1808.27\n",
      "[872]\tvalid_0's rmse: 1808.15\n",
      "[873]\tvalid_0's rmse: 1808.14\n",
      "[874]\tvalid_0's rmse: 1808.13\n",
      "[875]\tvalid_0's rmse: 1808.02\n",
      "[876]\tvalid_0's rmse: 1808.01\n",
      "[877]\tvalid_0's rmse: 1807.94\n",
      "[878]\tvalid_0's rmse: 1807.93\n",
      "[879]\tvalid_0's rmse: 1807.91\n",
      "[880]\tvalid_0's rmse: 1807.85\n",
      "[881]\tvalid_0's rmse: 1807.81\n",
      "[882]\tvalid_0's rmse: 1807.78\n",
      "[883]\tvalid_0's rmse: 1807.77\n",
      "[884]\tvalid_0's rmse: 1807.74\n",
      "[885]\tvalid_0's rmse: 1807.7\n",
      "[886]\tvalid_0's rmse: 1807.67\n",
      "[887]\tvalid_0's rmse: 1807.77\n",
      "[888]\tvalid_0's rmse: 1807.7\n",
      "[889]\tvalid_0's rmse: 1807.6\n",
      "[890]\tvalid_0's rmse: 1807.61\n",
      "[891]\tvalid_0's rmse: 1807.59\n",
      "[892]\tvalid_0's rmse: 1807.55\n",
      "[893]\tvalid_0's rmse: 1807.47\n",
      "[894]\tvalid_0's rmse: 1807.37\n",
      "[895]\tvalid_0's rmse: 1807.3\n",
      "[896]\tvalid_0's rmse: 1807.23\n",
      "[897]\tvalid_0's rmse: 1807.13\n",
      "[898]\tvalid_0's rmse: 1807.1\n",
      "[899]\tvalid_0's rmse: 1806.99\n",
      "[900]\tvalid_0's rmse: 1806.89\n",
      "[901]\tvalid_0's rmse: 1806.74\n",
      "[902]\tvalid_0's rmse: 1806.68\n",
      "[903]\tvalid_0's rmse: 1806.63\n",
      "[904]\tvalid_0's rmse: 1806.59\n",
      "[905]\tvalid_0's rmse: 1806.49\n",
      "[906]\tvalid_0's rmse: 1806.45\n",
      "[907]\tvalid_0's rmse: 1806.39\n",
      "[908]\tvalid_0's rmse: 1806.33\n",
      "[909]\tvalid_0's rmse: 1806.21\n",
      "[910]\tvalid_0's rmse: 1806.06\n",
      "[911]\tvalid_0's rmse: 1806.02\n",
      "[912]\tvalid_0's rmse: 1805.94\n",
      "[913]\tvalid_0's rmse: 1805.92\n",
      "[914]\tvalid_0's rmse: 1805.89\n",
      "[915]\tvalid_0's rmse: 1805.89\n",
      "[916]\tvalid_0's rmse: 1805.85\n",
      "[917]\tvalid_0's rmse: 1805.84\n",
      "[918]\tvalid_0's rmse: 1805.79\n",
      "[919]\tvalid_0's rmse: 1805.72\n",
      "[920]\tvalid_0's rmse: 1805.71\n",
      "[921]\tvalid_0's rmse: 1805.69\n",
      "[922]\tvalid_0's rmse: 1805.67\n",
      "[923]\tvalid_0's rmse: 1805.65\n",
      "[924]\tvalid_0's rmse: 1805.62\n",
      "[925]\tvalid_0's rmse: 1805.61\n",
      "[926]\tvalid_0's rmse: 1805.49\n",
      "[927]\tvalid_0's rmse: 1805.48\n",
      "[928]\tvalid_0's rmse: 1805.35\n",
      "[929]\tvalid_0's rmse: 1805\n",
      "[930]\tvalid_0's rmse: 1804.96\n",
      "[931]\tvalid_0's rmse: 1804.86\n",
      "[932]\tvalid_0's rmse: 1804.72\n",
      "[933]\tvalid_0's rmse: 1804.64\n",
      "[934]\tvalid_0's rmse: 1804.62\n",
      "[935]\tvalid_0's rmse: 1804.57\n",
      "[936]\tvalid_0's rmse: 1804.52\n",
      "[937]\tvalid_0's rmse: 1804.44\n",
      "[938]\tvalid_0's rmse: 1804.43\n",
      "[939]\tvalid_0's rmse: 1804.4\n",
      "[940]\tvalid_0's rmse: 1804.38\n",
      "[941]\tvalid_0's rmse: 1804.4\n",
      "[942]\tvalid_0's rmse: 1804.31\n",
      "[943]\tvalid_0's rmse: 1804.27\n",
      "[944]\tvalid_0's rmse: 1804.16\n",
      "[945]\tvalid_0's rmse: 1804.24\n",
      "[946]\tvalid_0's rmse: 1804.19\n",
      "[947]\tvalid_0's rmse: 1804.22\n",
      "[948]\tvalid_0's rmse: 1804.09\n",
      "[949]\tvalid_0's rmse: 1804.09\n",
      "[950]\tvalid_0's rmse: 1804.04\n",
      "[951]\tvalid_0's rmse: 1804\n",
      "[952]\tvalid_0's rmse: 1803.86\n",
      "[953]\tvalid_0's rmse: 1803.84\n",
      "[954]\tvalid_0's rmse: 1803.79\n",
      "[955]\tvalid_0's rmse: 1803.7\n",
      "[956]\tvalid_0's rmse: 1803.67\n",
      "[957]\tvalid_0's rmse: 1803.61\n",
      "[958]\tvalid_0's rmse: 1803.53\n",
      "[959]\tvalid_0's rmse: 1803.49\n",
      "[960]\tvalid_0's rmse: 1803.36\n",
      "[961]\tvalid_0's rmse: 1803.34\n",
      "[962]\tvalid_0's rmse: 1803.23\n",
      "[963]\tvalid_0's rmse: 1803.19\n",
      "[964]\tvalid_0's rmse: 1803.18\n",
      "[965]\tvalid_0's rmse: 1803.16\n",
      "[966]\tvalid_0's rmse: 1803.08\n",
      "[967]\tvalid_0's rmse: 1803.03\n",
      "[968]\tvalid_0's rmse: 1802.98\n",
      "[969]\tvalid_0's rmse: 1802.94\n",
      "[970]\tvalid_0's rmse: 1802.87\n",
      "[971]\tvalid_0's rmse: 1802.86\n",
      "[972]\tvalid_0's rmse: 1802.85\n",
      "[973]\tvalid_0's rmse: 1802.83\n",
      "[974]\tvalid_0's rmse: 1802.82\n",
      "[975]\tvalid_0's rmse: 1802.68\n",
      "[976]\tvalid_0's rmse: 1802.64\n",
      "[977]\tvalid_0's rmse: 1802.52\n",
      "[978]\tvalid_0's rmse: 1802.45\n",
      "[979]\tvalid_0's rmse: 1802.42\n",
      "[980]\tvalid_0's rmse: 1802.43\n",
      "[981]\tvalid_0's rmse: 1802.39\n",
      "[982]\tvalid_0's rmse: 1802.29\n",
      "[983]\tvalid_0's rmse: 1802.23\n",
      "[984]\tvalid_0's rmse: 1802.09\n",
      "[985]\tvalid_0's rmse: 1801.98\n",
      "[986]\tvalid_0's rmse: 1802.05\n",
      "[987]\tvalid_0's rmse: 1801.86\n",
      "[988]\tvalid_0's rmse: 1801.86\n",
      "[989]\tvalid_0's rmse: 1801.75\n",
      "[990]\tvalid_0's rmse: 1801.74\n",
      "[991]\tvalid_0's rmse: 1801.68\n",
      "[992]\tvalid_0's rmse: 1801.66\n",
      "[993]\tvalid_0's rmse: 1801.66\n",
      "[994]\tvalid_0's rmse: 1801.58\n",
      "[995]\tvalid_0's rmse: 1801.49\n",
      "[996]\tvalid_0's rmse: 1801.48\n",
      "[997]\tvalid_0's rmse: 1801.42\n",
      "[998]\tvalid_0's rmse: 1801.35\n",
      "[999]\tvalid_0's rmse: 1801.33\n",
      "[1000]\tvalid_0's rmse: 1801.3\n",
      "[1001]\tvalid_0's rmse: 1801.32\n",
      "[1002]\tvalid_0's rmse: 1801.27\n",
      "[1003]\tvalid_0's rmse: 1801.26\n",
      "[1004]\tvalid_0's rmse: 1801.23\n",
      "[1005]\tvalid_0's rmse: 1801.21\n",
      "[1006]\tvalid_0's rmse: 1801.21\n",
      "[1007]\tvalid_0's rmse: 1801.26\n",
      "[1008]\tvalid_0's rmse: 1801.19\n",
      "[1009]\tvalid_0's rmse: 1801.1\n",
      "[1010]\tvalid_0's rmse: 1801.04\n",
      "[1011]\tvalid_0's rmse: 1801\n",
      "[1012]\tvalid_0's rmse: 1800.98\n",
      "[1013]\tvalid_0's rmse: 1801.01\n",
      "[1014]\tvalid_0's rmse: 1800.91\n",
      "[1015]\tvalid_0's rmse: 1800.85\n",
      "[1016]\tvalid_0's rmse: 1800.82\n",
      "[1017]\tvalid_0's rmse: 1800.79\n",
      "[1018]\tvalid_0's rmse: 1800.72\n",
      "[1019]\tvalid_0's rmse: 1800.69\n",
      "[1020]\tvalid_0's rmse: 1800.65\n",
      "[1021]\tvalid_0's rmse: 1800.6\n",
      "[1022]\tvalid_0's rmse: 1800.57\n",
      "[1023]\tvalid_0's rmse: 1800.49\n",
      "[1024]\tvalid_0's rmse: 1800.47\n",
      "[1025]\tvalid_0's rmse: 1800.46\n",
      "[1026]\tvalid_0's rmse: 1800.43\n",
      "[1027]\tvalid_0's rmse: 1800.42\n",
      "[1028]\tvalid_0's rmse: 1800.34\n",
      "[1029]\tvalid_0's rmse: 1800.22\n",
      "[1030]\tvalid_0's rmse: 1800.22\n",
      "[1031]\tvalid_0's rmse: 1799.99\n",
      "[1032]\tvalid_0's rmse: 1799.98\n",
      "[1033]\tvalid_0's rmse: 1799.96\n",
      "[1034]\tvalid_0's rmse: 1799.82\n",
      "[1035]\tvalid_0's rmse: 1799.75\n",
      "[1036]\tvalid_0's rmse: 1799.66\n",
      "[1037]\tvalid_0's rmse: 1799.55\n",
      "[1038]\tvalid_0's rmse: 1799.5\n",
      "[1039]\tvalid_0's rmse: 1799.44\n",
      "[1040]\tvalid_0's rmse: 1799.43\n",
      "[1041]\tvalid_0's rmse: 1799.36\n",
      "[1042]\tvalid_0's rmse: 1799.3\n",
      "[1043]\tvalid_0's rmse: 1799.18\n",
      "[1044]\tvalid_0's rmse: 1799.15\n",
      "[1045]\tvalid_0's rmse: 1799.08\n",
      "[1046]\tvalid_0's rmse: 1798.98\n",
      "[1047]\tvalid_0's rmse: 1799.02\n",
      "[1048]\tvalid_0's rmse: 1798.96\n",
      "[1049]\tvalid_0's rmse: 1798.95\n",
      "[1050]\tvalid_0's rmse: 1798.95\n",
      "[1051]\tvalid_0's rmse: 1798.91\n",
      "[1052]\tvalid_0's rmse: 1798.84\n",
      "[1053]\tvalid_0's rmse: 1798.8\n",
      "[1054]\tvalid_0's rmse: 1798.7\n",
      "[1055]\tvalid_0's rmse: 1798.67\n",
      "[1056]\tvalid_0's rmse: 1798.65\n",
      "[1057]\tvalid_0's rmse: 1798.66\n",
      "[1058]\tvalid_0's rmse: 1798.6\n",
      "[1059]\tvalid_0's rmse: 1798.56\n",
      "[1060]\tvalid_0's rmse: 1798.45\n",
      "[1061]\tvalid_0's rmse: 1798.44\n",
      "[1062]\tvalid_0's rmse: 1798.43\n",
      "[1063]\tvalid_0's rmse: 1798.38\n",
      "[1064]\tvalid_0's rmse: 1798.34\n",
      "[1065]\tvalid_0's rmse: 1798.33\n",
      "[1066]\tvalid_0's rmse: 1798.35\n",
      "[1067]\tvalid_0's rmse: 1798.32\n",
      "[1068]\tvalid_0's rmse: 1798.32\n",
      "[1069]\tvalid_0's rmse: 1798.32\n",
      "[1070]\tvalid_0's rmse: 1798.29\n",
      "[1071]\tvalid_0's rmse: 1798.26\n",
      "[1072]\tvalid_0's rmse: 1798.24\n",
      "[1073]\tvalid_0's rmse: 1798.17\n",
      "[1074]\tvalid_0's rmse: 1798.17\n",
      "[1075]\tvalid_0's rmse: 1798.16\n",
      "[1076]\tvalid_0's rmse: 1798.12\n",
      "[1077]\tvalid_0's rmse: 1798.07\n",
      "[1078]\tvalid_0's rmse: 1798.05\n",
      "[1079]\tvalid_0's rmse: 1798.07\n",
      "[1080]\tvalid_0's rmse: 1798.05\n",
      "[1081]\tvalid_0's rmse: 1798.03\n",
      "[1082]\tvalid_0's rmse: 1797.98\n",
      "[1083]\tvalid_0's rmse: 1797.96\n",
      "[1084]\tvalid_0's rmse: 1797.9\n",
      "[1085]\tvalid_0's rmse: 1797.88\n",
      "[1086]\tvalid_0's rmse: 1797.88\n",
      "[1087]\tvalid_0's rmse: 1797.87\n",
      "[1088]\tvalid_0's rmse: 1797.87\n",
      "[1089]\tvalid_0's rmse: 1797.82\n",
      "[1090]\tvalid_0's rmse: 1797.76\n",
      "[1091]\tvalid_0's rmse: 1797.68\n",
      "[1092]\tvalid_0's rmse: 1797.67\n",
      "[1093]\tvalid_0's rmse: 1797.66\n",
      "[1094]\tvalid_0's rmse: 1797.62\n",
      "[1095]\tvalid_0's rmse: 1797.57\n",
      "[1096]\tvalid_0's rmse: 1797.53\n",
      "[1097]\tvalid_0's rmse: 1797.48\n",
      "[1098]\tvalid_0's rmse: 1797.43\n",
      "[1099]\tvalid_0's rmse: 1797.4\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1100]\tvalid_0's rmse: 1797.35\n",
      "[1101]\tvalid_0's rmse: 1797.32\n",
      "[1102]\tvalid_0's rmse: 1797.31\n",
      "[1103]\tvalid_0's rmse: 1797.28\n",
      "[1104]\tvalid_0's rmse: 1797.22\n",
      "[1105]\tvalid_0's rmse: 1797.16\n",
      "[1106]\tvalid_0's rmse: 1797.11\n",
      "[1107]\tvalid_0's rmse: 1797.1\n",
      "[1108]\tvalid_0's rmse: 1797.09\n",
      "[1109]\tvalid_0's rmse: 1797.01\n",
      "[1110]\tvalid_0's rmse: 1797.04\n",
      "[1111]\tvalid_0's rmse: 1797.06\n",
      "[1112]\tvalid_0's rmse: 1797.05\n",
      "[1113]\tvalid_0's rmse: 1797.06\n",
      "[1114]\tvalid_0's rmse: 1796.97\n",
      "[1115]\tvalid_0's rmse: 1796.96\n",
      "[1116]\tvalid_0's rmse: 1796.95\n",
      "[1117]\tvalid_0's rmse: 1796.89\n",
      "[1118]\tvalid_0's rmse: 1796.77\n",
      "[1119]\tvalid_0's rmse: 1796.75\n",
      "[1120]\tvalid_0's rmse: 1796.71\n",
      "[1121]\tvalid_0's rmse: 1796.64\n",
      "[1122]\tvalid_0's rmse: 1796.66\n",
      "[1123]\tvalid_0's rmse: 1796.64\n",
      "[1124]\tvalid_0's rmse: 1796.62\n",
      "[1125]\tvalid_0's rmse: 1796.58\n",
      "[1126]\tvalid_0's rmse: 1796.47\n",
      "[1127]\tvalid_0's rmse: 1796.44\n",
      "[1128]\tvalid_0's rmse: 1796.41\n",
      "[1129]\tvalid_0's rmse: 1796.38\n",
      "[1130]\tvalid_0's rmse: 1796.32\n",
      "[1131]\tvalid_0's rmse: 1796.29\n",
      "[1132]\tvalid_0's rmse: 1796.27\n",
      "[1133]\tvalid_0's rmse: 1796.16\n",
      "[1134]\tvalid_0's rmse: 1796.12\n",
      "[1135]\tvalid_0's rmse: 1796.08\n",
      "[1136]\tvalid_0's rmse: 1796.05\n",
      "[1137]\tvalid_0's rmse: 1796\n",
      "[1138]\tvalid_0's rmse: 1795.98\n",
      "[1139]\tvalid_0's rmse: 1795.92\n",
      "[1140]\tvalid_0's rmse: 1795.9\n",
      "[1141]\tvalid_0's rmse: 1795.88\n",
      "[1142]\tvalid_0's rmse: 1795.82\n",
      "[1143]\tvalid_0's rmse: 1795.83\n",
      "[1144]\tvalid_0's rmse: 1795.86\n",
      "[1145]\tvalid_0's rmse: 1795.8\n",
      "[1146]\tvalid_0's rmse: 1795.75\n",
      "[1147]\tvalid_0's rmse: 1795.73\n",
      "[1148]\tvalid_0's rmse: 1795.7\n",
      "[1149]\tvalid_0's rmse: 1795.68\n",
      "[1150]\tvalid_0's rmse: 1795.64\n",
      "[1151]\tvalid_0's rmse: 1795.59\n",
      "[1152]\tvalid_0's rmse: 1795.57\n",
      "[1153]\tvalid_0's rmse: 1795.54\n",
      "[1154]\tvalid_0's rmse: 1795.53\n",
      "[1155]\tvalid_0's rmse: 1795.45\n",
      "[1156]\tvalid_0's rmse: 1795.38\n",
      "[1157]\tvalid_0's rmse: 1795.28\n",
      "[1158]\tvalid_0's rmse: 1795.26\n",
      "[1159]\tvalid_0's rmse: 1795.23\n",
      "[1160]\tvalid_0's rmse: 1795.21\n",
      "[1161]\tvalid_0's rmse: 1795.22\n",
      "[1162]\tvalid_0's rmse: 1795.22\n",
      "[1163]\tvalid_0's rmse: 1795.21\n",
      "[1164]\tvalid_0's rmse: 1795.07\n",
      "[1165]\tvalid_0's rmse: 1795.06\n",
      "[1166]\tvalid_0's rmse: 1795.03\n",
      "[1167]\tvalid_0's rmse: 1794.97\n",
      "[1168]\tvalid_0's rmse: 1794.94\n",
      "[1169]\tvalid_0's rmse: 1794.85\n",
      "[1170]\tvalid_0's rmse: 1794.86\n",
      "[1171]\tvalid_0's rmse: 1794.84\n",
      "[1172]\tvalid_0's rmse: 1794.8\n",
      "[1173]\tvalid_0's rmse: 1794.76\n",
      "[1174]\tvalid_0's rmse: 1794.74\n",
      "[1175]\tvalid_0's rmse: 1794.71\n",
      "[1176]\tvalid_0's rmse: 1794.69\n",
      "[1177]\tvalid_0's rmse: 1794.65\n",
      "[1178]\tvalid_0's rmse: 1794.65\n",
      "[1179]\tvalid_0's rmse: 1794.56\n",
      "[1180]\tvalid_0's rmse: 1794.6\n",
      "[1181]\tvalid_0's rmse: 1794.58\n",
      "[1182]\tvalid_0's rmse: 1794.55\n",
      "[1183]\tvalid_0's rmse: 1794.53\n",
      "[1184]\tvalid_0's rmse: 1794.51\n",
      "[1185]\tvalid_0's rmse: 1794.48\n",
      "[1186]\tvalid_0's rmse: 1794.34\n",
      "[1187]\tvalid_0's rmse: 1794.21\n",
      "[1188]\tvalid_0's rmse: 1794.2\n",
      "[1189]\tvalid_0's rmse: 1794.2\n",
      "[1190]\tvalid_0's rmse: 1794.21\n",
      "[1191]\tvalid_0's rmse: 1794.09\n",
      "[1192]\tvalid_0's rmse: 1794.08\n",
      "[1193]\tvalid_0's rmse: 1794.01\n",
      "[1194]\tvalid_0's rmse: 1793.81\n",
      "[1195]\tvalid_0's rmse: 1793.78\n",
      "[1196]\tvalid_0's rmse: 1793.77\n",
      "[1197]\tvalid_0's rmse: 1793.71\n",
      "[1198]\tvalid_0's rmse: 1793.67\n",
      "[1199]\tvalid_0's rmse: 1793.67\n",
      "[1200]\tvalid_0's rmse: 1793.48\n",
      "[1201]\tvalid_0's rmse: 1793.48\n",
      "[1202]\tvalid_0's rmse: 1793.47\n",
      "[1203]\tvalid_0's rmse: 1793.43\n",
      "[1204]\tvalid_0's rmse: 1793.4\n",
      "[1205]\tvalid_0's rmse: 1793.36\n",
      "[1206]\tvalid_0's rmse: 1793.32\n",
      "[1207]\tvalid_0's rmse: 1793.24\n",
      "[1208]\tvalid_0's rmse: 1793.27\n",
      "[1209]\tvalid_0's rmse: 1793.19\n",
      "[1210]\tvalid_0's rmse: 1793.18\n",
      "[1211]\tvalid_0's rmse: 1793.19\n",
      "[1212]\tvalid_0's rmse: 1793.17\n",
      "[1213]\tvalid_0's rmse: 1793.16\n",
      "[1214]\tvalid_0's rmse: 1793.15\n",
      "[1215]\tvalid_0's rmse: 1793.15\n",
      "[1216]\tvalid_0's rmse: 1793.14\n",
      "[1217]\tvalid_0's rmse: 1793.11\n",
      "[1218]\tvalid_0's rmse: 1793.09\n",
      "[1219]\tvalid_0's rmse: 1793.08\n",
      "[1220]\tvalid_0's rmse: 1793.06\n",
      "[1221]\tvalid_0's rmse: 1793\n",
      "[1222]\tvalid_0's rmse: 1792.99\n",
      "[1223]\tvalid_0's rmse: 1792.97\n",
      "[1224]\tvalid_0's rmse: 1792.98\n",
      "[1225]\tvalid_0's rmse: 1792.92\n",
      "[1226]\tvalid_0's rmse: 1792.94\n",
      "[1227]\tvalid_0's rmse: 1792.9\n",
      "[1228]\tvalid_0's rmse: 1792.9\n",
      "[1229]\tvalid_0's rmse: 1792.83\n",
      "[1230]\tvalid_0's rmse: 1792.74\n",
      "[1231]\tvalid_0's rmse: 1792.7\n",
      "[1232]\tvalid_0's rmse: 1792.66\n",
      "[1233]\tvalid_0's rmse: 1792.59\n",
      "[1234]\tvalid_0's rmse: 1792.57\n",
      "[1235]\tvalid_0's rmse: 1792.47\n",
      "[1236]\tvalid_0's rmse: 1792.4\n",
      "[1237]\tvalid_0's rmse: 1792.37\n",
      "[1238]\tvalid_0's rmse: 1792.32\n",
      "[1239]\tvalid_0's rmse: 1792.28\n",
      "[1240]\tvalid_0's rmse: 1792.27\n",
      "[1241]\tvalid_0's rmse: 1792.26\n",
      "[1242]\tvalid_0's rmse: 1792.21\n",
      "[1243]\tvalid_0's rmse: 1792.18\n",
      "[1244]\tvalid_0's rmse: 1792.14\n",
      "[1245]\tvalid_0's rmse: 1792.14\n",
      "[1246]\tvalid_0's rmse: 1792.15\n",
      "[1247]\tvalid_0's rmse: 1792.13\n",
      "[1248]\tvalid_0's rmse: 1792.11\n",
      "[1249]\tvalid_0's rmse: 1792.1\n",
      "[1250]\tvalid_0's rmse: 1792.03\n",
      "[1251]\tvalid_0's rmse: 1791.97\n",
      "[1252]\tvalid_0's rmse: 1791.93\n",
      "[1253]\tvalid_0's rmse: 1791.92\n",
      "[1254]\tvalid_0's rmse: 1791.92\n",
      "[1255]\tvalid_0's rmse: 1791.92\n",
      "[1256]\tvalid_0's rmse: 1791.93\n",
      "[1257]\tvalid_0's rmse: 1791.91\n",
      "[1258]\tvalid_0's rmse: 1791.9\n",
      "[1259]\tvalid_0's rmse: 1791.9\n",
      "[1260]\tvalid_0's rmse: 1791.87\n",
      "[1261]\tvalid_0's rmse: 1791.84\n",
      "[1262]\tvalid_0's rmse: 1791.82\n",
      "[1263]\tvalid_0's rmse: 1791.75\n",
      "[1264]\tvalid_0's rmse: 1791.75\n",
      "[1265]\tvalid_0's rmse: 1791.74\n",
      "[1266]\tvalid_0's rmse: 1791.65\n",
      "[1267]\tvalid_0's rmse: 1791.6\n",
      "[1268]\tvalid_0's rmse: 1791.58\n",
      "[1269]\tvalid_0's rmse: 1791.4\n",
      "[1270]\tvalid_0's rmse: 1791.33\n",
      "[1271]\tvalid_0's rmse: 1791.31\n",
      "[1272]\tvalid_0's rmse: 1791.29\n",
      "[1273]\tvalid_0's rmse: 1791.28\n",
      "[1274]\tvalid_0's rmse: 1791.24\n",
      "[1275]\tvalid_0's rmse: 1791.24\n",
      "[1276]\tvalid_0's rmse: 1791.22\n",
      "[1277]\tvalid_0's rmse: 1791.13\n",
      "[1278]\tvalid_0's rmse: 1791.05\n",
      "[1279]\tvalid_0's rmse: 1791.05\n",
      "[1280]\tvalid_0's rmse: 1791.04\n",
      "[1281]\tvalid_0's rmse: 1790.99\n",
      "[1282]\tvalid_0's rmse: 1790.96\n",
      "[1283]\tvalid_0's rmse: 1790.95\n",
      "[1284]\tvalid_0's rmse: 1790.94\n",
      "[1285]\tvalid_0's rmse: 1790.92\n",
      "[1286]\tvalid_0's rmse: 1790.89\n",
      "[1287]\tvalid_0's rmse: 1790.86\n",
      "[1288]\tvalid_0's rmse: 1790.86\n",
      "[1289]\tvalid_0's rmse: 1790.77\n",
      "[1290]\tvalid_0's rmse: 1790.78\n",
      "[1291]\tvalid_0's rmse: 1790.77\n",
      "[1292]\tvalid_0's rmse: 1790.79\n",
      "[1293]\tvalid_0's rmse: 1790.77\n",
      "[1294]\tvalid_0's rmse: 1790.77\n",
      "[1295]\tvalid_0's rmse: 1790.74\n",
      "[1296]\tvalid_0's rmse: 1790.71\n",
      "[1297]\tvalid_0's rmse: 1790.69\n",
      "[1298]\tvalid_0's rmse: 1790.7\n",
      "[1299]\tvalid_0's rmse: 1790.7\n",
      "[1300]\tvalid_0's rmse: 1790.66\n",
      "[1301]\tvalid_0's rmse: 1790.64\n",
      "[1302]\tvalid_0's rmse: 1790.62\n",
      "[1303]\tvalid_0's rmse: 1790.62\n",
      "[1304]\tvalid_0's rmse: 1790.62\n",
      "[1305]\tvalid_0's rmse: 1790.62\n",
      "[1306]\tvalid_0's rmse: 1790.6\n",
      "[1307]\tvalid_0's rmse: 1790.54\n",
      "[1308]\tvalid_0's rmse: 1790.5\n",
      "[1309]\tvalid_0's rmse: 1790.48\n",
      "[1310]\tvalid_0's rmse: 1790.48\n",
      "[1311]\tvalid_0's rmse: 1790.39\n",
      "[1312]\tvalid_0's rmse: 1790.34\n",
      "[1313]\tvalid_0's rmse: 1790.35\n",
      "[1314]\tvalid_0's rmse: 1790.34\n",
      "[1315]\tvalid_0's rmse: 1790.32\n",
      "[1316]\tvalid_0's rmse: 1790.32\n",
      "[1317]\tvalid_0's rmse: 1790.28\n",
      "[1318]\tvalid_0's rmse: 1790.23\n",
      "[1319]\tvalid_0's rmse: 1790.14\n",
      "[1320]\tvalid_0's rmse: 1790.12\n",
      "[1321]\tvalid_0's rmse: 1790.06\n",
      "[1322]\tvalid_0's rmse: 1790.04\n",
      "[1323]\tvalid_0's rmse: 1789.98\n",
      "[1324]\tvalid_0's rmse: 1789.88\n",
      "[1325]\tvalid_0's rmse: 1789.84\n",
      "[1326]\tvalid_0's rmse: 1789.85\n",
      "[1327]\tvalid_0's rmse: 1789.82\n",
      "[1328]\tvalid_0's rmse: 1789.73\n",
      "[1329]\tvalid_0's rmse: 1789.65\n",
      "[1330]\tvalid_0's rmse: 1789.69\n",
      "[1331]\tvalid_0's rmse: 1789.65\n",
      "[1332]\tvalid_0's rmse: 1789.6\n",
      "[1333]\tvalid_0's rmse: 1789.56\n",
      "[1334]\tvalid_0's rmse: 1789.51\n",
      "[1335]\tvalid_0's rmse: 1789.44\n",
      "[1336]\tvalid_0's rmse: 1789.37\n",
      "[1337]\tvalid_0's rmse: 1789.35\n",
      "[1338]\tvalid_0's rmse: 1789.28\n",
      "[1339]\tvalid_0's rmse: 1789.22\n",
      "[1340]\tvalid_0's rmse: 1789.13\n",
      "[1341]\tvalid_0's rmse: 1789.11\n",
      "[1342]\tvalid_0's rmse: 1789.11\n",
      "[1343]\tvalid_0's rmse: 1789.06\n",
      "[1344]\tvalid_0's rmse: 1789.05\n",
      "[1345]\tvalid_0's rmse: 1789.05\n",
      "[1346]\tvalid_0's rmse: 1789.05\n",
      "[1347]\tvalid_0's rmse: 1788.98\n",
      "[1348]\tvalid_0's rmse: 1788.98\n",
      "[1349]\tvalid_0's rmse: 1788.96\n",
      "[1350]\tvalid_0's rmse: 1788.94\n",
      "[1351]\tvalid_0's rmse: 1788.94\n",
      "[1352]\tvalid_0's rmse: 1788.87\n",
      "[1353]\tvalid_0's rmse: 1788.84\n",
      "[1354]\tvalid_0's rmse: 1788.82\n",
      "[1355]\tvalid_0's rmse: 1788.82\n",
      "[1356]\tvalid_0's rmse: 1788.79\n",
      "[1357]\tvalid_0's rmse: 1788.79\n",
      "[1358]\tvalid_0's rmse: 1788.8\n",
      "[1359]\tvalid_0's rmse: 1788.78\n",
      "[1360]\tvalid_0's rmse: 1788.74\n",
      "[1361]\tvalid_0's rmse: 1788.75\n",
      "[1362]\tvalid_0's rmse: 1788.66\n",
      "[1363]\tvalid_0's rmse: 1788.71\n",
      "[1364]\tvalid_0's rmse: 1788.66\n",
      "[1365]\tvalid_0's rmse: 1788.65\n",
      "[1366]\tvalid_0's rmse: 1788.62\n",
      "[1367]\tvalid_0's rmse: 1788.5\n",
      "[1368]\tvalid_0's rmse: 1788.48\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1369]\tvalid_0's rmse: 1788.47\n",
      "[1370]\tvalid_0's rmse: 1788.45\n",
      "[1371]\tvalid_0's rmse: 1788.36\n",
      "[1372]\tvalid_0's rmse: 1788.34\n",
      "[1373]\tvalid_0's rmse: 1788.33\n",
      "[1374]\tvalid_0's rmse: 1788.32\n",
      "[1375]\tvalid_0's rmse: 1788.31\n",
      "[1376]\tvalid_0's rmse: 1788.25\n",
      "[1377]\tvalid_0's rmse: 1788.21\n",
      "[1378]\tvalid_0's rmse: 1788.18\n",
      "[1379]\tvalid_0's rmse: 1788.12\n",
      "[1380]\tvalid_0's rmse: 1788.11\n",
      "[1381]\tvalid_0's rmse: 1788.07\n",
      "[1382]\tvalid_0's rmse: 1788.06\n",
      "[1383]\tvalid_0's rmse: 1788.07\n",
      "[1384]\tvalid_0's rmse: 1788.07\n",
      "[1385]\tvalid_0's rmse: 1788.07\n",
      "[1386]\tvalid_0's rmse: 1788.06\n",
      "[1387]\tvalid_0's rmse: 1788.05\n",
      "[1388]\tvalid_0's rmse: 1788.03\n",
      "[1389]\tvalid_0's rmse: 1788.01\n",
      "[1390]\tvalid_0's rmse: 1788.02\n",
      "[1391]\tvalid_0's rmse: 1787.89\n",
      "[1392]\tvalid_0's rmse: 1787.88\n",
      "[1393]\tvalid_0's rmse: 1787.87\n",
      "[1394]\tvalid_0's rmse: 1787.86\n",
      "[1395]\tvalid_0's rmse: 1787.82\n",
      "[1396]\tvalid_0's rmse: 1787.74\n",
      "[1397]\tvalid_0's rmse: 1787.69\n",
      "[1398]\tvalid_0's rmse: 1787.68\n",
      "[1399]\tvalid_0's rmse: 1787.63\n",
      "[1400]\tvalid_0's rmse: 1787.61\n",
      "[1401]\tvalid_0's rmse: 1787.57\n",
      "[1402]\tvalid_0's rmse: 1787.54\n",
      "[1403]\tvalid_0's rmse: 1787.4\n",
      "[1404]\tvalid_0's rmse: 1787.32\n",
      "[1405]\tvalid_0's rmse: 1787.32\n",
      "[1406]\tvalid_0's rmse: 1787.3\n",
      "[1407]\tvalid_0's rmse: 1787.24\n",
      "[1408]\tvalid_0's rmse: 1787.24\n",
      "[1409]\tvalid_0's rmse: 1787.21\n",
      "[1410]\tvalid_0's rmse: 1787.17\n",
      "[1411]\tvalid_0's rmse: 1787.17\n",
      "[1412]\tvalid_0's rmse: 1787.16\n",
      "[1413]\tvalid_0's rmse: 1787.16\n",
      "[1414]\tvalid_0's rmse: 1787.14\n",
      "[1415]\tvalid_0's rmse: 1787.14\n",
      "[1416]\tvalid_0's rmse: 1787.15\n",
      "[1417]\tvalid_0's rmse: 1787.11\n",
      "[1418]\tvalid_0's rmse: 1787.08\n",
      "[1419]\tvalid_0's rmse: 1787.07\n",
      "[1420]\tvalid_0's rmse: 1787.03\n",
      "[1421]\tvalid_0's rmse: 1786.97\n",
      "[1422]\tvalid_0's rmse: 1786.99\n",
      "[1423]\tvalid_0's rmse: 1786.97\n",
      "[1424]\tvalid_0's rmse: 1786.96\n",
      "[1425]\tvalid_0's rmse: 1786.94\n",
      "[1426]\tvalid_0's rmse: 1786.88\n",
      "[1427]\tvalid_0's rmse: 1786.87\n",
      "[1428]\tvalid_0's rmse: 1786.83\n",
      "[1429]\tvalid_0's rmse: 1786.78\n",
      "[1430]\tvalid_0's rmse: 1786.77\n",
      "[1431]\tvalid_0's rmse: 1786.74\n",
      "[1432]\tvalid_0's rmse: 1786.69\n",
      "[1433]\tvalid_0's rmse: 1786.64\n",
      "[1434]\tvalid_0's rmse: 1786.6\n",
      "[1435]\tvalid_0's rmse: 1786.53\n",
      "[1436]\tvalid_0's rmse: 1786.53\n",
      "[1437]\tvalid_0's rmse: 1786.51\n",
      "[1438]\tvalid_0's rmse: 1786.48\n",
      "[1439]\tvalid_0's rmse: 1786.49\n",
      "[1440]\tvalid_0's rmse: 1786.33\n",
      "[1441]\tvalid_0's rmse: 1786.35\n",
      "[1442]\tvalid_0's rmse: 1786.34\n",
      "[1443]\tvalid_0's rmse: 1786.26\n",
      "[1444]\tvalid_0's rmse: 1786.26\n",
      "[1445]\tvalid_0's rmse: 1786.26\n",
      "[1446]\tvalid_0's rmse: 1786.24\n",
      "[1447]\tvalid_0's rmse: 1786.19\n",
      "[1448]\tvalid_0's rmse: 1786.14\n",
      "[1449]\tvalid_0's rmse: 1786.09\n",
      "[1450]\tvalid_0's rmse: 1786.07\n",
      "[1451]\tvalid_0's rmse: 1786.06\n",
      "[1452]\tvalid_0's rmse: 1786.05\n",
      "[1453]\tvalid_0's rmse: 1786.05\n",
      "[1454]\tvalid_0's rmse: 1786.04\n",
      "[1455]\tvalid_0's rmse: 1786.01\n",
      "[1456]\tvalid_0's rmse: 1785.96\n",
      "[1457]\tvalid_0's rmse: 1785.87\n",
      "[1458]\tvalid_0's rmse: 1785.85\n",
      "[1459]\tvalid_0's rmse: 1785.82\n",
      "[1460]\tvalid_0's rmse: 1785.77\n",
      "[1461]\tvalid_0's rmse: 1785.74\n",
      "[1462]\tvalid_0's rmse: 1785.69\n",
      "[1463]\tvalid_0's rmse: 1785.62\n",
      "[1464]\tvalid_0's rmse: 1785.65\n",
      "[1465]\tvalid_0's rmse: 1785.61\n",
      "[1466]\tvalid_0's rmse: 1785.62\n",
      "[1467]\tvalid_0's rmse: 1785.55\n",
      "[1468]\tvalid_0's rmse: 1785.54\n",
      "[1469]\tvalid_0's rmse: 1785.49\n",
      "[1470]\tvalid_0's rmse: 1785.5\n",
      "[1471]\tvalid_0's rmse: 1785.48\n",
      "[1472]\tvalid_0's rmse: 1785.5\n",
      "[1473]\tvalid_0's rmse: 1785.48\n",
      "[1474]\tvalid_0's rmse: 1785.46\n",
      "[1475]\tvalid_0's rmse: 1785.46\n",
      "[1476]\tvalid_0's rmse: 1785.45\n",
      "[1477]\tvalid_0's rmse: 1785.45\n",
      "[1478]\tvalid_0's rmse: 1785.44\n",
      "[1479]\tvalid_0's rmse: 1785.37\n",
      "[1480]\tvalid_0's rmse: 1785.3\n",
      "[1481]\tvalid_0's rmse: 1785.29\n",
      "[1482]\tvalid_0's rmse: 1785.26\n",
      "[1483]\tvalid_0's rmse: 1785.26\n",
      "[1484]\tvalid_0's rmse: 1785.28\n",
      "[1485]\tvalid_0's rmse: 1785.32\n",
      "[1486]\tvalid_0's rmse: 1785.3\n",
      "[1487]\tvalid_0's rmse: 1785.26\n",
      "[1488]\tvalid_0's rmse: 1785.24\n",
      "[1489]\tvalid_0's rmse: 1785.21\n",
      "[1490]\tvalid_0's rmse: 1785.2\n",
      "[1491]\tvalid_0's rmse: 1785.18\n",
      "[1492]\tvalid_0's rmse: 1785.18\n",
      "[1493]\tvalid_0's rmse: 1785.21\n",
      "[1494]\tvalid_0's rmse: 1785.2\n",
      "[1495]\tvalid_0's rmse: 1785.14\n",
      "[1496]\tvalid_0's rmse: 1785.03\n",
      "[1497]\tvalid_0's rmse: 1785.05\n",
      "[1498]\tvalid_0's rmse: 1785.02\n",
      "[1499]\tvalid_0's rmse: 1785.02\n",
      "[1500]\tvalid_0's rmse: 1784.99\n",
      "[1501]\tvalid_0's rmse: 1784.98\n",
      "[1502]\tvalid_0's rmse: 1784.98\n",
      "[1503]\tvalid_0's rmse: 1784.96\n",
      "[1504]\tvalid_0's rmse: 1784.94\n",
      "[1505]\tvalid_0's rmse: 1784.93\n",
      "[1506]\tvalid_0's rmse: 1784.82\n",
      "[1507]\tvalid_0's rmse: 1784.77\n",
      "[1508]\tvalid_0's rmse: 1784.73\n",
      "[1509]\tvalid_0's rmse: 1784.7\n",
      "[1510]\tvalid_0's rmse: 1784.68\n",
      "[1511]\tvalid_0's rmse: 1784.61\n",
      "[1512]\tvalid_0's rmse: 1784.56\n",
      "[1513]\tvalid_0's rmse: 1784.51\n",
      "[1514]\tvalid_0's rmse: 1784.51\n",
      "[1515]\tvalid_0's rmse: 1784.51\n",
      "[1516]\tvalid_0's rmse: 1784.5\n",
      "[1517]\tvalid_0's rmse: 1784.42\n",
      "[1518]\tvalid_0's rmse: 1784.44\n",
      "[1519]\tvalid_0's rmse: 1784.43\n",
      "[1520]\tvalid_0's rmse: 1784.44\n",
      "[1521]\tvalid_0's rmse: 1784.41\n",
      "[1522]\tvalid_0's rmse: 1784.4\n",
      "[1523]\tvalid_0's rmse: 1784.37\n",
      "[1524]\tvalid_0's rmse: 1784.32\n",
      "[1525]\tvalid_0's rmse: 1784.33\n",
      "[1526]\tvalid_0's rmse: 1784.32\n",
      "[1527]\tvalid_0's rmse: 1784.29\n",
      "[1528]\tvalid_0's rmse: 1784.28\n",
      "[1529]\tvalid_0's rmse: 1784.32\n",
      "[1530]\tvalid_0's rmse: 1784.31\n",
      "[1531]\tvalid_0's rmse: 1784.26\n",
      "[1532]\tvalid_0's rmse: 1784.24\n",
      "[1533]\tvalid_0's rmse: 1784.2\n",
      "[1534]\tvalid_0's rmse: 1784.16\n",
      "[1535]\tvalid_0's rmse: 1784.16\n",
      "[1536]\tvalid_0's rmse: 1784.11\n",
      "[1537]\tvalid_0's rmse: 1784.1\n",
      "[1538]\tvalid_0's rmse: 1784.1\n",
      "[1539]\tvalid_0's rmse: 1784.03\n",
      "[1540]\tvalid_0's rmse: 1783.95\n",
      "[1541]\tvalid_0's rmse: 1783.98\n",
      "[1542]\tvalid_0's rmse: 1783.88\n",
      "[1543]\tvalid_0's rmse: 1783.87\n",
      "[1544]\tvalid_0's rmse: 1783.84\n",
      "[1545]\tvalid_0's rmse: 1783.85\n",
      "[1546]\tvalid_0's rmse: 1783.83\n",
      "[1547]\tvalid_0's rmse: 1783.78\n",
      "[1548]\tvalid_0's rmse: 1783.74\n",
      "[1549]\tvalid_0's rmse: 1783.71\n",
      "[1550]\tvalid_0's rmse: 1783.64\n",
      "[1551]\tvalid_0's rmse: 1783.63\n",
      "[1552]\tvalid_0's rmse: 1783.61\n",
      "[1553]\tvalid_0's rmse: 1783.59\n",
      "[1554]\tvalid_0's rmse: 1783.53\n",
      "[1555]\tvalid_0's rmse: 1783.37\n",
      "[1556]\tvalid_0's rmse: 1783.39\n",
      "[1557]\tvalid_0's rmse: 1783.36\n",
      "[1558]\tvalid_0's rmse: 1783.36\n",
      "[1559]\tvalid_0's rmse: 1783.34\n",
      "[1560]\tvalid_0's rmse: 1783.31\n",
      "[1561]\tvalid_0's rmse: 1783.28\n",
      "[1562]\tvalid_0's rmse: 1783.23\n",
      "[1563]\tvalid_0's rmse: 1783.21\n",
      "[1564]\tvalid_0's rmse: 1783.17\n",
      "[1565]\tvalid_0's rmse: 1783.19\n",
      "[1566]\tvalid_0's rmse: 1783.19\n",
      "[1567]\tvalid_0's rmse: 1783.18\n",
      "[1568]\tvalid_0's rmse: 1783.18\n",
      "[1569]\tvalid_0's rmse: 1783.15\n",
      "[1570]\tvalid_0's rmse: 1783.15\n",
      "[1571]\tvalid_0's rmse: 1783.07\n",
      "[1572]\tvalid_0's rmse: 1782.98\n",
      "[1573]\tvalid_0's rmse: 1782.92\n",
      "[1574]\tvalid_0's rmse: 1782.89\n",
      "[1575]\tvalid_0's rmse: 1782.85\n",
      "[1576]\tvalid_0's rmse: 1782.81\n",
      "[1577]\tvalid_0's rmse: 1782.79\n",
      "[1578]\tvalid_0's rmse: 1782.75\n",
      "[1579]\tvalid_0's rmse: 1782.73\n",
      "[1580]\tvalid_0's rmse: 1782.7\n",
      "[1581]\tvalid_0's rmse: 1782.68\n",
      "[1582]\tvalid_0's rmse: 1782.61\n",
      "[1583]\tvalid_0's rmse: 1782.57\n",
      "[1584]\tvalid_0's rmse: 1782.59\n",
      "[1585]\tvalid_0's rmse: 1782.54\n",
      "[1586]\tvalid_0's rmse: 1782.52\n",
      "[1587]\tvalid_0's rmse: 1782.49\n",
      "[1588]\tvalid_0's rmse: 1782.47\n",
      "[1589]\tvalid_0's rmse: 1782.46\n",
      "[1590]\tvalid_0's rmse: 1782.45\n",
      "[1591]\tvalid_0's rmse: 1782.45\n",
      "[1592]\tvalid_0's rmse: 1782.45\n",
      "[1593]\tvalid_0's rmse: 1782.42\n",
      "[1594]\tvalid_0's rmse: 1782.34\n",
      "[1595]\tvalid_0's rmse: 1782.31\n",
      "[1596]\tvalid_0's rmse: 1782.31\n",
      "[1597]\tvalid_0's rmse: 1782.29\n",
      "[1598]\tvalid_0's rmse: 1782.29\n",
      "[1599]\tvalid_0's rmse: 1782.24\n",
      "[1600]\tvalid_0's rmse: 1782.21\n",
      "[1601]\tvalid_0's rmse: 1782.18\n",
      "[1602]\tvalid_0's rmse: 1782.11\n",
      "[1603]\tvalid_0's rmse: 1782.1\n",
      "[1604]\tvalid_0's rmse: 1782.16\n",
      "[1605]\tvalid_0's rmse: 1782.1\n",
      "[1606]\tvalid_0's rmse: 1782.04\n",
      "[1607]\tvalid_0's rmse: 1782.04\n",
      "[1608]\tvalid_0's rmse: 1781.96\n",
      "[1609]\tvalid_0's rmse: 1781.96\n",
      "[1610]\tvalid_0's rmse: 1781.95\n",
      "[1611]\tvalid_0's rmse: 1781.94\n",
      "[1612]\tvalid_0's rmse: 1781.9\n",
      "[1613]\tvalid_0's rmse: 1781.88\n",
      "[1614]\tvalid_0's rmse: 1781.86\n",
      "[1615]\tvalid_0's rmse: 1781.86\n",
      "[1616]\tvalid_0's rmse: 1781.79\n",
      "[1617]\tvalid_0's rmse: 1781.79\n",
      "[1618]\tvalid_0's rmse: 1781.77\n",
      "[1619]\tvalid_0's rmse: 1781.74\n",
      "[1620]\tvalid_0's rmse: 1781.72\n",
      "[1621]\tvalid_0's rmse: 1781.72\n",
      "[1622]\tvalid_0's rmse: 1781.71\n",
      "[1623]\tvalid_0's rmse: 1781.7\n",
      "[1624]\tvalid_0's rmse: 1781.66\n",
      "[1625]\tvalid_0's rmse: 1781.57\n",
      "[1626]\tvalid_0's rmse: 1781.6\n",
      "[1627]\tvalid_0's rmse: 1781.59\n",
      "[1628]\tvalid_0's rmse: 1781.58\n",
      "[1629]\tvalid_0's rmse: 1781.57\n",
      "[1630]\tvalid_0's rmse: 1781.51\n",
      "[1631]\tvalid_0's rmse: 1781.53\n",
      "[1632]\tvalid_0's rmse: 1781.51\n",
      "[1633]\tvalid_0's rmse: 1781.53\n",
      "[1634]\tvalid_0's rmse: 1781.54\n",
      "[1635]\tvalid_0's rmse: 1781.53\n",
      "[1636]\tvalid_0's rmse: 1781.51\n",
      "[1637]\tvalid_0's rmse: 1781.51\n",
      "[1638]\tvalid_0's rmse: 1781.51\n",
      "[1639]\tvalid_0's rmse: 1781.52\n",
      "[1640]\tvalid_0's rmse: 1781.51\n",
      "Early stopping, best iteration is:\n",
      "[1630]\tvalid_0's rmse: 1781.51\n"
     ]
    }
   ],
   "source": [
    "num_round = 10000\n",
    "bst = lgb.train(params, train_data, num_round, valid_sets=[ \n",
    "                test_data], early_stopping_rounds=10) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LGBMRegressor(metric='rmse')"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# LightGBM Regression\n",
    "lgbm_reg = LGBMRegressor(metric='rmse')\n",
    "lgbm_reg.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [],
   "source": [
    "Y_train = lgbm_reg.predict(X_train) \n",
    "Y_val = lgbm_reg.predict(X_val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training RMSE:  1860.581890013172\n",
      "Validation RMSE:  1896.2320020063732\n"
     ]
    }
   ],
   "source": [
    "# Calculate and print the Root Mean Squared Error (RMSE) for training and validation predictions. \n",
    "print(\"Training RMSE: \", np.sqrt(mse(Y_train, y_train))) \n",
    "print(\"Validation RMSE: \", np.sqrt(mse(Y_val, y_val))) "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Analisis Model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Kualitas Prediksi"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [],
   "source": [
    "models = [\n",
    "    ('Linear Regression', LinearRegression()),\n",
    "    ('Random Forest', RandomForestRegressor(n_estimators=20, random_state=42)),\n",
    "    ('Xgb Regression', XGBRegressor(n_estimators=20)),\n",
    "    ('Light GBM Regression', LGBMRegressor(metric='rmse'))\n",
    "    \n",
    "]\n",
    "\n",
    "evaluation_metrics = {'Model': [], 'RMSE': []}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [],
   "source": [
    "evaluation_metrics = {'Model': [], 'RMSE': [], 'train_time' : [], 'pred_time': []}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [],
   "source": [
    "for name, model in models:\n",
    "    start_train_time = time.time_ns()\n",
    "    model.fit(X_train, y_train)\n",
    "    end_train_time = time.time_ns()\n",
    "    \n",
    "    start_pred_time = time.time_ns()\n",
    "    y_pred = model.predict(X_val)\n",
    "    end_pred_time = time.time_ns()\n",
    "    \n",
    "    \n",
    "    RMSE = np.sqrt(mse(y_val, y_pred))\n",
    "    train_timer = start_train_time - end_train_time\n",
    "    pred_timer = start_pred_time - end_pred_time\n",
    "    \n",
    "    \n",
    "    \n",
    "    evaluation_metrics['Model'].append(name)\n",
    "    evaluation_metrics['RMSE'].append(RMSE)\n",
    "    evaluation_metrics['train_time'].append(train_timer)\n",
    "    evaluation_metrics['pred_time'].append(pred_timer)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "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>Model</th>\n",
       "      <th>RMSE</th>\n",
       "      <th>train_time</th>\n",
       "      <th>pred_time</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Linear Regression</td>\n",
       "      <td>3419.900580</td>\n",
       "      <td>-2186822053</td>\n",
       "      <td>-5063349</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Random Forest</td>\n",
       "      <td>1825.632383</td>\n",
       "      <td>-39728244886</td>\n",
       "      <td>-895595271</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Xgb Regression</td>\n",
       "      <td>1960.596259</td>\n",
       "      <td>-24304780514</td>\n",
       "      <td>-186805694</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Light GBM Regression</td>\n",
       "      <td>1896.232002</td>\n",
       "      <td>-6078309476</td>\n",
       "      <td>-822331089</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  Model         RMSE   train_time  pred_time\n",
       "0     Linear Regression  3419.900580  -2186822053   -5063349\n",
       "1         Random Forest  1825.632383 -39728244886 -895595271\n",
       "2        Xgb Regression  1960.596259 -24304780514 -186805694\n",
       "3  Light GBM Regression  1896.232002  -6078309476 -822331089"
      ]
     },
     "execution_count": 94,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "evaluation_metrics_df = pd.DataFrame(evaluation_metrics)\n",
    "evaluation_metrics_df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert-info\">\n",
    "    \n",
    "\n",
    "<b>Insight</b>\n",
    "\n",
    "- RMSE Terkecil ditemukan pada model Random Forest, Namun memerlukan waktu lebih lama dalam proses pelatihan dan prediksi yang berbanding terbalik dengan Model Linear Regression. Linear Regression Memiliki waktu lebih cepat namun memiliki nilai RMSE yang paling banyak.\n",
    "    \n",
    "- Model Light GBM Regression memiliki value paling optimal, dimana nilai RMSE tidak jauh berbeda dengan Random Forest dan memiliki waktu proses pelatihan dan prediksi lebih cepat."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.19"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": true,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {
    "height": "calc(100% - 180px)",
    "left": "10px",
    "top": "150px",
    "width": "416.098px"
   },
   "toc_section_display": true,
   "toc_window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}