Serie temporal - Skforecast - OLOT.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "2eede2c0",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd \n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from sklearn import preprocessing\n",
    "from sklearn.model_selection import train_test_split\n",
    "from scipy import stats\n",
    "from statsmodels.tsa.stattools import adfuller\n",
    "from statsmodels.graphics.tsaplots import plot_acf, plot_pacf\n",
    "from pmdarima.arima import auto_arima\n",
    "\n",
    "# Tratamiento de datos\n",
    "# ==============================================================================\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "\n",
    "# Gráficos\n",
    "# ==============================================================================\n",
    "import matplotlib.pyplot as plt\n",
    "plt.style.use('fivethirtyeight')\n",
    "plt.rcParams['lines.linewidth'] = 1.5\n",
    "\n",
    "# Modelado y Forecasting\n",
    "# ==============================================================================\n",
    "from sklearn.linear_model import LinearRegression\n",
    "from sklearn.linear_model import Lasso\n",
    "from sklearn.ensemble import RandomForestRegressor\n",
    "from sklearn.metrics import mean_squared_error\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.pipeline import make_pipeline\n",
    "\n",
    "\n",
    "from skforecast.ForecasterAutoreg import ForecasterAutoreg\n",
    "from skforecast.ForecasterAutoregCustom import ForecasterAutoregCustom\n",
    "from skforecast.ForecasterAutoregDirect import ForecasterAutoregDirect\n",
    "from skforecast.model_selection import grid_search_forecaster\n",
    "from skforecast.model_selection import backtesting_forecaster\n",
    "from skforecast.utils import save_forecaster\n",
    "from skforecast.utils import load_forecaster\n",
    "\n",
    "# Configuración warnings\n",
    "# ==============================================================================\n",
    "import warnings\n",
    "# warnings.filterwarnings('ignore')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "cb9399a0",
   "metadata": {},
   "outputs": [],
   "source": [
    "lugar_nacimiento=pd.read_csv('t674mun_171143OLOT.csv', sep=';')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "f4d86f08",
   "metadata": {},
   "outputs": [],
   "source": [
    "lugar_nacimiento=lugar_nacimiento.rename(columns={'Unnamed: 0': 'Año'})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "0d5fff16",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['Año', 'Catalunya', 'Mateixa comarca', 'Altra comarca',\n",
       "       'Resta d'Espanya', 'Estranger', 'Total'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lugar_nacimiento.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "ed7c87e3",
   "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>Año</th>\n",
       "      <th>Catalunya</th>\n",
       "      <th>Mateixa comarca</th>\n",
       "      <th>Altra comarca</th>\n",
       "      <th>Resta d'Espanya</th>\n",
       "      <th>Estranger</th>\n",
       "      <th>Total</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>2000</td>\n",
       "      <td>22790</td>\n",
       "      <td>18364</td>\n",
       "      <td>4426</td>\n",
       "      <td>4174</td>\n",
       "      <td>1003</td>\n",
       "      <td>27967</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>2001</td>\n",
       "      <td>22856</td>\n",
       "      <td>18396</td>\n",
       "      <td>4460</td>\n",
       "      <td>4156</td>\n",
       "      <td>1327</td>\n",
       "      <td>28339</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>2002</td>\n",
       "      <td>22887</td>\n",
       "      <td>18382</td>\n",
       "      <td>4505</td>\n",
       "      <td>4142</td>\n",
       "      <td>1861</td>\n",
       "      <td>28890</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>2003</td>\n",
       "      <td>22950</td>\n",
       "      <td>18325</td>\n",
       "      <td>4625</td>\n",
       "      <td>4105</td>\n",
       "      <td>2491</td>\n",
       "      <td>29546</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>2004</td>\n",
       "      <td>22999</td>\n",
       "      <td>18292</td>\n",
       "      <td>4707</td>\n",
       "      <td>4057</td>\n",
       "      <td>3250</td>\n",
       "      <td>30306</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>2005</td>\n",
       "      <td>23138</td>\n",
       "      <td>18310</td>\n",
       "      <td>4828</td>\n",
       "      <td>4048</td>\n",
       "      <td>4085</td>\n",
       "      <td>31271</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>2006</td>\n",
       "      <td>23067</td>\n",
       "      <td>18226</td>\n",
       "      <td>4841</td>\n",
       "      <td>3978</td>\n",
       "      <td>4887</td>\n",
       "      <td>31932</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>2007</td>\n",
       "      <td>22993</td>\n",
       "      <td>18193</td>\n",
       "      <td>4800</td>\n",
       "      <td>3920</td>\n",
       "      <td>5424</td>\n",
       "      <td>32337</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>2008</td>\n",
       "      <td>22973</td>\n",
       "      <td>18210</td>\n",
       "      <td>4763</td>\n",
       "      <td>3893</td>\n",
       "      <td>6037</td>\n",
       "      <td>32903</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>2009</td>\n",
       "      <td>23023</td>\n",
       "      <td>18212</td>\n",
       "      <td>4811</td>\n",
       "      <td>3840</td>\n",
       "      <td>6661</td>\n",
       "      <td>33524</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>2010</td>\n",
       "      <td>23032</td>\n",
       "      <td>18244</td>\n",
       "      <td>4788</td>\n",
       "      <td>3754</td>\n",
       "      <td>6803</td>\n",
       "      <td>33589</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>2011</td>\n",
       "      <td>23092</td>\n",
       "      <td>18254</td>\n",
       "      <td>4838</td>\n",
       "      <td>3680</td>\n",
       "      <td>6953</td>\n",
       "      <td>33725</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>2012</td>\n",
       "      <td>23115</td>\n",
       "      <td>18252</td>\n",
       "      <td>4863</td>\n",
       "      <td>3618</td>\n",
       "      <td>7080</td>\n",
       "      <td>33813</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>2013</td>\n",
       "      <td>23228</td>\n",
       "      <td>18328</td>\n",
       "      <td>4900</td>\n",
       "      <td>3582</td>\n",
       "      <td>7171</td>\n",
       "      <td>33981</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>2014</td>\n",
       "      <td>23245</td>\n",
       "      <td>18299</td>\n",
       "      <td>4946</td>\n",
       "      <td>3512</td>\n",
       "      <td>7156</td>\n",
       "      <td>33913</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>2015</td>\n",
       "      <td>23298</td>\n",
       "      <td>18325</td>\n",
       "      <td>4973</td>\n",
       "      <td>3509</td>\n",
       "      <td>7137</td>\n",
       "      <td>33944</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>2016</td>\n",
       "      <td>23316</td>\n",
       "      <td>18317</td>\n",
       "      <td>4999</td>\n",
       "      <td>3471</td>\n",
       "      <td>7213</td>\n",
       "      <td>34000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>2017</td>\n",
       "      <td>23346</td>\n",
       "      <td>18308</td>\n",
       "      <td>5038</td>\n",
       "      <td>3436</td>\n",
       "      <td>7412</td>\n",
       "      <td>34194</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2018</td>\n",
       "      <td>23323</td>\n",
       "      <td>18224</td>\n",
       "      <td>5099</td>\n",
       "      <td>3420</td>\n",
       "      <td>7743</td>\n",
       "      <td>34486</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2019</td>\n",
       "      <td>23346</td>\n",
       "      <td>18186</td>\n",
       "      <td>5160</td>\n",
       "      <td>3366</td>\n",
       "      <td>8516</td>\n",
       "      <td>35228</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2020</td>\n",
       "      <td>23314</td>\n",
       "      <td>18116</td>\n",
       "      <td>5198</td>\n",
       "      <td>3336</td>\n",
       "      <td>9276</td>\n",
       "      <td>35926</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2021</td>\n",
       "      <td>23309</td>\n",
       "      <td>18072</td>\n",
       "      <td>5237</td>\n",
       "      <td>3251</td>\n",
       "      <td>9739</td>\n",
       "      <td>36299</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2022</td>\n",
       "      <td>23316</td>\n",
       "      <td>18029</td>\n",
       "      <td>5287</td>\n",
       "      <td>3234</td>\n",
       "      <td>10166</td>\n",
       "      <td>36716</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     Año  Catalunya  Mateixa comarca  Altra comarca  Resta d'Espanya  \\\n",
       "22  2000      22790            18364           4426             4174   \n",
       "21  2001      22856            18396           4460             4156   \n",
       "20  2002      22887            18382           4505             4142   \n",
       "19  2003      22950            18325           4625             4105   \n",
       "18  2004      22999            18292           4707             4057   \n",
       "17  2005      23138            18310           4828             4048   \n",
       "16  2006      23067            18226           4841             3978   \n",
       "15  2007      22993            18193           4800             3920   \n",
       "14  2008      22973            18210           4763             3893   \n",
       "13  2009      23023            18212           4811             3840   \n",
       "12  2010      23032            18244           4788             3754   \n",
       "11  2011      23092            18254           4838             3680   \n",
       "10  2012      23115            18252           4863             3618   \n",
       "9   2013      23228            18328           4900             3582   \n",
       "8   2014      23245            18299           4946             3512   \n",
       "7   2015      23298            18325           4973             3509   \n",
       "6   2016      23316            18317           4999             3471   \n",
       "5   2017      23346            18308           5038             3436   \n",
       "4   2018      23323            18224           5099             3420   \n",
       "3   2019      23346            18186           5160             3366   \n",
       "2   2020      23314            18116           5198             3336   \n",
       "1   2021      23309            18072           5237             3251   \n",
       "0   2022      23316            18029           5287             3234   \n",
       "\n",
       "    Estranger  Total  \n",
       "22       1003  27967  \n",
       "21       1327  28339  \n",
       "20       1861  28890  \n",
       "19       2491  29546  \n",
       "18       3250  30306  \n",
       "17       4085  31271  \n",
       "16       4887  31932  \n",
       "15       5424  32337  \n",
       "14       6037  32903  \n",
       "13       6661  33524  \n",
       "12       6803  33589  \n",
       "11       6953  33725  \n",
       "10       7080  33813  \n",
       "9        7171  33981  \n",
       "8        7156  33913  \n",
       "7        7137  33944  \n",
       "6        7213  34000  \n",
       "5        7412  34194  \n",
       "4        7743  34486  \n",
       "3        8516  35228  \n",
       "2        9276  35926  \n",
       "1        9739  36299  \n",
       "0       10166  36716  "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lugar_nacimiento_olot=lugar_nacimiento.sort_values(by='Año', ascending=True)\n",
    "lugar_nacimiento_olot"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "5f7fb82b",
   "metadata": {},
   "outputs": [],
   "source": [
    "serie_extranjeros_olot = lugar_nacimiento_olot[['Año', 'Estranger']]\n",
    "serie_extranjeros_olot= lugar_nacimiento_olot.set_index('Año')\n",
    "serie_extranjeros_olot.index = pd.to_datetime(serie_extranjeros_olot.index, format='%Y')\n",
    "serie_extranjeros_olot = serie_extranjeros_olot['Estranger']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "b0461a00",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Número de filas con missing values: 0.0\n"
     ]
    }
   ],
   "source": [
    "print(f'Número de filas con missing values: {lugar_nacimiento.isnull().any(axis=1).mean()}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "a9305386",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fechas train : 2000-01-01 00:00:00 --- 2015-01-01 00:00:00  (n=16)\n",
      "Fechas test  : 2016-01-01 00:00:00 --- 2022-01-01 00:00:00  (n=7)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Separación datos train-test\n",
    "# ==============================================================================\n",
    "steps = 7 #se utilizan los últimos 7 años como conjunto test para evaluar la capacidad del modelo\n",
    "datos_train = serie_extranjeros_olot[:-steps]\n",
    "datos_test  = serie_extranjeros_olot[-steps:]\n",
    "\n",
    "print(f\"Fechas train : {datos_train.index.min()} --- {datos_train.index.max()}  (n={len(datos_train)})\")\n",
    "print(f\"Fechas test  : {datos_test.index.min()} --- {datos_test.index.max()}  (n={len(datos_test)})\")\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(10, 6))\n",
    "datos_train.plot(ax=ax, label='train')\n",
    "datos_test.plot(ax=ax, label='test')\n",
    "ax.legend();"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e8eea12e",
   "metadata": {},
   "source": [
    "### ForecasterAutoreg\n",
    "\n",
    "Se crea y entrena un modelo ForecasterAutoreg a partir de un regresor RandomForestRegressor y una ventana temporal de 7 lags. Esto último significa que, el modelo, utiliza como predictores los 7 años anteriores."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "0d20f04c",
   "metadata": {},
   "outputs": [],
   "source": [
    "#datos_test.index = pd.to_datetime(datos_test.index, format='%Y-%m-%d')\n",
    "#datos_train.index = pd.to_datetime(datos_train.index, format='%Y-%m-%d')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "281d1c24",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Año\n",
       "2016-01-01     7213\n",
       "2017-01-01     7412\n",
       "2018-01-01     7743\n",
       "2019-01-01     8516\n",
       "2020-01-01     9276\n",
       "2021-01-01     9739\n",
       "2022-01-01    10166\n",
       "Name: Estranger, dtype: int64"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "datos_test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "7cdc83f3",
   "metadata": {},
   "outputs": [],
   "source": [
    "freq = pd.infer_freq(datos_train.index)\n",
    "datos_train.index = pd.date_range(start=datos_train.index[0], periods=len(datos_train), freq=freq)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "597eaf93",
   "metadata": {},
   "outputs": [],
   "source": [
    "freq = pd.infer_freq(datos_test.index)\n",
    "datos_test.index = pd.date_range(start=datos_test.index[0], periods=len(datos_test), freq=freq)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "22dcdce1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "================= \n",
       "ForecasterAutoreg \n",
       "================= \n",
       "Regressor: RandomForestRegressor(random_state=123) \n",
       "Lags: [1 2 3 4 5 6 7] \n",
       "Transformer for y: None \n",
       "Transformer for exog: None \n",
       "Window size: 7 \n",
       "Weight function included: False \n",
       "Exogenous included: False \n",
       "Type of exogenous variable: None \n",
       "Exogenous variables names: None \n",
       "Training range: [Timestamp('2000-01-01 00:00:00'), Timestamp('2015-01-01 00:00:00')] \n",
       "Training index type: DatetimeIndex \n",
       "Training index frequency: AS-JAN \n",
       "Regressor parameters: {'bootstrap': True, 'ccp_alpha': 0.0, 'criterion': 'squared_error', 'max_depth': None, 'max_features': 1.0, 'max_leaf_nodes': None, 'max_samples': None, 'min_impurity_decrease': 0.0, 'min_samples_leaf': 1, 'min_samples_split': 2, 'min_weight_fraction_leaf': 0.0, 'n_estimators': 100, 'n_jobs': None, 'oob_score': False, 'random_state': 123, 'verbose': 0, 'warm_start': False} \n",
       "Creation date: 2023-04-11 17:20:41 \n",
       "Last fit date: 2023-04-11 17:20:41 \n",
       "Skforecast version: 0.7.0 \n",
       "Python version: 3.10.5 \n",
       "Forecaster id: None "
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Crear y entrenar forecaster\n",
    "# ==============================================================================\n",
    "forecaster = ForecasterAutoreg(\n",
    "                regressor = RandomForestRegressor(random_state=123),\n",
    "                lags = 7\n",
    "             )\n",
    "\n",
    "forecaster.fit(y=datos_train)\n",
    "forecaster"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "21cae648",
   "metadata": {},
   "source": [
    "### Predicciones\n",
    "\n",
    "Una vez entrenado el modelo, se predicen los datos de test (5 años a futuro)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "f09f9082",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2016-01-01    7140.58\n",
       "2017-01-01    7140.58\n",
       "2018-01-01    7140.58\n",
       "2019-01-01    7140.58\n",
       "2020-01-01    7140.58\n",
       "2021-01-01    7140.58\n",
       "2022-01-01    7140.58\n",
       "Freq: AS-JAN, Name: pred, dtype: float64"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "steps = 7\n",
    "predicciones = forecaster.predict(steps=steps)\n",
    "predicciones.head(7)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "5e846c71",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Gráfico\n",
    "fig, ax = plt.subplots(figsize=(10, 6))\n",
    "datos_train.plot(ax=ax, label='train')\n",
    "datos_test.plot(ax=ax, label='test')\n",
    "predicciones.plot(ax=ax, label='predicciones')\n",
    "ax.legend();"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "87e2eb62",
   "metadata": {},
   "source": [
    "### Error de las predicciones en el conjunto de test\n",
    "\n",
    "Se cuantifica el error que comete el modelo en sus predicciones. En este caso, se emplea como métrica el mean squared error (mse)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "b519ed48",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Error de test (mse): 3256939.250685714\n"
     ]
    }
   ],
   "source": [
    "# Error test\n",
    "\n",
    "error_mse = mean_squared_error(\n",
    "                y_true = datos_test,\n",
    "                y_pred = predicciones\n",
    "            )\n",
    "\n",
    "print(f\"Error de test (mse): {error_mse}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b79906e4",
   "metadata": {},
   "source": [
    "### Ajuste de hiperparámetros (tuning)\n",
    "\n",
    "El ForecasterAutoreg entrenado ha utilizado una ventana temporal de 6 lags y un modelo Random Forest con los hiperparámetros por defecto. Sin embargo, no hay ninguna razón por la que estos valores sean los más adecuados. Para identificar la mejor combinación de lags e hiperparámetros, la librería Skforecast dispone de la función grid_search_forecaster con la que comparar los resultados obtenidos con cada configuración del modelo."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "2bf4c609",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of models compared: 12.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "loop lags_grid:   0%|                                               | 0/2 [00:00<?, ?it/s]\n",
      "loop param_grid:   0%|                                              | 0/6 [00:00<?, ?it/s]\u001b[A\n",
      "loop param_grid:  17%|██████▎                               | 1/6 [00:00<00:00,  5.70it/s]\u001b[A\n",
      "loop param_grid:  33%|████████████▋                         | 2/6 [00:01<00:02,  1.76it/s]\u001b[A\n",
      "loop param_grid:  50%|███████████████████                   | 3/6 [00:01<00:01,  2.58it/s]\u001b[A\n",
      "loop param_grid:  67%|█████████████████████████▎            | 4/6 [00:02<00:01,  1.80it/s]\u001b[A\n",
      "loop param_grid:  83%|███████████████████████████████▋      | 5/6 [00:02<00:00,  2.43it/s]\u001b[A\n",
      "loop param_grid: 100%|██████████████████████████████████████| 6/6 [00:02<00:00,  1.81it/s]\u001b[A\n",
      "loop lags_grid:  50%|███████████████████▌                   | 1/2 [00:02<00:02,  3.00s/it]\u001b[A\n",
      "loop param_grid:   0%|                                              | 0/6 [00:00<?, ?it/s]\u001b[A\n",
      "loop param_grid:  17%|██████▎                               | 1/6 [00:00<00:00,  6.11it/s]\u001b[A\n",
      "loop param_grid:  33%|████████████▋                         | 2/6 [00:00<00:02,  1.98it/s]\u001b[A\n",
      "loop param_grid:  50%|███████████████████                   | 3/6 [00:01<00:01,  2.90it/s]\u001b[A\n",
      "loop param_grid:  67%|█████████████████████████▎            | 4/6 [00:01<00:00,  2.02it/s]\u001b[A\n",
      "loop param_grid:  83%|███████████████████████████████▋      | 5/6 [00:01<00:00,  2.63it/s]\u001b[A\n",
      "loop param_grid: 100%|██████████████████████████████████████| 6/6 [00:02<00:00,  1.99it/s]\u001b[A\n",
      "loop lags_grid: 100%|███████████████████████████████████████| 2/2 [00:05<00:00,  2.86s/it]\u001b[A\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "`Forecaster` refitted using the best-found lags and parameters, and the whole data set: \n",
      "  Lags: [1 2 3 4 5] \n",
      "  Parameters: {'max_depth': 3, 'n_estimators': 500}\n",
      "  Backtesting metric: 113.06565112112357\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Grid search de hiperparámetros\n",
    "\n",
    "steps = 7\n",
    "forecaster = ForecasterAutoreg(\n",
    "                regressor = RandomForestRegressor(random_state=123),\n",
    "                lags      = 7 # Este valor será remplazado en el grid search\n",
    "             )\n",
    "\n",
    "# Lags utilizados como predictores\n",
    "lags_grid = [5,10]\n",
    "\n",
    "# Hiperparámetros del regresor\n",
    "param_grid = {'n_estimators': [100, 500],\n",
    "              'max_depth': [3, 5, 7]}\n",
    "\n",
    "resultados_grid = grid_search_forecaster(\n",
    "                        forecaster         = forecaster,\n",
    "                        y                  = datos_train,\n",
    "                        param_grid         = param_grid,\n",
    "                        lags_grid          = lags_grid,\n",
    "                        steps              = steps,\n",
    "                        refit              = True,\n",
    "                        metric             = 'mean_squared_error',\n",
    "                        initial_train_size = 15,\n",
    "                        fixed_train_size   = False,\n",
    "                        return_best        = True,\n",
    "                        verbose            = False\n",
    "                  )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "0d858489",
   "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>lags</th>\n",
       "      <th>params</th>\n",
       "      <th>mean_squared_error</th>\n",
       "      <th>max_depth</th>\n",
       "      <th>n_estimators</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>[1, 2, 3, 4, 5]</td>\n",
       "      <td>{'max_depth': 5, 'n_estimators': 100}</td>\n",
       "      <td>882.090000</td>\n",
       "      <td>5</td>\n",
       "      <td>100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>[1, 2, 3, 4, 5]</td>\n",
       "      <td>{'max_depth': 7, 'n_estimators': 100}</td>\n",
       "      <td>882.090000</td>\n",
       "      <td>7</td>\n",
       "      <td>100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>[1, 2, 3, 4, 5]</td>\n",
       "      <td>{'max_depth': 5, 'n_estimators': 500}</td>\n",
       "      <td>949.995684</td>\n",
       "      <td>5</td>\n",
       "      <td>500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>[1, 2, 3, 4, 5]</td>\n",
       "      <td>{'max_depth': 7, 'n_estimators': 500}</td>\n",
       "      <td>949.995684</td>\n",
       "      <td>7</td>\n",
       "      <td>500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>[1, 2, 3, 4, 5, 6, 7, 8, 9, 10]</td>\n",
       "      <td>{'max_depth': 3, 'n_estimators': 500}</td>\n",
       "      <td>1001.849104</td>\n",
       "      <td>3</td>\n",
       "      <td>500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>[1, 2, 3, 4, 5, 6, 7, 8, 9, 10]</td>\n",
       "      <td>{'max_depth': 5, 'n_estimators': 500}</td>\n",
       "      <td>1001.849104</td>\n",
       "      <td>5</td>\n",
       "      <td>500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>[1, 2, 3, 4, 5, 6, 7, 8, 9, 10]</td>\n",
       "      <td>{'max_depth': 7, 'n_estimators': 500}</td>\n",
       "      <td>1001.849104</td>\n",
       "      <td>7</td>\n",
       "      <td>500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>[1, 2, 3, 4, 5, 6, 7, 8, 9, 10]</td>\n",
       "      <td>{'max_depth': 3, 'n_estimators': 100}</td>\n",
       "      <td>1017.610000</td>\n",
       "      <td>3</td>\n",
       "      <td>100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>[1, 2, 3, 4, 5, 6, 7, 8, 9, 10]</td>\n",
       "      <td>{'max_depth': 5, 'n_estimators': 100}</td>\n",
       "      <td>1017.610000</td>\n",
       "      <td>5</td>\n",
       "      <td>100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>[1, 2, 3, 4, 5, 6, 7, 8, 9, 10]</td>\n",
       "      <td>{'max_depth': 7, 'n_estimators': 100}</td>\n",
       "      <td>1017.610000</td>\n",
       "      <td>7</td>\n",
       "      <td>100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>[1, 2, 3, 4, 5]</td>\n",
       "      <td>{'max_depth': 3, 'n_estimators': 500}</td>\n",
       "      <td>1239.683070</td>\n",
       "      <td>3</td>\n",
       "      <td>500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>[1, 2, 3, 4, 5]</td>\n",
       "      <td>{'max_depth': 3, 'n_estimators': 100}</td>\n",
       "      <td>1324.377664</td>\n",
       "      <td>3</td>\n",
       "      <td>100</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                               lags                                 params  \\\n",
       "2                   [1, 2, 3, 4, 5]  {'max_depth': 5, 'n_estimators': 100}   \n",
       "4                   [1, 2, 3, 4, 5]  {'max_depth': 7, 'n_estimators': 100}   \n",
       "3                   [1, 2, 3, 4, 5]  {'max_depth': 5, 'n_estimators': 500}   \n",
       "5                   [1, 2, 3, 4, 5]  {'max_depth': 7, 'n_estimators': 500}   \n",
       "7   [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]  {'max_depth': 3, 'n_estimators': 500}   \n",
       "9   [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]  {'max_depth': 5, 'n_estimators': 500}   \n",
       "11  [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]  {'max_depth': 7, 'n_estimators': 500}   \n",
       "6   [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]  {'max_depth': 3, 'n_estimators': 100}   \n",
       "8   [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]  {'max_depth': 5, 'n_estimators': 100}   \n",
       "10  [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]  {'max_depth': 7, 'n_estimators': 100}   \n",
       "1                   [1, 2, 3, 4, 5]  {'max_depth': 3, 'n_estimators': 500}   \n",
       "0                   [1, 2, 3, 4, 5]  {'max_depth': 3, 'n_estimators': 100}   \n",
       "\n",
       "    mean_squared_error  max_depth  n_estimators  \n",
       "2           882.090000          5           100  \n",
       "4           882.090000          7           100  \n",
       "3           949.995684          5           500  \n",
       "5           949.995684          7           500  \n",
       "7          1001.849104          3           500  \n",
       "9          1001.849104          5           500  \n",
       "11         1001.849104          7           500  \n",
       "6          1017.610000          3           100  \n",
       "8          1017.610000          5           100  \n",
       "10         1017.610000          7           100  \n",
       "1          1239.683070          3           500  \n",
       "0          1324.377664          3           100  "
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Resultados Grid Search\n",
    "\n",
    "resultados_grid"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a089cf30",
   "metadata": {},
   "source": [
    "Los mejores resultados se obtienen si se utiliza una ventana temporal de 5 lags y una configuración de Random Forest {'max_depth': 3, 'n_estimators': 500}."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b91422f3",
   "metadata": {},
   "source": [
    "### Modelo final\n",
    "\n",
    "Finalmente, se entrena de nuevo un ForecasterAutoreg con la configuración óptima encontrada mediante validación. Este paso no es necesario si se indica return_best = True en la función grid_search_forecaster."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "056fe98b",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Crear y entrenar forecaster con mejores hiperparámetros\n",
    "\n",
    "regressor = RandomForestRegressor(max_depth=3, n_estimators=500, random_state=123)\n",
    "forecaster = ForecasterAutoreg(\n",
    "                regressor = regressor,\n",
    "                lags      = 5\n",
    "             )\n",
    "\n",
    "forecaster.fit(y=datos_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "9592c584",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Predicciones\n",
    "predicciones = forecaster.predict(steps=steps)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "6845e75e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Gráfico\n",
    "fig, ax = plt.subplots(figsize=(10, 6))\n",
    "datos_train.plot(ax=ax, label='train')\n",
    "datos_test.plot(ax=ax, label='test')\n",
    "predicciones.plot(ax=ax, label='predicciones')\n",
    "ax.legend();"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "13dd830b",
   "metadata": {},
   "source": [
    "### Error del Test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "f186f120",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Error de test (mse) 33859.10939047616\n"
     ]
    }
   ],
   "source": [
    "error_mse = mean_squared_error(\n",
    "                y_true = datos_test,\n",
    "                y_pred = predicciones\n",
    "            )\n",
    "\n",
    "print(f\"Error de test (mse) {error_mse}\")"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}