3.1-regression.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Clasificación útilizando aprendizaje de máquinas"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "En aprendizaje de máquinas, la regresión es la tarea de identificar a qué valor continuo una observación nueva pertenece, tomando como base un entrenamiento previo sobre un grupo de observaciones donde sí se conocían sus valores continuos correspondientes."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import sklearn.preprocessing \n",
    "import sklearn.ensemble\n",
    "import sklearn.linear_model\n",
    "import sklearn.svm\n",
    "import sklearn.dummy\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns # Paquete de visualización que funciona sobre matplotlib"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Para entrenar y probar los modelos de clasificación, utilizaremos un conjunto de datos que contiene valores de propiedades junto con características asociadas a las mismas."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "real_estate = pd.read_csv(\"data/real_estate.csv\", delimiter=\";\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "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>transaction_date</th>\n",
       "      <th>house_age</th>\n",
       "      <th>distance_nearest_station</th>\n",
       "      <th>number_convenience_stores</th>\n",
       "      <th>latitude</th>\n",
       "      <th>longitude</th>\n",
       "      <th>price_unit_area</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2012.916667</td>\n",
       "      <td>32.0</td>\n",
       "      <td>84.87882</td>\n",
       "      <td>10</td>\n",
       "      <td>24.98298</td>\n",
       "      <td>121.54024</td>\n",
       "      <td>37.9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2012.916667</td>\n",
       "      <td>19.5</td>\n",
       "      <td>306.59470</td>\n",
       "      <td>9</td>\n",
       "      <td>24.98034</td>\n",
       "      <td>121.53951</td>\n",
       "      <td>42.2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2013.583333</td>\n",
       "      <td>13.3</td>\n",
       "      <td>561.98450</td>\n",
       "      <td>5</td>\n",
       "      <td>24.98746</td>\n",
       "      <td>121.54391</td>\n",
       "      <td>47.3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2013.500000</td>\n",
       "      <td>13.3</td>\n",
       "      <td>561.98450</td>\n",
       "      <td>5</td>\n",
       "      <td>24.98746</td>\n",
       "      <td>121.54391</td>\n",
       "      <td>54.8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2012.833333</td>\n",
       "      <td>5.0</td>\n",
       "      <td>390.56840</td>\n",
       "      <td>5</td>\n",
       "      <td>24.97937</td>\n",
       "      <td>121.54245</td>\n",
       "      <td>43.1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   transaction_date  house_age  distance_nearest_station  \\\n",
       "0       2012.916667       32.0                  84.87882   \n",
       "1       2012.916667       19.5                 306.59470   \n",
       "2       2013.583333       13.3                 561.98450   \n",
       "3       2013.500000       13.3                 561.98450   \n",
       "4       2012.833333        5.0                 390.56840   \n",
       "\n",
       "   number_convenience_stores  latitude  longitude  price_unit_area  \n",
       "0                         10  24.98298  121.54024             37.9  \n",
       "1                          9  24.98034  121.53951             42.2  \n",
       "2                          5  24.98746  121.54391             47.3  \n",
       "3                          5  24.98746  121.54391             54.8  \n",
       "4                          5  24.97937  121.54245             43.1  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "real_estate.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Verificamos con el método pandas.DataFrame.describe() que las variables no tienen el mismo rango, por lo que seria necesario reescalarlas."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "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>transaction_date</th>\n",
       "      <th>house_age</th>\n",
       "      <th>distance_nearest_station</th>\n",
       "      <th>number_convenience_stores</th>\n",
       "      <th>latitude</th>\n",
       "      <th>longitude</th>\n",
       "      <th>price_unit_area</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>414.000000</td>\n",
       "      <td>414.000000</td>\n",
       "      <td>414.000000</td>\n",
       "      <td>414.000000</td>\n",
       "      <td>414.000000</td>\n",
       "      <td>414.000000</td>\n",
       "      <td>414.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>2013.148953</td>\n",
       "      <td>17.712560</td>\n",
       "      <td>1083.885689</td>\n",
       "      <td>4.094203</td>\n",
       "      <td>24.969030</td>\n",
       "      <td>121.533361</td>\n",
       "      <td>37.980193</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.281995</td>\n",
       "      <td>11.392485</td>\n",
       "      <td>1262.109595</td>\n",
       "      <td>2.945562</td>\n",
       "      <td>0.012410</td>\n",
       "      <td>0.015347</td>\n",
       "      <td>13.606488</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>2012.666667</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>23.382840</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>24.932070</td>\n",
       "      <td>121.473530</td>\n",
       "      <td>7.600000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>2012.916667</td>\n",
       "      <td>9.025000</td>\n",
       "      <td>289.324800</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>24.963000</td>\n",
       "      <td>121.528085</td>\n",
       "      <td>27.700000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>2013.166667</td>\n",
       "      <td>16.100000</td>\n",
       "      <td>492.231300</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>24.971100</td>\n",
       "      <td>121.538630</td>\n",
       "      <td>38.450000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>2013.416667</td>\n",
       "      <td>28.150000</td>\n",
       "      <td>1454.279000</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>24.977455</td>\n",
       "      <td>121.543305</td>\n",
       "      <td>46.600000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>2013.583333</td>\n",
       "      <td>43.800000</td>\n",
       "      <td>6488.021000</td>\n",
       "      <td>10.000000</td>\n",
       "      <td>25.014590</td>\n",
       "      <td>121.566270</td>\n",
       "      <td>117.500000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       transaction_date   house_age  distance_nearest_station  \\\n",
       "count        414.000000  414.000000                414.000000   \n",
       "mean        2013.148953   17.712560               1083.885689   \n",
       "std            0.281995   11.392485               1262.109595   \n",
       "min         2012.666667    0.000000                 23.382840   \n",
       "25%         2012.916667    9.025000                289.324800   \n",
       "50%         2013.166667   16.100000                492.231300   \n",
       "75%         2013.416667   28.150000               1454.279000   \n",
       "max         2013.583333   43.800000               6488.021000   \n",
       "\n",
       "       number_convenience_stores    latitude   longitude  price_unit_area  \n",
       "count                 414.000000  414.000000  414.000000       414.000000  \n",
       "mean                    4.094203   24.969030  121.533361        37.980193  \n",
       "std                     2.945562    0.012410    0.015347        13.606488  \n",
       "min                     0.000000   24.932070  121.473530         7.600000  \n",
       "25%                     1.000000   24.963000  121.528085        27.700000  \n",
       "50%                     4.000000   24.971100  121.538630        38.450000  \n",
       "75%                     6.000000   24.977455  121.543305        46.600000  \n",
       "max                    10.000000   25.014590  121.566270       117.500000  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "real_estate.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(414, 7)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "real_estate.shape # Obtenemos las dimensiones de nuestro conjunto de datos."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Para obtener una intuición general de las relaciones que existen entre variables, una práctica habitual es evaluar la correlación que existe entre las variables, sobretodo la correlación que existe entre las características y la etiqueta. Para esto utilizamos indicadores de estadística clásica como es el coeficiente de correlación de Pearson o Spearman."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "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>transaction_date</th>\n",
       "      <th>house_age</th>\n",
       "      <th>distance_nearest_station</th>\n",
       "      <th>number_convenience_stores</th>\n",
       "      <th>latitude</th>\n",
       "      <th>longitude</th>\n",
       "      <th>price_unit_area</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>transaction_date</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.017542</td>\n",
       "      <td>0.060880</td>\n",
       "      <td>0.009544</td>\n",
       "      <td>0.035016</td>\n",
       "      <td>-0.041065</td>\n",
       "      <td>0.087529</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>house_age</th>\n",
       "      <td>0.017542</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.025622</td>\n",
       "      <td>0.049593</td>\n",
       "      <td>0.054420</td>\n",
       "      <td>-0.048520</td>\n",
       "      <td>-0.210567</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>distance_nearest_station</th>\n",
       "      <td>0.060880</td>\n",
       "      <td>0.025622</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.602519</td>\n",
       "      <td>-0.591067</td>\n",
       "      <td>-0.806317</td>\n",
       "      <td>-0.673613</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>number_convenience_stores</th>\n",
       "      <td>0.009544</td>\n",
       "      <td>0.049593</td>\n",
       "      <td>-0.602519</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.444143</td>\n",
       "      <td>0.449099</td>\n",
       "      <td>0.571005</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>latitude</th>\n",
       "      <td>0.035016</td>\n",
       "      <td>0.054420</td>\n",
       "      <td>-0.591067</td>\n",
       "      <td>0.444143</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.412924</td>\n",
       "      <td>0.546307</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>longitude</th>\n",
       "      <td>-0.041065</td>\n",
       "      <td>-0.048520</td>\n",
       "      <td>-0.806317</td>\n",
       "      <td>0.449099</td>\n",
       "      <td>0.412924</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.523287</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>price_unit_area</th>\n",
       "      <td>0.087529</td>\n",
       "      <td>-0.210567</td>\n",
       "      <td>-0.673613</td>\n",
       "      <td>0.571005</td>\n",
       "      <td>0.546307</td>\n",
       "      <td>0.523287</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                           transaction_date  house_age  \\\n",
       "transaction_date                   1.000000   0.017542   \n",
       "house_age                          0.017542   1.000000   \n",
       "distance_nearest_station           0.060880   0.025622   \n",
       "number_convenience_stores          0.009544   0.049593   \n",
       "latitude                           0.035016   0.054420   \n",
       "longitude                         -0.041065  -0.048520   \n",
       "price_unit_area                    0.087529  -0.210567   \n",
       "\n",
       "                           distance_nearest_station  \\\n",
       "transaction_date                           0.060880   \n",
       "house_age                                  0.025622   \n",
       "distance_nearest_station                   1.000000   \n",
       "number_convenience_stores                 -0.602519   \n",
       "latitude                                  -0.591067   \n",
       "longitude                                 -0.806317   \n",
       "price_unit_area                           -0.673613   \n",
       "\n",
       "                           number_convenience_stores  latitude  longitude  \\\n",
       "transaction_date                            0.009544  0.035016  -0.041065   \n",
       "house_age                                   0.049593  0.054420  -0.048520   \n",
       "distance_nearest_station                   -0.602519 -0.591067  -0.806317   \n",
       "number_convenience_stores                   1.000000  0.444143   0.449099   \n",
       "latitude                                    0.444143  1.000000   0.412924   \n",
       "longitude                                   0.449099  0.412924   1.000000   \n",
       "price_unit_area                             0.571005  0.546307   0.523287   \n",
       "\n",
       "                           price_unit_area  \n",
       "transaction_date                  0.087529  \n",
       "house_age                        -0.210567  \n",
       "distance_nearest_station         -0.673613  \n",
       "number_convenience_stores         0.571005  \n",
       "latitude                          0.546307  \n",
       "longitude                         0.523287  \n",
       "price_unit_area                   1.000000  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "correlation = real_estate.corr() # Por defecto este método calcula la correlación de Pearson.\n",
    "correlation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x24ca0ab3c08>"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.heatmap(\n",
    "    correlation,\n",
    "    cmap=\"RdYlGn\" # Utilizamos una paleta de colores divergente para visualizar diferencias en los valores positivos y negativos\n",
    ") # Utilizamos seaborn para visualizar un mapa de calor de las correlaciones."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Preprocesamiento"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "feature_names = [\n",
    "    'transaction_date', \n",
    "    'house_age', \n",
    "    'distance_nearest_station', \n",
    "    'number_convenience_stores', \n",
    "    'latitude', \n",
    "    'longitude'\n",
    "]\n",
    "\n",
    "features = real_estate[feature_names]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "label = real_estate.price_unit_area"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Los algoritmos para entrenar modelos de aprendizaje de máquinas no se comportan bien en su mayoría con diferencias en los rangos de valores de sus características, por lo que es siempre una buena práctica reescalar los valores hacia un rango predeterminado."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "scaler = sklearn.preprocessing.MinMaxScaler() # Instanciamos la clase del scaler que realizará una normalización de los datos\n",
    "features = scaler.fit_transform(features) # Ajustamos el scaler y al mismo tiempo transformamos nuestras características"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Verificamos que el rango de todas nuestras características ahora esté entre 0 y 1."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "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>transaction_date</th>\n",
       "      <th>house_age</th>\n",
       "      <th>distance_nearest_station</th>\n",
       "      <th>number_convenience_stores</th>\n",
       "      <th>latitude</th>\n",
       "      <th>longitude</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>414.000000</td>\n",
       "      <td>414.000000</td>\n",
       "      <td>414.000000</td>\n",
       "      <td>414.000000</td>\n",
       "      <td>414.000000</td>\n",
       "      <td>414.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>0.526131</td>\n",
       "      <td>0.404396</td>\n",
       "      <td>0.164047</td>\n",
       "      <td>0.409420</td>\n",
       "      <td>0.447892</td>\n",
       "      <td>0.645149</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.307631</td>\n",
       "      <td>0.260102</td>\n",
       "      <td>0.195233</td>\n",
       "      <td>0.294556</td>\n",
       "      <td>0.150390</td>\n",
       "      <td>0.165486</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>0.272727</td>\n",
       "      <td>0.206050</td>\n",
       "      <td>0.041138</td>\n",
       "      <td>0.100000</td>\n",
       "      <td>0.374818</td>\n",
       "      <td>0.588257</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>0.545455</td>\n",
       "      <td>0.367580</td>\n",
       "      <td>0.072525</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>0.472976</td>\n",
       "      <td>0.701962</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>0.818182</td>\n",
       "      <td>0.642694</td>\n",
       "      <td>0.221342</td>\n",
       "      <td>0.600000</td>\n",
       "      <td>0.549988</td>\n",
       "      <td>0.752372</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       transaction_date   house_age  distance_nearest_station  \\\n",
       "count        414.000000  414.000000                414.000000   \n",
       "mean           0.526131    0.404396                  0.164047   \n",
       "std            0.307631    0.260102                  0.195233   \n",
       "min            0.000000    0.000000                  0.000000   \n",
       "25%            0.272727    0.206050                  0.041138   \n",
       "50%            0.545455    0.367580                  0.072525   \n",
       "75%            0.818182    0.642694                  0.221342   \n",
       "max            1.000000    1.000000                  1.000000   \n",
       "\n",
       "       number_convenience_stores    latitude   longitude  \n",
       "count                 414.000000  414.000000  414.000000  \n",
       "mean                    0.409420    0.447892    0.645149  \n",
       "std                     0.294556    0.150390    0.165486  \n",
       "min                     0.000000    0.000000    0.000000  \n",
       "25%                     0.100000    0.374818    0.588257  \n",
       "50%                     0.400000    0.472976    0.701962  \n",
       "75%                     0.600000    0.549988    0.752372  \n",
       "max                     1.000000    1.000000    1.000000  "
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame(features,columns=feature_names).describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "features_train, features_test, label_train, label_test = sklearn.model_selection.train_test_split( \n",
    "    features,\n",
    "    label,\n",
    "    test_size=0.30, \n",
    "    random_state = 11, \n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Modelamiento"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Para evaluar el rendimiento de nuestro modelos sobre el conjunto de prueba debemos utilizar métricas distintas a las que usábamos en clasificación. sklearn nos ofrece una serie de métricas especificas para regresión, las cuales se encuentran todas descritas en la documentación https://scikit-learn.org/stable/modules/model_evaluation.html ."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "def regression_report(y_true, y_pred):\n",
    "    \"\"\"\n",
    "    Esta función recibe un arreglo de valores reales y predichos para \n",
    "    retornar un diccionario con una serie de métricas de regresión\n",
    "    \"\"\"\n",
    "    return {\n",
    "        'mae': sklearn.metrics.mean_absolute_error(y_true, y_pred),\n",
    "        'rmse': sklearn.metrics.mean_squared_error(y_true, y_pred) ** 0.5,\n",
    "        'r2': sklearn.metrics.r2_score(y_true, y_pred)\n",
    "    }"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Regresión lineal"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LinearRegression(copy_X=True, fit_intercept=True, n_jobs=None, normalize=False)"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lr = sklearn.linear_model.LinearRegression()\n",
    "lr.fit(features_train, label_train) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "lr_predictions = lr.predict(features_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'mae': 7.098167201942863,\n",
       " 'rmse': 11.154714425923883,\n",
       " 'r2': 0.5040618675072719}"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lr_regression_report = regression_report(label_test, lr_predictions)\n",
    "lr_regression_report"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Support vector machine"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "SVR(C=1.0, cache_size=200, coef0=0.0, degree=3, epsilon=0.1, gamma='scale',\n",
       "    kernel='rbf', max_iter=-1, shrinking=True, tol=0.001, verbose=False)"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "svm = sklearn.svm.SVR() # Instanciamos una support vector machine con un kernel lineal\n",
    "svm.fit(features_train, label_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "svm_predictions = svm.predict(features_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'mae': 7.621752619669486, 'rmse': 11.98334176277971, 'r2': 0.4276436826302662}"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "svm_regression_report = regression_report(label_test, svm_predictions)\n",
    "svm_regression_report"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Random forests"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestRegressor(bootstrap=True, ccp_alpha=0.0, criterion='mse',\n",
       "                      max_depth=None, max_features='auto', max_leaf_nodes=None,\n",
       "                      max_samples=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=None, oob_score=False,\n",
       "                      random_state=None, verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rf = sklearn.ensemble.RandomForestRegressor()\n",
    "rf.fit(features_train, label_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "rf_predictions = rf.predict(features_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'mae': 5.064831819047617, 'rmse': 9.73887919364118, 'r2': 0.6219680147638262}"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rf_regression_report = regression_report(label_test, rf_predictions)\n",
    "rf_regression_report"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Regresor tonto"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DummyRegressor(constant=None, quantile=None, strategy='mean')"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dummy = sklearn.dummy.DummyRegressor()\n",
    "dummy.fit(features_train, label_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "dummy_predictions = dummy.predict(features_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'mae': 12.224824913494809,\n",
       " 'rmse': 15.84202098213816,\n",
       " 'r2': -0.000303430094670043}"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dummy_regression_report = regression_report(label_test, dummy_predictions)\n",
    "dummy_regression_report"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Resumen"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Evaluemos cuál fue el modelo que tuvo el mejor rendimiento en cada una de las métricas."
   ]
  },
  {
   "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>mae</th>\n",
       "      <th>rmse</th>\n",
       "      <th>r2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Random Forest</th>\n",
       "      <td>5.064832</td>\n",
       "      <td>9.738879</td>\n",
       "      <td>0.621968</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Linear Regression</th>\n",
       "      <td>7.098167</td>\n",
       "      <td>11.154714</td>\n",
       "      <td>0.504062</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Support Vector Machine</th>\n",
       "      <td>7.621753</td>\n",
       "      <td>11.983342</td>\n",
       "      <td>0.427644</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Regresor Tonto</th>\n",
       "      <td>12.224825</td>\n",
       "      <td>15.842021</td>\n",
       "      <td>-0.000303</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                              mae       rmse        r2\n",
       "Random Forest            5.064832   9.738879  0.621968\n",
       "Linear Regression        7.098167  11.154714  0.504062\n",
       "Support Vector Machine   7.621753  11.983342  0.427644\n",
       "Regresor Tonto          12.224825  15.842021 -0.000303"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "performances = pd.DataFrame( # Consolidamos todas las métricas en un DatFrame\n",
    "    data = [\n",
    "        lr_regression_report,\n",
    "        svm_regression_report,\n",
    "        rf_regression_report,\n",
    "        dummy_regression_report\n",
    "    ],\n",
    "    index = [\n",
    "        \"Linear Regression\",\n",
    "        \"Support Vector Machine\",\n",
    "        \"Random Forest\",\n",
    "        \"Regresor Tonto\"\n",
    "    ]\n",
    ").sort_values( # Ordenamos los valores\n",
    "    by=\"rmse\"\n",
    ")\n",
    "performances"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "([0, 1, 2, 3], <a list of 4 Text xticklabel objects>)"
      ]
     },
     "execution_count": 27,
     "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": [
    "plt.bar( # Visualizamos los resultados como un gráfico de barras\n",
    "    x=performances.index,\n",
    "    height=performances.mae,\n",
    ")\n",
    "plt.title(\"Mean Absolute Error\")\n",
    "plt.xlabel(\"Model\")\n",
    "plt.ylabel(\"Value\")\n",
    "plt.xticks(rotation=45)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "([0, 1, 2, 3], <a list of 4 Text xticklabel objects>)"
      ]
     },
     "execution_count": 28,
     "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": [
    "plt.bar(\n",
    "    x=performances.index,\n",
    "    height=performances.rmse,\n",
    ")\n",
    "plt.title(\"Root Mean Squared Error\")\n",
    "plt.xlabel(\"Model\")\n",
    "plt.ylabel(\"Value\")\n",
    "plt.xticks(rotation=45)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "([0, 1, 2, 3], <a list of 4 Text xticklabel objects>)"
      ]
     },
     "execution_count": 29,
     "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": [
    "plt.bar(\n",
    "    x=performances.index,\n",
    "    height=performances.r2,\n",
    ")\n",
    "plt.title(\"R$^2$\")\n",
    "plt.xlabel(\"Model\")\n",
    "plt.ylabel(\"Value\")\n",
    "plt.xticks(rotation=45)"
   ]
  },
  {
   "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": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}