P1-UsedVehiclePricePrediction2 (1).ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Project 1\n",
    "\n",
    "# Used Vehicle Price Prediction"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Introduction\n",
    "\n",
    "- 1.2 Million listings scraped from TrueCar.com - Price, Mileage, Make, Model dataset from Kaggle: [data](https://www.kaggle.com/jpayne/852k-used-car-listings)\n",
    "- Each observation represents the price of an used car"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Desarrollo de modelo para predicción de precios de autos\n",
    "\n",
    "Para el desarrollo de la predicción de precios de autos se probaron dos tipos de modelos ensamblados, boosting y random forest.\n",
    "\n",
    "Dado el tamaño de la muestra se redujo la base de datos a 20.000 registros, lo que generó parámetros óptimos para la estimación de los modelos."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "##Parameters\n",
    "size_sample=20000\n",
    "estimator_range = range(300, 1000, 50)\n",
    "max_depth_range = range(10, 50, 5)\n",
    "feature_range = range(2, 13, 1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Las librerías utilizadas para la estimación de los modelos fueron las siguientes."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "from sklearn.ensemble import AdaBoostRegressor\n",
    "from sklearn.ensemble import RandomForestRegressor\n",
    "from sklearn import preprocessing\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn import metrics\n",
    "from sklearn.externals import joblib\n",
    "import category_encoders as ce\n",
    "from sklearn.model_selection import cross_val_score\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = pd.read_csv('https://github.com/albahnsen/PracticalMachineLearningClass/raw/master/datasets/dataTrain_carListings.zip')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(500000, 6)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    500000.000000\n",
       "mean      21144.186304\n",
       "std       10753.259704\n",
       "min        5001.000000\n",
       "25%       13499.000000\n",
       "50%       18450.000000\n",
       "75%       26998.000000\n",
       "max       79999.000000\n",
       "Name: Price, dtype: float64"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.Price.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x274b3216dd8>"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "data.plot(kind='scatter', y='Price', x='Year')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x1c1b0eabc18>"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "data.plot(kind='scatter', y='Price', x='Mileage')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Price</th>\n",
       "      <th>Year</th>\n",
       "      <th>Mileage</th>\n",
       "      <th>State</th>\n",
       "      <th>Make</th>\n",
       "      <th>Model</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>21490</td>\n",
       "      <td>2014</td>\n",
       "      <td>31909</td>\n",
       "      <td>MD</td>\n",
       "      <td>Nissan</td>\n",
       "      <td>MuranoAWD</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>21250</td>\n",
       "      <td>2016</td>\n",
       "      <td>25741</td>\n",
       "      <td>KY</td>\n",
       "      <td>Chevrolet</td>\n",
       "      <td>CamaroCoupe</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>20925</td>\n",
       "      <td>2016</td>\n",
       "      <td>24633</td>\n",
       "      <td>SC</td>\n",
       "      <td>Hyundai</td>\n",
       "      <td>Santa</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>14500</td>\n",
       "      <td>2012</td>\n",
       "      <td>84026</td>\n",
       "      <td>OK</td>\n",
       "      <td>Jeep</td>\n",
       "      <td>Grand</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>32488</td>\n",
       "      <td>2013</td>\n",
       "      <td>22816</td>\n",
       "      <td>TN</td>\n",
       "      <td>Jeep</td>\n",
       "      <td>Wrangler</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Price  Year  Mileage State       Make        Model\n",
       "0  21490  2014    31909    MD     Nissan    MuranoAWD\n",
       "1  21250  2016    25741    KY  Chevrolet  CamaroCoupe\n",
       "2  20925  2016    24633    SC    Hyundai        Santa\n",
       "3  14500  2012    84026    OK       Jeep        Grand\n",
       "4  32488  2013    22816    TN       Jeep     Wrangler"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.State = data.State.str.lstrip()\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Teniendo en cuenta las características de los datos se decidió crear un índice ponderado de precio interestatal, que permite identificar si los precios por modelos en cada Estado se encuentran por encima o por deajo del promedio nacional. \n",
    "\n",
    "**Nota:** Este índice se desarrollo en Tableau y se exportó a un CSV para ser utilizado tanto en el entrenamiento del modelo, como en el web service.\n",
    "\n",
    "La variable **Year** se modificó para que sea la diferencia entre 2019 y el año de referencia del carro; así que en estos momentos esta variable nos explica cuantos años de antigüedad tiene el carro al día de hoy."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "def createFeatures(data):\n",
    "    data.State = data.State.str.lstrip()\n",
    "    data = data.reset_index()\n",
    "    stateRate = pd.read_csv(\"StatesRate.csv\")\n",
    "    data = data.set_index('State').join(stateRate.set_index('State')).reset_index()\n",
    "    data.Year = 2019-data.Year\n",
    "    data[\"MakeModel4\"] = data[\"Make\"] + data[\"Model\"].str[:4]\n",
    "    data=data.drop(['State'], axis=1)\n",
    "    return data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>State</th>\n",
       "      <th>RateState</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>FL</td>\n",
       "      <td>0.962969</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>MI</td>\n",
       "      <td>0.965131</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>OH</td>\n",
       "      <td>0.968848</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>OK</td>\n",
       "      <td>0.973919</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>GA</td>\n",
       "      <td>0.974889</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  State  RateState\n",
       "0    FL   0.962969\n",
       "1    MI   0.965131\n",
       "2    OH   0.968848\n",
       "3    OK   0.973919\n",
       "4    GA   0.974889"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "stateRate = pd.read_csv(\"StatesRate.csv\")\n",
    "stateRate.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Price</th>\n",
       "      <th>Year</th>\n",
       "      <th>Mileage</th>\n",
       "      <th>Make</th>\n",
       "      <th>Model</th>\n",
       "      <th>RateState</th>\n",
       "      <th>MakeModel4</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>index</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>757</th>\n",
       "      <td>6700</td>\n",
       "      <td>7</td>\n",
       "      <td>107728</td>\n",
       "      <td>Nissan</td>\n",
       "      <td>Sentra4dr</td>\n",
       "      <td>1.042688</td>\n",
       "      <td>NissanSent</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1174</th>\n",
       "      <td>14625</td>\n",
       "      <td>7</td>\n",
       "      <td>94715</td>\n",
       "      <td>Subaru</td>\n",
       "      <td>Outback2.5i</td>\n",
       "      <td>1.042688</td>\n",
       "      <td>SubaruOutb</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1495</th>\n",
       "      <td>17194</td>\n",
       "      <td>5</td>\n",
       "      <td>32977</td>\n",
       "      <td>Buick</td>\n",
       "      <td>RegalPremium</td>\n",
       "      <td>1.042688</td>\n",
       "      <td>BuickRega</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1521</th>\n",
       "      <td>9995</td>\n",
       "      <td>8</td>\n",
       "      <td>94837</td>\n",
       "      <td>Buick</td>\n",
       "      <td>Regal4dr</td>\n",
       "      <td>1.042688</td>\n",
       "      <td>BuickRega</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2194</th>\n",
       "      <td>9194</td>\n",
       "      <td>10</td>\n",
       "      <td>66501</td>\n",
       "      <td>Toyota</td>\n",
       "      <td>Corolla4dr</td>\n",
       "      <td>1.042688</td>\n",
       "      <td>ToyotaCoro</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       Price  Year  Mileage    Make         Model  RateState  MakeModel4\n",
       "index                                                                   \n",
       "757     6700     7   107728  Nissan     Sentra4dr   1.042688  NissanSent\n",
       "1174   14625     7    94715  Subaru   Outback2.5i   1.042688  SubaruOutb\n",
       "1495   17194     5    32977   Buick  RegalPremium   1.042688   BuickRega\n",
       "1521    9995     8    94837   Buick      Regal4dr   1.042688   BuickRega\n",
       "2194    9194    10    66501  Toyota    Corolla4dr   1.042688  ToyotaCoro"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data = createFeatures(data).set_index(\"index\")\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Las variables categóricas fueron transformadas con el codificador binario; el archivo **BinEncoder.pkl** fue utilizado para estimar las predicciones del modelo y para el correcto funcionamiento del web service."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Origninal Data:  (500000, 7)\n",
      "Y sample:  (20000,)\n",
      "X sample:  (20000, 30)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Make_0</th>\n",
       "      <th>Make_1</th>\n",
       "      <th>Make_2</th>\n",
       "      <th>Make_3</th>\n",
       "      <th>Make_4</th>\n",
       "      <th>Make_5</th>\n",
       "      <th>Make_6</th>\n",
       "      <th>Model_0</th>\n",
       "      <th>Model_1</th>\n",
       "      <th>Model_2</th>\n",
       "      <th>...</th>\n",
       "      <th>MakeModel4_2</th>\n",
       "      <th>MakeModel4_3</th>\n",
       "      <th>MakeModel4_4</th>\n",
       "      <th>MakeModel4_5</th>\n",
       "      <th>MakeModel4_6</th>\n",
       "      <th>MakeModel4_7</th>\n",
       "      <th>MakeModel4_8</th>\n",
       "      <th>Year</th>\n",
       "      <th>Mileage</th>\n",
       "      <th>RateState</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>index</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>28961</th>\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",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>10</td>\n",
       "      <td>57407</td>\n",
       "      <td>1.080224</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>205531</th>\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>1</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>10421</td>\n",
       "      <td>0.994189</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>308395</th>\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",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>8</td>\n",
       "      <td>60726</td>\n",
       "      <td>0.978508</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>117495</th>\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",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>41495</td>\n",
       "      <td>0.994189</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>308461</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</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>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>13504</td>\n",
       "      <td>0.978508</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 30 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        Make_0  Make_1  Make_2  Make_3  Make_4  Make_5  Make_6  Model_0  \\\n",
       "index                                                                     \n",
       "28961        0       0       0       0       1       0       0        0   \n",
       "205531       0       0       0       1       0       0       0        0   \n",
       "308395       0       0       0       1       0       0       0        0   \n",
       "117495       0       0       0       0       1       0       0        0   \n",
       "308461       0       0       0       0       1       1       0        0   \n",
       "\n",
       "        Model_1  Model_2    ...      MakeModel4_2  MakeModel4_3  MakeModel4_4  \\\n",
       "index                       ...                                                 \n",
       "28961         0        0    ...                 0             1             0   \n",
       "205531        0        1    ...                 1             0             0   \n",
       "308395        0        0    ...                 0             0             1   \n",
       "117495        0        0    ...                 0             1             0   \n",
       "308461        0        0    ...                 0             1             1   \n",
       "\n",
       "        MakeModel4_5  MakeModel4_6  MakeModel4_7  MakeModel4_8  Year  Mileage  \\\n",
       "index                                                                           \n",
       "28961              1             0             0             0    10    57407   \n",
       "205531             1             1             1             0     2    10421   \n",
       "308395             0             1             0             1     8    60726   \n",
       "117495             1             1             1             1     3    41495   \n",
       "308461             1             1             0             1     3    13504   \n",
       "\n",
       "        RateState  \n",
       "index              \n",
       "28961    1.080224  \n",
       "205531   0.994189  \n",
       "308395   0.978508  \n",
       "117495   0.994189  \n",
       "308461   0.978508  \n",
       "\n",
       "[5 rows x 30 columns]"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Create the binary encoder and save it\n",
    "BinEncoder = ce.BinaryEncoder().fit(data.drop(['Price'], axis=1))\n",
    "joblib.dump(BinEncoder, 'BinEncoder.pkl', compress=3)\n",
    "print(\"Origninal Data: \", data.shape)\n",
    "\n",
    "#Create sample that it will be use for find the hiperparameters, after select features X and Y\n",
    "data_sample = data.sample(n=size_sample, random_state=1)\n",
    "Y = data_sample.Price\n",
    "X = BinEncoder.transform(data_sample.drop(['Price'], axis=1))\n",
    "print(\"Y sample: \", Y.shape)\n",
    "print(\"X sample: \", X.shape)\n",
    "X.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# Split the data that is going to be use to find the random forest parameters\n",
    "size = 0.3\n",
    "X_train, X_validation, Y_train, Y_validation = train_test_split(X, Y,test_size=size).copy()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Se dividieron los datos en **Train (0.7)** y **Test (0,3)** para encontrar los hiper parámetros adecuados para el modelo."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Finding max parameters"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Finding max estimators"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "best_n_estimator:  950\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'NEG_RMSE')"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZsAAAESCAYAAAArJ3joAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAIABJREFUeJzt3Xl4VOXZx/HvnZ1A2MMm+xZ2AoTNFQGruKFYEKu4FMVq1WprXV6XVqW1atVa21pw1yoqIqIiUkVU6gIGCBAIYV8iEIIsAUJClvv945zoEAcygZmcmeT+XNe5mPOcZX6JgzfnnGeeR1QVY4wxJpSivA5gjDGm5rNiY4wxJuSs2BhjjAk5KzbGGGNCzoqNMcaYkLNiY4wxJuSs2PgQkRdEZKeIZAaw75MikuEua0Rkb3VkNMaYSCT2PZsficjpwAHgFVXtVYXjbgb6qeovQxbOGGMimF3Z+FDVL4Ddvm0i0klEPhKRxSKyQES6+Tn0MmBatYQ0xpgIFON1gAgwFfiVqq4VkcHAv4Dh5RtFpB3QAfjUo3zGGBP2rNgcg4jUA04GpotIeXN8hd3GA2+raml1ZjPGmEhixebYooC9qpp6jH3GA7+upjzGGBOR7JnNMahqPrBRRMYCiKNv+XYRSQEaAV97FNEYYyKCFRsfIjINp3CkiEiOiEwELgcmisgyYCUw2ueQy4A31Lr0GWPMMVnXZ2OMMSFnVzbGGGNCzjoIuJo2bart27f3OoYxxkSUxYsX71LV5Mr2s2Ljat++Penp6V7HMMaYiCIimwPZz26jGWOMCTkrNsYYY0LOio0xxpiQs2JjjDEm5KzYGGOMCTkrNsYYY0LOio0xxpiQs+/ZGGNMLaCq7C8qYV9BMfsOHbl0Sq7HoA6NQ/r+VmyMMSZCqCqFxWXsPXTYKRQFxewtLxo+RWSvbzEpcPbNLyyhtMz/WJhXDm1nxcYYY2qzJVv2cO/MTHbuLyL/UDGHS8uOum+UQIM6sT8uiXG0bZxIgzoxNKwT59Pu/NnQ/bNRYlzIfw4rNsYYE6b2FhzmpteWoMBZPZrRwKdglBcK3wJSLy6GqCip9LxesGJjjDFhSFW5a8YKdu4v4p0bT6ZP64ZeRzoh1hvNGGPC0OuLtvDRyh3ccU5KxBcasGJjjDFhJ3vHfh58fxWndWnKtad29DpOUHhWbERkrIisFJEyEUk7yj4pIpLhs+SLyK3utjd92jeJSIbb3l5EDvls+3d1/lzGGHMiCotLuXnaEpISYnh8XN+wfQZTVV4+s8kExgBTjraDqmYDqQAiEg18B8x0t11avp+IPA7s8zl0vaqmhiCzMcaE1OTZq1iTe4CXrhlIs6QEr+MEjWfFRlWzAEQCrtojcIrIERP1iHOCccDwoAY0xphq9lHmDv7zzRauO60Dw1KaeR0nqCLpmc14YJqf9tOAXFVd69PWQUSWisjnInLa0U4oIpNEJF1E0vPy8oKd1xhjArZt7yHunLGc3ic14Pdnd/M6TtCF9MpGRD4BWvjZdI+qzqrCeeKAC4G7/Wy+jCOL0Hagrap+LyIDgHdFpKeq5lc8UFWnAlMB0tLS/H+11hhjQqy0TLn1jQxKSsv4+2X9iIuJpOuAwIS02KjqyCCdahSwRFVzfRtFJAbnuc8An/csAorc14tFZD3QFUgPUhZjjAmqf3y6jkWbdvPEuL50aFrX6zghESnls+LVS7mRwGpVzSlvEJFktzMBItIR6AJsqJaUxhhTRd9u2s1T89Zwcb+TGNO/tddxQsbLrs8Xi0gOMBSYLSJz3fZWIvKhz36JwFnAO35O4+85zunAchFZBrwN/EpVd4fiZzDGmBOxr6CY30xbSpvGiTw4uqfXcULKy95oM3G7MVdo3wac67NeADQ5yjmu9tM2A5gRtKDGGBMCqspd7yxn5/4iZtxwMkkJsV5HCqlIuY1mjDE1yuuLtjAncwe/PzuFvm0ifziaylixMcaYarYm98fhaK47rWYMR1MZKzbGGFONCotLufn1pTVuOJrK2BQDxhhTjf40O4vs3P01bjiaytiVjTHGVJO5K3fw6jeba+RwNJWxYmOMMdVg295D3PF2zR2OpjJWbIwxJsRKy5Rb38yguAYPR1MZe2ZjjDEh9s/561i0cTePj625w9FUpvaVV2OMqUbpm3bzt0/WcFFqK8b0P8nrOJ6xYmOMMSGyr6CY37yRQetGiTx0Ua+qzN9V49htNGOMCYHy4Why8wt5uxYMR1MZu7IxxpgQmLZoK3Myd3D72Smk1oLhaCpjxcYYY4JsTe5+Hnh/Jad1acqkWjIcTWWs2BhjTBAVFpdyy7Sl1IuvXcPRVMae2RhjTBD9+cMsVu/Yz4u1bDiaytiVjTHGBMl/V+7gla83c+2pHTizlg1HUxkrNsYYEwTb9x3ijhnL6XVSfX5/TorXccKOFRtjjDlBpWXKrW9kcLikjL+P70d8TLTXkcKOPbMxxpgT9PSna1m4cTd/HduXjsn1vI4Tljy7shGRsSKyUkTKRCTtKPukiEiGz5IvIre62970ad8kIhk+x/URka/d868QEXtKZ4wJiTcWbeFvn6xlTL+TuKQWD0dTGS+vbDKBMcCUo+2gqtlAKoCIRAPfATPdbZeW7ycijwP73NcxwH+ACaq6TESaAMUh+hmMMbXYnBXb+b+ZKzijazJ/uaRPrR6OpjKeFRtVzQKq8h9nBLBeVTf7NopzgnHAcLfpZ8ByVV3mvs/3QQlsjDE+FqzN4zdvZNCvbSOeuaJ/rZw2oCoi6bczHpjmp/00IFdV17rrXQEVkbkiskRE7jjaCUVkkoiki0h6Xl5eCCIbY2qiJVv2cP2ri+mYXJcXrhpIYpw9/q5MSH9DIvIJ0MLPpntUdVYVzhMHXAjc7WfzZRxZhGKAU4GBQAEwT0QWq+q8igeq6lRgKkBaWpoGmscYU3tl79jPNS9+S9N68bzyy0E0SKzdA2wGKqTFRlVHBulUo4Alqprr2+g+nxkDDPBpzgE+V9Vd7j4fAv2BnxQbY4ypiq27C5jw/ELiY6L4z8TBNKtvfY8CFSm30SpevZQbCaxW1RyftrlAHxFJdIvRGcCqashojKnBdu4v5IrnF1JUUsarEwfTtkmi15Eiipddny8WkRxgKDBbROa67a3cq5Hy/RKBs4B3/JzmJ89xVHUP8ATwLZCBc0U0OzQ/hTGmNthXUMyVzy8ib38RL14zkJQWSV5Hijiiao8qwHlmk56e7nUMY0yYKThcwoTnF7E8Zy8vXD2Q07okex0prLjPxP1+V9JXpNxGM8aYane4pIxf/WcJS7fs4anx/azQnADrr2eMMX6Ulim/fSuDL9bk8ZcxvTm3d0uvI0U0u7IxxpgKVJX7ZmXywfLt3D2qG+MHtfU6UsSzYmOMMRX89b/ZvL5wCzcM68T1Z3TyOk6NYMXGGGN8TP1iPf+cv57LBrXljrNtXppgsWJjjDGut77dyp8/XM15vVsy+aJeNrBmEFmxMcYY4KPM7dz1znJO69KUJy9NJTrKCk0wWbExxtR6/1u7i1umZZDapiFTJgywEZxDwH6jxphabemWPUx6NZ2OyXV58epBNoJziFixMcbUWmty93PNSzaCc3WwYmOMqZXKR3COi7YRnKuDXS8aY2qdnfsLmfD8QgqLy3jr+qE2gnM1sCsbY0ytsu+QM4Jzbn4RL1xtIzhXFys2xpha49DhUia+9C3r8w4wZcIABrRr5HWkWsOKjTGmVjhcUsYNry1miTuC8+ldbQTn6mTPbIwxNV5pmfK76cv4LDuPh20EZ0/YlY0xpkYrK1PumbmC95dt485zunGZjeDsCSs2xpgaS1V54P2VvPHtVm4e3pkbhtkIzl6xYmOMqZFUlb/MWc3LX2/m2lM78NuzunodqVbzrNiIyFgRWSkiZSLid/5qEUkRkQyfJV9EbnW3venTvklEMtz2yyscUyYiqdX5sxljvPfUvLVM+WIDVwxpyz3ndbcRnD3mZQeBTGAMMOVoO6hqNpAKICLRwHfATHfbpeX7icjjwD63/TXgNbe9NzBLVTNC8yMYY8LRvz9fz98+WcvPB7TmwQttqoBw4FmxUdUsoCofghHAelXd7NsozgnGAcP9HHMZMO0EYhpjIsxLX27kL3NWc0HfVjxySR+ibKqAsBBJz2zG479wnAbkqupaP9suPcoxAIjIJBFJF5H0vLy8IMU0xnjljUVb+OP7q/hZj+Y8Ma6vzUkTRkJabETkExHJ9LOMruJ54oALgel+Nvu9ehGRwUCBqmYe7byqOlVV01Q1LTnZvuBlTCR7d+l33D1zBWd0TebpX/QjNjqS/i1d84X0NpqqjgzSqUYBS1Q117dRRGJwnvsM8HPM0a6EjDE1zJwV2/nd9GUM6dCEKRMGEB8T7XUkU0GkjCBwtGcvI4HVqprj2ygiUcBY4PRqyGaM8dCnq3O5edpSUts05Lmr0kiItUITjrzs+nyxiOQAQ4HZIjLXbW8lIh/67JcInAW84+c0R7t6OR3IUdUNwU9ujAkX/1u7i1/9ZwndW9bnxWsGUjc+Uv79XPuIqnqdISykpaVpenq61zGMMQFatHE3V76wkPZN6jLtuiE0qhvndaRaSUQWq6rf70r6sidoxpiIs3TLHq55cREnNazDqxMHW6GJAFZsjDERZeW2fVz1wiKa1IvntWuHkJwU73UkEwArNsaYiLE2dz8Tnl9EvfgYXrt2MC0aJHgdyQTIio0xJiJs3HWQXzy3kOgo4bXrhtCmcaLXkUwVVFpsROQtn9ePVNj231CEMsYYX1t3F3D5s99QWqa8fu1gOjSt63UkU0WBXNl08Xl9VoVt9rV7Y0xI7dhXyOXPLeRAUQmvThxEl+ZJXkcyxyGQTunH6htt/aaNMSGTt7+IXzz3DbsPHuY/1w6mZ6sGXkcyxymQYpMoIv1wroLquK/FXeqEMpwxpvbac/AwE55fyLa9h3jll4NJbdPQ60jmBARSbHYAT/h5Xb5ujDFBlV9YzJUvLGLDroO8cNVABnVo7HUkc4IqLTaqOqwachhjDAAHi0q45sVvWb0jnykTBnBql6ZeRzJBEEhvtIEi0sJn/UoRmSUifxcR++eGMSZoCotLmfjytyzdsoe/j+/H8G7NvY5kgiSQ3mhTgMMAInI68BfgFZxpmKeGLpoxpjYpKinl+lcXs3Djbp4Yl8qo3i29jmSCKJBnNtGqutt9fSkwVVVnADNEJCN00YwxtcnkD7L4fE0ej1zSm4v6neR1HBNkgVzZRLuTlAGMAD712WbjeRtjTtjqHfm8tnAzV5/cnksHtvU6jgmBQIrFNOBzEdkFHAIWAIhIZ5xbacYYc9xUlT/NziIpIZZbR3ap/AATkQLpjfYnEZkHtAT+qz9OgBMF3BzKcMaYmu+z7DwWrN3F/ef3oGGiTRVQU1VabNweZ2vcJV5Eysfz3uUuxhhzXIpLy5g8exUdmtbliiHtvI5jQiiQ22i7gBygxF0Xn20KdAx2KGNM7TBt0RbW5x3k2SvTiIuxQehrskCKzdPAMOBLnOc3/1ObS9oYc4L2FRTz5MdrOLlTE0Z2b+Z1HBNilf5TQlV/A6QC04EJwFIReVREOpzIG4vIWBFZKSJlIuJ3/moRSRGRDJ8lX0Rudbe96dO+qbwbtojEisjLIrJCRLJE5O4TyWmMCY2nP13L3kPF3HNed0Sk8gNMRAuo67J7JTNfRJYC44GHgLXAsyfw3pnAGJwvjR7tfbNxCh0iEg18B8x0t11avp+IPM6PPePGAvGq2ltEEoFVIjJNVTedQFZjTBBt2nWQl7/exLgBbWwk51oikA4CdYHROF/oTAbeAfqr6tYTeWNVzXLPH+ghI4D1qrq5Qj4BxgHDy08N1HW/G1QHZ/SD/BPJaowJrofnZBEbHcXvzu7qdRRTTQK5stmJcxUzDViH8z/zgSIyEEBV3wldvCOMdzNUdBqQq6pr3fW3cYrjdiARuM1nBIQjiMgkYBJA27b2RTJjqsPX679n7spcbv9ZV5olJXgdx1STQIrNdJwC081dfCnOlY5fIvIJ0MLPpntUdVagIUUkDrgQ8Pf85TKOLEKDgFKgFdAIWCAin6jqhooHqupU3PHd0tLSrNODMSFWVqZMnr2KkxrW4drTrCNrbRLIlzqvPt6Tq+rI4z22glHAElXN9W10b5WNAQb4NP8C+EhVi4GdIvIlkAb8pNgYY6rXjCU5rNyWz1PjU0mIjfY6jqlGAXVsF5FoEWnqsx4nIpNEJCt00Y5Q8eql3Ehgtarm+LRtAYaLoy4wBFhdDRmNMcdwsKiEx+Zmk9qmIRf2beV1HFPNApnPZjywG1guIp+LyJk4VwmjgMuP941F5GIRyQGGArNFZK7b3kpEPvTZLxE4C/+36/w9x/knUA+nt9u3wIuquvx4cxpjgmPK5+vZub+I+87vYV2da6FAntncCwxQ1XUi0h/4GhivqjNP5I3d439yDlXdBpzrs14ANDnKOa7203YAp/uzMSZMbNt7iKkLNnBB31YMaNfI6zjGA4HcRjusqusAVHUJsPFEC40xpnZ5bG42ZQp3npPidRTjkUCubJqJyG991uv5rqvqE8GPZYypKTK27mXm0u+4cVgnWjdK9DqO8UggxeZZIOkY68YY45eqMvmDVTStF8eNZ3b2Oo7xUCBdnx8I5EQicreqPnzikYwxNcWHK3aQvnkPD4/pTb14m9i3NgvmmN72UN4Y84PC4lIenpNFtxZJjEtr43Uc47FgFhvry2iM+cFLX20iZ88h7j2vB9FR9r+H2i6YxcaGezHGALDrQBH/+HQdI7o149QuTSs/wNR4dmVjjAm6Jz5eQ2FxKf93Xnevo5gwEcxiMz2I5zLGRKjVO/J5Y9EWrhjSjk7J9byOY8JEIPPZ9AQ6qep77vqTQPlsR/9wv+iJqv45ZCmNMRFBVfnT7CySEmL5zYguXscxYSSQK5u/ALt81s8GZgPzgftDEcoYE5k+y85jwdpd3DKiC43qxnkdx4SRQDq+t1TVr3zW81V1BoCIXB+aWMaYSFNcWsbk2avo0LQuE4a08zqOCTOBXNkcMVqAqg7xWW0W3DjGmEg1bdEW1ucd5O5R3YiLCebjYFMTBPKJ2CYigys2isgQYFvwIxljIs2+gmKe/HgNQzs24awezb2OY8JQILfR7gTeFJGXgCVu2wDgKuDSEOUyxkSQpz9dy95Dxdx7fnebq8b4VemVjaouAgYD0cDV7hIFDHG3GWNqsU27DvLy15sYO6A1PVs1qHR/UzsF0vW5vqruxE/PMxFpq6pbQpLMGBMRHp6TRWx0FLf/zOaqMUcXyDObz8pfiMi8CtveDWoaY0xE+Xr998xdmcuNwzrRrH6C13FMGAvkmY3vDdjGx9hmqqisTPn+4GF27i9kZ34RO/cXkptfRG5+ITv3F7Ezv5C8/UVcPqQdv7a5QEyYKStTJs9eRasGCVx7Wkev45gwF0ix0aO89rceMBEZC/wR6A4MUtV0P/ukAG/6NHUE7lfVv4nIm0D5dXtDYK+qpopIHDAFSAPKgN+o6mfHm/N4VCwi5cXDt4jk5hex60ARJWU//RU2Soylef0EkpPiUeDJj9cwsntzUlrYnHUmfMxYksPKbfk8NT6VhNhor+OYMFeVaaGFI6eIFiD5BN47ExiDUxj8UtVsIBVARKKB74CZ7rYfesKJyOPAPnf1Ond7bxFpBswRkYGqWnYCWY9q1bZ8Xlu4mdz8IvL2B15EujRPonn9eJolJdC8fjzJP/wZT3zMj39xdx88zPDHP+O+dzN58/oh1tPHhIWDRSU8Njeb1DYNubBvK6/jmAhQ1WmhK04J/dzxvrGqZgFV+Z/nCGC9qm72bRTnBOOA4W5TD2Ce+x47RWQvzlVOSHrO7TpQxIcrttMsKYFm9Z0i0iwpnub1j11EAtW4bhx3j+rGnTNW8PbiHMbaJFQmDEz5fD079xfxzBX97R9AJiBBmxa6GowHpvlpPw3IVdW17voyYLSIvAG0wflOUBv8FBsRmQRMAmjbtu1xhTq9azJL7//ZcR0bqLED2vBWeg4Pz1nNyO7Nbcwp46ltew8xdcEGzu/TkgHtKj7GNca/QLo+H2uwTVXVh45x7CdACz+b7lHVWQHkKz9PHHAhcLefzZdxZBF6Aec5UDqwGfgKKPF3XlWdCkwFSEtLC9vJ36KihMkX9eL8p//Ho3NX8/CYPl5HMrXYY3OzKVO485xuXkcxESSQ22gH/bTVBSYCTYCjFhtVHXmcuSoaBSxR1VzfRhGJwXnuM8DnPUuA23z2+QpYS4Tr3rI+vzylPc8u2MjPB7RhQLtGXkcytVDG1r3MXPodNwzrRJvGiV7HMREkkBEEHi9fcK4C6gDXAG/g9A6rDhWvXsqNBFarak55g4gkikhd9/VZQImqrqqemKF168iutGyQwL3vZlJSGpL+DsYclaoy+YNVNK0Xx43DOnkdx0SYgIZmFZHGIjIZWI5zNdRfVe90RxY4LiJysYjkAEOB2SIy121vJSIf+uyXCJwFvOPnNP6e4zQDlohIFs64bhOON2O4qRsfwx8u6EHW9nxe+mqT13FMLTNz6Xekb97D7T9LISkh1us4JsIE8szmMZxbVVOB3qp6IBhvrKozcbsxV2jfBpzrs16Ac7vO3zmu9tO2iR+/f1PjnN2zBWemJPPkx2s4r09LWjao43UkUwvsO1TMnz/Mom+bhoyzHpHmOARyZfM7oBVwL850A/nusl9E8kMbz1QkIjxwYS9KypSHPqgRdwdNBHjy4zV8f/Awk0f3IirKujqbqgvkmU2UqtZR1SRVre+zJKlq/eoIaY7UtkkiNw/vzIcrdvBZ9nHfyTQmICu37eOVrzdxxeB29G5tozqb42PT6UWo607vSMfkutw/ayWFxaVexzE1VFmZct+7mTRKjLNRnc0JsWIToeJjopk8uhdbdhfwr/nrvI5jaqi3l+SwZMte7hrVjQaJ1inAHD8rNhHs5M5NuSi1Ff/+fAMb8oLSb8OYH+wtOMxf5qxmQLtGXNK/tddxTISzYhPh7jmvB/GxUdw3KxPVsB0EwUSgx+Zms7fgMA9ZpwATBFZsIlxyUjx3nJ3Cl+u+571l27yOY2qI5Tl7eX3RFq46uT09Wlk/IHPirNjUAL8Y3I4+rRsweXYW+YXFXscxEa7U7RTQtF48t53V1es4poawYlMDREcJf7qoN98fKOLxudlexzER7s1vt7IsZx/3nNud+jZSgAkSKzY1RO/WDZgwpB2vfrOZFTn7Kj/AGD92HzzMo3NXM7hDY0an2qRoJnis2NQgvzs7hSb14rnn3RWU+pkptDqt23mAa15cxPzV9qXTSPLInNUcKCzhoYt62aRoJqis2NQg9RNiufe87izP2cfrCzdXfkCIfLB8G6P/8T/mZ+fx69eXsHKbXWlFgiVb9vBm+lZ+eWoHujZPqvwAY6rAik0Nc2HfVpzSuQmPzs1m5/7Can3v4tIyHnx/FTe9vpSUFkm8d9MpNKgTy3Uvp1d7FlM15Z0CmteP55YRXbyOY2ogKzY1jIjw0OheFBWX8efZWdX2vjv2FTJ+6je88OVGrjmlPW9MGkqf1g159so09hQUc/2ri21YnTD22sLNrNyWz33n96BefCBzKhpTNVZsaqCOyfX41RkdeTdjG1+t2xXy9/tq3S7Of3oBWdvzefqyfvzhgp7ExTgfrV4nNeDJS/uydMte7pqx3L54Goby9hfx2NxsTu3clPN6t/Q6jqmhrNjUUDee2Zm2jRO5d1YmRSWhuaIoK1P+9dk6rnh+IQ0T43jvplO4oO9PezCd06slt/+sK+9mbONfn60PSRZz/B6ek0VhcSkPjO5pnQJMyFixqaESYqN5cHRPNuQd5NkvNgT9/PsOFTPp1cU8+lE25/Zuyaxfn0LnZkd/qPzrMzszOrUVj83N5qPM7UHPY47Poo27eWfJd1x3Wkc6JdfzOo6pwazY1GDDUppxbu8WPP3pOrZ8XxC0867cto8Lnv4fn2Xv5I8X9ODpy/pRt5L7/CLCI5f0IbVNQ257cxmZ31kPNa+VlJZx/6xMTmpYh5uGd/Y6jqnhrNjUcPed34OYKOEP7wVnoM630rcy5l9fcbikjDevH8LVp3QI+NZLQmw0U68cQKPEWK57xXqoee3lrzezesd+7ju/B4lx1inAhJZnxUZExorIShEpE5G0o+yTIiIZPku+iNzqbksVkW/c9nQRGeS2i4j8XUTWichyEelfnT9XuGnZoA63ndWV+dl5zF2547jPU1hcyp1vL+eOt5czoF0jPrjlVAa0a1zl8zRLSuDZq9LYW1DMpFesh5pXcvMLefLjNZzRNZmzezb3Oo6pBby8sskExgBfHG0HVc1W1VRVTQUGAAXATHfzo8AD7rb73XWAUUAXd5kEPBOa+JHj6pPb061FEg+8v4qDRSVVPn7L9wVc8sxXvJm+lZvO7MyrEwfTtF78cefp2aoBT16aSsbWvdxpPdQ88ecPszhcWsYDF1qnAFM9PCs2qpqlqlUZNXIEsF5Vy78ar0D52OcNgPLx9UcDr6jjG6ChiNTq/pwx0VH86eJebN9XyN8+WVOlY+dl5XL+0wvYuruA569K4/azU4gOwtwm5/Rqwe/PTmFWxjb+aTONVquv1u9iVsY2fnVGJ9o3ret1HFNLRNKN2vHANJ/1W4G5IvJXnKJ5stt+ErDVZ78ct+0nXaBEZBLO1Q9t27YNQeTwMaBdY8YPbMMLX25iTP/WdG957DlKSsuUJz7O5p/z19OzVX2euXwAbZskBjXTjcM6sW7nAf763zV0blaPc3rV6n8TVIvDJWXcP2slbRrX4cZhnbyOY2qRkF7ZiMgnIpLpZxldxfPEARcC032abwBuU9U2wG3A8+W7+zmF3/s0qjpVVdNUNS05ObkqkSLSned0o0GdWO59N5OyYwzUuetAEVe+sJB/zl/PpWltmHHDyUEvNOD0UHt4TG/6tbUeatXlxS83sm7nAf54QU8SYqO9jmNqkZAWG1Udqaq9/CyzqniqUcASVc31absKeMd9PR0Y5L7OAdr47NeaH2+x1WqN6sZx16huLN68h+mLt/rdZ/HmPZz/9/+RvmkPj17Sh0d+3iek/1OODQrxAAAUAElEQVRKiI1m6oQ0GiXGcu3L6ezMtx5qobJ93yGemreWkd2bMaK7dQow1StSuj5fxpG30MApIGe4r4cDa93X7wFXur3ShgD7VNW+Rej6ef/WDGzfiIfnrGb3wcM/tKsqL365kUunfE1cTBQzbjiZcQPbHONMwZOcFM9zVw0kv7CY62wMtZCZ/EEWpWXKHy7o6XUUUwt52fX5YhHJAYYCs0VkrtveSkQ+9NkvETiLH69iyl0HPC4iy4A/4z57AT4ENgDrgGeBG0P6g0SYqChh8kW9OVBYwl/mOAN1Hiwq4eZpS3ng/VUMS0nm/ZtOpddJDao1V49W9Xny0lSW5+zl929bD7Vg+2JNHrNXbOemMzvTpnHwb4kaUxnPOgio6kx+7Mbs274NONdnvQBo4me//+F0h67YrsCvgxq2hklpkcTEUzsw5YsNpLVrzNQFG9iQd4A7zknhV6d3IioIvc2Ox9k9nR5qj36UTddm9bjZhroPiqKSUv7w3kraN0nkutM7eh3H1FKR1BvNBNEtI7rw/rJt3DFjOU3qxvGfiYM5uXNTr2NxwxmdWJd7gMc/dnqojbJRiE/Ycws2snHXQV66ZqB1CjCesWJTS9WNj+GJS1N5Y9EW7hrVnRYNEryOBDg91P48pjebvj/IbW9l0KZxYrXf0qtJtu4u4OlP13JOzxYMS2nmdRxTi0VKBwETAkM6NuFv4/uFTaEplxAbzZQJaTSpG8+1L6eTaz3UjttDH6xCEO6/oIfXUUwtZ8XGhCWnh1oa+YXFTHol3XqoHYf5q3fy31W53DKiC60a1vE6jqnlrNiYsNW9ZX2eGt+P5d/t4/bpy6yHWhUUFjudAjol12XiqR28jmOMFRsT3s7q0Zw7z+nGB8u38/d5NoZaoP79+Xq27C7gwdG9fpii2xgvWQcBE/auP70ja3L38+QnTg+18/pYD7Vj2fz9Qf712XrO79OSU8Kgh6ExYFc2JgKUj6GW1q4Rv5uewYocG0PtaFSVP763ktgo4d7zrFOACR9WbExEiI+J5t8TBjg91F751nqoHcXclTuYn53HbWd1DbtehqZ2s2JjIkbTek4PtQOFJVz3SjqHDlsPNV+7DhRxz8xMeraqz1Unt/c6jjFHsGJjIkp5D7UV3+3jd9MzKD3GVAm1iapyz8wV7C8s4YlxqcRG219tE17sE2kizsgezbnn3O58uGIH976baV2igXeWfMfclbncfnZXUlokeR3HmJ+w3mgmIl17Wkf2FBzmn/PXUz8hhrtGdUPEmwFEvfbd3kP88b2VDGrfmImn2kCbJjxZsTER6/afpZB/qIQpX2ygfp1Yfn1mZ68jVbuyMuWOt5dRqspfx/Yl2qMRu42pjBUbE7FEhAcu7MmBohIem5tN/YQYJgxt73WsavXK15v4ct33PDymd0im7jYmWKzYmIgWFSU8+vM+7C8s5r5ZK0lKiOWifid5HatarM87wMNzVnNmSjLjq2lWVWOOl3UQMBEvNjqKf/yiP0M7NuF305fxyapcryOFXElpGb99axl14qJ55JI+tfZ5lYkcVmxMjZAQG82zV6XRq1V9bnx9CV+t3+V1pJB65rP1LNu6l8kX9aJZffvypgl/nhUbERkrIitFpExE0o6yT4qIZPgs+SJyq7stVUS+cdvTRWSQ295NRL4WkSIRub06fybjrXrxMbx0zSBn+uOX08nYutfrSCGR+d0+npq3lgv7tuL8Pq28jmNMQLy8sskExgBfHG0HVc1W1VRVTQUGAAXATHfzo8AD7rb73XWA3cAtwF9DFdyEr0Z143h14mCa1Ivn6hcXkb1jv9eRgqqwuJTfvpVB47pxPDi6p9dxjAmYZ8VGVbNUNbsKh4wA1qvq5vJTAPXd1w2Abe55d6rqt0Bx0MKaiNK8fgKvXTuY+JgoJjy/kC3fF3gdKWie+HgNa3IP8OjP+9AwMc7rOMYELJKe2YwHpvms3wo8JiJbca5i7vYklQlLbRon8urEwRwuLePy57+pEQN3LtzwPc8u2MDlg9syLKWZ13GMqZKQFhsR+UREMv0so6t4njjgQmC6T/MNwG2q2ga4DXj+OPJNcp/3pOfl5VX1cBPmujZP4uVrBrH7wGGueG4hew4e9jrScTtQVMLtby+jTaNE/u/c7l7HMabKQlpsVHWkqvbys8yq4qlGAUtU1bdP61XAO+7r6cCg48g3VVXTVDUtOTm5qoebCNC3TUOeu2ogm3cXcNWLi9hfGJl3V/80exU5ew7xxLi+1I23r8eZyBMpt9Eu48hbaOA8oznDfT0cWFutiUzEGNqpCc9c3p9V2/K59uV0Cosja2qC+at3Mm3RVq4/vRNp7Rt7HceY4+Jl1+eLRSQHGArMFpG5bnsrEfnQZ79E4Cx+vIopdx3wuIgsA/4MTHL3b+Ge97fAvSKSIyL1MbXaiO7NeXxcXxZt2s2vX1tCcWmZ15ECsufgYe6YsZxuLZK47awuXscx5rh5dj2uqjP5sRuzb/s24Fyf9QKgiZ/9/ofTHbpi+w6gdVDDmhphdOpJ7C8s4d53M7l9+jKeHJdKVJgPXHnvrEz2FhzmpWsGEh8T7XUcY46b3fw1tcoVQ9qRX1jMox9lk5QQw0Oje4XtUC/vLdvG7OXb+f3ZKfRs1cDrOMacECs2pta5cVhn8g+V8O/P11M/IZY7zunmdaSf2LGvkPvezaRf24Zcf7rNUWMinxUbUyvdeU4K+YXF/Ouz9SQlxHLDsE5eR/qBqnLnjOUcLinjiXGpxNgUz6YGsGJjaiUR4aHRvThQWMIjH62mfp0YLh/czutYALy+aAufr8njwdE96dC0rtdxjAkKKzam1oqOEh4f15cDRU6ngXrxMYxO9XYunE27DjL5gyxO69KUK8Kk+BkTDHZ9bmq12Ogo/nV5fwa2b8zv3lrGp6u9mwuntEy5ffoyYqKdCeHCvaecMVVhxcbUegmx0Tx/VRrdW9bnhv8s4ZsN33uS49kFG0jfvIcHR/ekZYM6nmQwJlSs2BgDJCXE8vIvB9GmcSLXvpzO8pzqnQsna3s+T/x3DaN6teAij2/lGRMKVmyMcTWuG8d/Jg6mYWIsV76wiCc/XsPynL2UlWlI3/dwiTPFc/06sUy+KHy/92PMibAOAsb4aNHAmQvn928v5++fruWpeWtplhTPiO7NGNGtOad0bkqduOB+k/+peWvI2p7Pc1em0aRefFDPbUy4sGJjTAXtmtTlreuHsvvgYeav3sm81bm8v2w70xZtJT4milM6N/2h+LRokHBC77V48x6e+Ww949JaM7JH8yD9BMaEH1EN7S2CSJGWlqbp6elexzBh6nBJGYs27uaTrFzmrc5l6+5DAPRsVZ8R3ZszsnszerVqUKUeZAWHSzj3qQUUlyof3XoaSQmxoYpvTMiIyGJVTat0Pys2Dis2JlCqytqdB5iXtZN5Wbks2bKHMoVmSfEM79aMEd2bc2oAt9vun5XJK19vZtp1Qxja6SdjzRoTEQItNnYbzZgqEhG6Nk+ia/MkbhjW6YfbbZ+u3skHy7fzxrdH3m4b3q3ZT7oyf7Emj1e+3szEUztYoTG1gl3ZuOzKxgRDILfb2jZO5Jy/LaBeQgwf3HwqCbE2dYCJXHYbrYqs2JhgU1XW7TzAJxVut8XHRFFSpsy88WT6tG7odUxjTojdRjPGYyJCl+ZJdPG53fZZ9k7mZ+cxpGNjKzSmVrFiY0w1aVw3jjH9WzOmv00ka2ofG0HAGGNMyHlWbERkrIisFJEyEfF7v09EUkQkw2fJF5Fb3W2pIvKN254uIoPc9stFZLm7fCUifavz5zLGGPNTXt5GywTGAFOOtoOqZgOpACISDXwHzHQ3Pwo8oKpzRORcd30YsBE4Q1X3iMgoYCowOFQ/hDHGmMp5VmxUNQuoyqCDI4D1qrq5/BRAffd1A2Cbe96vfI75BrAb5MYY47FI6iAwHpjms34rMFdE/opzO/BkP8dMBOZUQzZjjDHHENJnNiLyiYhk+llGV/E8ccCFwHSf5huA21S1DXAb8HyFY87EKTZ3HuO8k9znPel5eXlViWSMMaYKQnplo6ojg3SqUcASVfWds/cq4Dfu6+nAc+UbRKSPuz5KVY867aKqTsV5pkNaWpp9u9UYY0IkUro+X8aRt9DAeUZzhvt6OLAWQETaAu8AE1R1TbUlNMYYc1SeDVcjIhcDTwPJwF4gQ1XPFpFWwHOqeq67XyKwFeioqvt8jj8VeArn6qwQuFFVF4vIc8AlQHlHgpKAxu0RyfM5pqqaAruO81ivWXZvWHZvRGr2cM7dTlWTK9vJxkYLAhFJD6SghSPL7g3L7o1IzR6puX1Fym00Y4wxEcyKjTHGmJCzYhMcU70OcAIsuzcsuzciNXuk5v6BPbMxxhgTcnZlY4wxJuSs2BhjjAk5KzYBEJEEEVkkIsvcaREecNs7iMhCEVkrIm+6w+ogIvHu+jp3e3uP80eLyFIR+SDCcm8SkRXl00i4bY1F5GM3+8ci0shtFxH5u5t9uYj09zh7QxF5W0RWi0iWiAyNhOxHm9YjErK7eW5z/45misg09+9upHzef+PmXik/TqUSEb/3QFixCUwRMFxV++JMeXCOiAwBHgGeVNUuwB6csdhw/9yjqp2BJ939vPQbIMtnPVJyA5ypqqk+3zG4C5jnZp/nroMzpFEXd5kEPFPtSY/0FPCRqnYD+uL8/sM+u6pmu7/vVGAAUIAzrUfYZxeRk4BbgDRV7QVE4wzgG/afdxHpBVwHDML5vJwvIl2IgN97wFTVliosQCKwBGeOnF1AjNs+FJjrvp4LDHVfx7j7iUd5W+N8SIcDHwASCbndDJuAphXasoGW7uuWQLb7egpwmb/9PMhdH2deJanQHvbZK+T9GfBlpGQHTsIZbaSx+/n9ADg7Ej7vwFickVPK1+8D7oiE33ugi13ZBMi9FZUB7AQ+BtYDe1W1xN0lB+fDDj9+6HG37wOaVG/iH/wN50Nb5q43ITJygzNn0X9FZLGITHLbmqvqdgD3z2Zu+w/ZXb4/V3XrCOQBL7q3L58TkbpERnZfvtN6hH12Vf0O+CuwBdiO8/ldTGR83jOB00WkiThDdJ0LtCECfu+BsmITIFUtVefWQmucS93u/nZz//Q3I1y19zEXkfOBnaq62LfZz65hldvHKaraH+eWwa9F5PRj7BtO2WOA/sAzqtoPOMiPtz/8CafswFGn9fC7q582T7K7zzNGAx2AVkBdnM9ORWH3eVdnMslHcP4h+xGwDCg5xiFhkz1QVmyqSFX3Ap8BQ4CGIlI+TUNr3NlCcf6V0QbA3d4A2F29SQE4BbhQRDYBb+DcSvsb4Z8bAFUtn311J85zg0FAroi0dDO2xLnSBJ/sLt+fq7rlADmqutBdfxun+ERC9nIVp/WIhOwjgY2qmqeqxTijv59M5Hzen1fV/qp6uptjLZHxew+IFZsAiEiyiDR0X9fB+VBnAfOBn7u7XQXMcl+/567jbv9U3Rur1UlV71bV1qraHueWyKeqejlhnhtAROqKSFL5a5znB5kVMlbMfqXbS2cIsK/89kN1U9UdwFYRSXGbRgCriIDsPipO6xEJ2bcAQ0QkUUSEH3/vYf95BxCRZu6fbYExOL//SPi9B8brh0aRsAB9gKXAcpz/4d3vtncEFgHrcG43xLvtCe76Ond7xzD4GYYBH0RKbjfjMndZCdzjtjfB6fCw1v2zsdsuwD9xnqWtwOmR5OXvOxVIdz8z7wKNIih7IvA90MCnLVKyPwCsdv+evgrER8Ln3c2zAKc4LgNGRNLvPZDFhqsxxhgTcnYbzRhjTMhZsTHGGBNyVmyMMcaEnBUbY4wxIWfFxhhjTMhZsTHGGBNyVmyM8ZCIpIrIuT7rF4rIsYa2qcq5b3XH2TLGc/Y9G2M8JCJX43wh76YQnHuTe+5dVTgmWlVLg53FGLuyMSYAItJenEnQnnUnt/qvO3SRv307ichH7mjVC0Skm9s+1p0ca5mIfOEOdvkgcKk4E5VdKiJXi8g/3P1fEpFnRGS+iGwQkTNE5AU3x0s+7/eMiKTLkRP73YIzGOV8EZnvtl0mzmR0mSLyiM/xB0TkQRFZCAwVkb+IyCp3Uq6/huY3amodr4cwsMWWSFiA9jij8Ka6628BVxxl33lAF/f1YJwxt8AZVuQk93VD98+rgX/4HPvDOvASzgCqgjOacT7QG+cfiYt9spQPYRKNM0hsH3d9E+58QDiFZwuQjDMq9afARe42BcaVnwtnbhTxzWmLLSe62JWNMYHbqKoZ7uvFOAXoCCJSD2ek4enu/EdTcCa9AvgSeElErsMpDIF4X1UVp1DlquoKVS3DGS+u/P3HicgSnPH7egI9/JxnIPCZOiMilwCvAeVTNpQCM9zX+UAh8JyIjMGZqdOYExZT+S7GGFeRz+tSwN9ttCicybpSK25Q1V+JyGDgPCBDRH6yzzHes6zC+5cBMSLSAbgdGKiqe9zbawl+zuNv/pNyheo+p1HVEhEZhDNi8njgJpypKYw5IXZlY0wQqWo+sFFExgK4Q8D3dV93UtWFqno/zhTEbYD9QNIJvGV9nMnZ9olIc46cLMz33AuBM0SkqYhE40wh8HnFk7lXZg1U9UPgVpzRq405YXZlY0zwXQ48IyL3ArE4z12WAY+JSBecq4x5btsW4C73ltvDVX0jVV0mIktxbqttwLlVV24qMEdEtqvqmSJyN87cLgJ8qKqzfnpGkoBZIpLg7ndbVTMZ4491fTbGGBNydhvNGGNMyNltNGOOk4j8EzilQvNTqvqiF3mMCWd2G80YY0zI2W00Y4wxIWfFxhhjTMhZsTHGGBNyVmyMMcaE3P8DvRq9ppBQqeMAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#list of values to try for n_estimators\n",
    "RMSE_scores_estim = []\n",
    "\n",
    "# use 5-fold cross-validation with each value of n_estimators (WARNING: SLOW!)\n",
    "for estimator in estimator_range:\n",
    "    rf  = RandomForestRegressor (n_estimators=estimator,max_features=8, max_depth=15, random_state=1, n_jobs=-1)\n",
    "    RMSE_scores_estim.append(cross_val_score(rf, X_train, Y_train, cv=5, scoring='neg_mean_squared_error').mean())\n",
    "\n",
    "best_n_estimator = estimator_range[RMSE_scores_estim.index(max(RMSE_scores_estim))]\n",
    "print (\"best_n_estimator: \", best_n_estimator)\n",
    "plt.plot(estimator_range, RMSE_scores_estim)\n",
    "plt.xlabel('n_estimators')\n",
    "plt.ylabel('NEG_RMSE')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Finding max depth"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#list of values to try for n_estimators\n",
    "RMSE_scores_depth = []\n",
    "# use 5-fold cross-validation with each value of n_estimators (WARNING: SLOW!)\n",
    "for depth in max_depth_range:\n",
    "    rf  = RandomForestRegressor (n_estimators=400, max_depth=depth, random_state=1, n_jobs=-1)\n",
    "    RMSE_scores_depth.append(cross_val_score(rf, X_train, Y_train, cv=5, scoring='neg_mean_squared_error').mean())\n",
    "\n",
    "best_depth = max_depth_range[RMSE_scores_depth.index(max(RMSE_scores_depth))]\n",
    "print (\"best_depth: \", best_depth)\n",
    "plt.plot(max_depth_range, RMSE_scores_depth)\n",
    "plt.xlabel('max_depth')\n",
    "plt.ylabel('RMSE')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Finding max features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "best_maxfeatures:  4\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'NEG_RMSE')"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# list of values to try for max_features\n",
    "\n",
    "RMSE_scores_featu = []\n",
    "# use 10-fold cross-validation with each value of max_features (WARNING: SLOW!)\n",
    "for feature in feature_range:\n",
    "    rf  = RandomForestRegressor(n_estimators=500, max_features=feature, max_depth=15, random_state=1, n_jobs=-1)\n",
    "    RMSE_scores_featu.append(cross_val_score(rf, X, Y, cv=5, scoring='neg_mean_squared_error').mean())\n",
    "\n",
    "best_maxfeatures = feature_range[RMSE_scores_featu.index(max(RMSE_scores_featu))]\n",
    "print (\"best_maxfeatures: \", best_maxfeatures)\n",
    "plt.plot(feature_range, RMSE_scores_featu)\n",
    "plt.xlabel('MAX FEATURES')\n",
    "plt.ylabel('NEG_RMSE')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Train the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Y:  (500000,)\n",
      "X:  (500000, 30)\n",
      "Rows train:  350000\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Make_0</th>\n",
       "      <th>Make_1</th>\n",
       "      <th>Make_2</th>\n",
       "      <th>Make_3</th>\n",
       "      <th>Make_4</th>\n",
       "      <th>Make_5</th>\n",
       "      <th>Make_6</th>\n",
       "      <th>Model_0</th>\n",
       "      <th>Model_1</th>\n",
       "      <th>Model_2</th>\n",
       "      <th>...</th>\n",
       "      <th>MakeModel4_2</th>\n",
       "      <th>MakeModel4_3</th>\n",
       "      <th>MakeModel4_4</th>\n",
       "      <th>MakeModel4_5</th>\n",
       "      <th>MakeModel4_6</th>\n",
       "      <th>MakeModel4_7</th>\n",
       "      <th>MakeModel4_8</th>\n",
       "      <th>Year</th>\n",
       "      <th>Mileage</th>\n",
       "      <th>RateState</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>index</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>377727</th>\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",
       "      <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>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>7</td>\n",
       "      <td>57432</td>\n",
       "      <td>0.999828</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>212934</th>\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",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</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>1</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>59909</td>\n",
       "      <td>0.974889</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>341392</th>\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",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>5</td>\n",
       "      <td>26535</td>\n",
       "      <td>1.080224</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>269561</th>\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",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>29053</td>\n",
       "      <td>0.974889</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>396570</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</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>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>7</td>\n",
       "      <td>85056</td>\n",
       "      <td>1.007226</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 30 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        Make_0  Make_1  Make_2  Make_3  Make_4  Make_5  Make_6  Model_0  \\\n",
       "index                                                                     \n",
       "377727       0       0       0       0       1       0       1        0   \n",
       "212934       0       0       0       0       1       0       1        0   \n",
       "341392       0       0       0       1       0       0       0        0   \n",
       "269561       0       0       0       1       0       0       0        0   \n",
       "396570       0       0       1       1       0       0       0        0   \n",
       "\n",
       "        Model_1  Model_2    ...      MakeModel4_2  MakeModel4_3  MakeModel4_4  \\\n",
       "index                       ...                                                 \n",
       "377727        0        0    ...                 0             0             0   \n",
       "212934        0        0    ...                 0             1             0   \n",
       "341392        0        0    ...                 1             0             0   \n",
       "269561        0        0    ...                 0             0             1   \n",
       "396570        0        0    ...                 1             0             0   \n",
       "\n",
       "        MakeModel4_5  MakeModel4_6  MakeModel4_7  MakeModel4_8  Year  Mileage  \\\n",
       "index                                                                           \n",
       "377727             0             1             1             1     7    57432   \n",
       "212934             0             0             1             0     4    59909   \n",
       "341392             1             1             1             0     5    26535   \n",
       "269561             0             1             0             0     2    29053   \n",
       "396570             1             1             0             0     7    85056   \n",
       "\n",
       "        RateState  \n",
       "index              \n",
       "377727   0.999828  \n",
       "212934   0.974889  \n",
       "341392   1.080224  \n",
       "269561   0.974889  \n",
       "396570   1.007226  \n",
       "\n",
       "[5 rows x 30 columns]"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Create de features\n",
    "Y = data.Price\n",
    "X = BinEncoder.transform(data.drop(['Price'], axis=1))\n",
    "print(\"Y: \", Y.shape)\n",
    "print(\"X: \", X.shape)\n",
    "\n",
    "#Train the model selecting the best found parameters\n",
    "size = 0.3\n",
    "X_train, X_validation, Y_train, Y_validation = train_test_split(X, Y,test_size=size).copy()\n",
    "print (\"Rows train: \", len(X_train))\n",
    "X_train.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Se estiman un RandomForest con los hiper parámetros encontrados y se hacen las predicciones correspondientes."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestRegressor(bootstrap=True, criterion='mse', max_depth=15,\n",
       "           max_features=8, max_leaf_nodes=None, min_impurity_decrease=0.0,\n",
       "           min_impurity_split=None, min_samples_leaf=1,\n",
       "           min_samples_split=2, min_weight_fraction_leaf=0.0,\n",
       "           n_estimators=100, n_jobs=-1, oob_score=False, random_state=1,\n",
       "           verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rf = RandomForestRegressor(n_estimators=950, max_depth=15, max_features=8, random_state=1, n_jobs=-1)\n",
    "rf.fit(X_train, Y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE: 3473.774499301229\n"
     ]
    }
   ],
   "source": [
    "y_pred_train = rf.predict(X_validation)\n",
    "print('RMSE:', np.sqrt(metrics.mean_squared_error(Y_validation, y_pred_train)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Train the model with all data\n",
    "\n",
    "Se estima un modelo con toda la base de datos y se hacen las predicciones del caso."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestRegressor(bootstrap=True, criterion='mse', max_depth=15,\n",
       "           max_features=8, max_leaf_nodes=None, min_impurity_decrease=0.0,\n",
       "           min_impurity_split=None, min_samples_leaf=1,\n",
       "           min_samples_split=2, min_weight_fraction_leaf=0.0,\n",
       "           n_estimators=100, n_jobs=-1, oob_score=False, random_state=1,\n",
       "           verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rf.fit(X, Y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_test = pd.read_csv('https://github.com/albahnsen/PracticalMachineLearningClass/raw/master/datasets/dataTest_carListings.zip', index_col=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Year</th>\n",
       "      <th>Mileage</th>\n",
       "      <th>Make</th>\n",
       "      <th>Model</th>\n",
       "      <th>RateState</th>\n",
       "      <th>MakeModel4</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ID</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>134</th>\n",
       "      <td>16</td>\n",
       "      <td>189346</td>\n",
       "      <td>Toyota</td>\n",
       "      <td>Tundra4WD</td>\n",
       "      <td>1.042688</td>\n",
       "      <td>ToyotaTund</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1108</th>\n",
       "      <td>2</td>\n",
       "      <td>2186</td>\n",
       "      <td>Chevrolet</td>\n",
       "      <td>TraverseAWD</td>\n",
       "      <td>1.042688</td>\n",
       "      <td>ChevroletTrav</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1418</th>\n",
       "      <td>7</td>\n",
       "      <td>62756</td>\n",
       "      <td>Ford</td>\n",
       "      <td>Fusion4dr</td>\n",
       "      <td>1.042688</td>\n",
       "      <td>FordFusi</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2311</th>\n",
       "      <td>4</td>\n",
       "      <td>30595</td>\n",
       "      <td>Nissan</td>\n",
       "      <td>Pathfinder4WD</td>\n",
       "      <td>1.042688</td>\n",
       "      <td>NissanPath</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2432</th>\n",
       "      <td>4</td>\n",
       "      <td>16094</td>\n",
       "      <td>Chevrolet</td>\n",
       "      <td>Silverado</td>\n",
       "      <td>1.042688</td>\n",
       "      <td>ChevroletSilv</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      Year  Mileage       Make          Model  RateState     MakeModel4\n",
       "ID                                                                     \n",
       "134     16   189346     Toyota      Tundra4WD   1.042688     ToyotaTund\n",
       "1108     2     2186  Chevrolet    TraverseAWD   1.042688  ChevroletTrav\n",
       "1418     7    62756       Ford      Fusion4dr   1.042688       FordFusi\n",
       "2311     4    30595     Nissan  Pathfinder4WD   1.042688     NissanPath\n",
       "2432     4    16094  Chevrolet      Silverado   1.042688  ChevroletSilv"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_test = createFeatures(data_test).set_index('ID')\n",
    "data_test.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "X_test:  (250000, 30)\n"
     ]
    }
   ],
   "source": [
    "#Codify the test dataset with \"BinEncoder\"\n",
    "X_test = BinEncoder.transform(data_test)\n",
    "print(\"X_test: \", X_test.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Predict the price for the test dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The data was exported to CSV file call test_submission.csv\n"
     ]
    }
   ],
   "source": [
    "y_pred = pd.DataFrame(rf.predict(X_test),index=data_test.index,columns=['Price'])\n",
    "y_pred.to_csv('test_submission.csv', index_label='ID')\n",
    "print(\"The data was exported to CSV file call test_submission.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['RF_modelcars.pkl']"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "joblib.dump(rf, 'RF_modelcars.pkl', compress=3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Se estima un modelo tipo boosting, pero no resulta mejor que el Random Forest."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "from sklearn.ensemble import GradientBoostingRegressor\n",
    "clfgrad = GradientBoostingRegressor(n_estimators=400, max_depth=15, max_features=8, random_state=1)\n",
    "clfgrad.fit(X_train, Y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "y_pred_clfgrad = clfgrad.predict(X_validation)\n",
    "print('RMSE:', np.sqrt(metrics.mean_squared_error(Y_validation, y_pred_clfgrad)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "clfAda = AdaBoostRegressor(RandomForestRegressor(n_estimators=800, max_depth=15, max_features=8, random_state=1, n_jobs=-1), n_estimators=10)\n",
    "clfAda.fit(X_train, Y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "y_pred_clfAda = clfAda.predict(X_validation)\n",
    "print('RMSE:', np.sqrt(metrics.mean_squared_error(Y_validation, y_pred_clfAda)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Exercise P1.2 (50%)\n",
    "\n",
    "Create an API of the model.\n",
    "\n",
    "![](https://github.com/gomezivan/Uniandes/blob/master/Screenshot%20WebService.png)\n",
    "\n",
    "https://github.com/gomezivan/Uniandes/blob/master/Screenshot%20WebService.png\n",
    "\n",
    "#### Evaluation:\n",
    "- 40% - API hosted on a cloud service\n",
    "- 10% - Show screenshots of the model doing the predictions on the local machine]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}