ENSAMBLE_CON_TEXT.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "from xgboost import XGBRegressor\n",
    "from sklearn.ensemble import RandomForestRegressor\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "train = pd.read_csv('normalized2_train.csv', index_col='id', parse_dates=['fecha'])\n",
    "test = pd.read_csv('normalized2_test.csv', index_col='id', parse_dates=['fecha'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Casteamos todo a int, esto es porque XgBoost dio problemas usando el set de datos con los datos heredados.\n",
    "const = ['antiguedad', 'banos', 'garages', 'metroscubiertos', 'metrostotales', 'metros']\n",
    "every_column_except_y= [col for col in train.columns if col not in ['preciomt2', 'precio','id', 'logpreciomt2', 'fecha', 'descripcion', 'titulo', 'logprecio']]\n",
    "columns = [col for col in train[every_column_except_y].columns if col not in const]\n",
    "# columns\n",
    "train.loc[:, columns] = train[columns].fillna(0)\n",
    "for column in columns:\n",
    "    train[column] = train[column].astype('int32')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "xgb = XGBRegressor(objective ='reg:squarederror',\\\n",
    "                   colsample_bytree = 0.8,\\\n",
    "                   learning_rate = 0.1,\\\n",
    "                   max_depth = 10,\\\n",
    "                   reg_alpha = 1.2,\\\n",
    "                   n_estimators = 1000,\\\n",
    "                   reg_lambda = 1.4,\\\n",
    "                   subsample=0.8, seed=30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "X, y = train.loc[:,every_column_except_y], train['logprecio']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/ggparente95/.local/lib/python3.7/site-packages/xgboost/core.py:587: FutureWarning: Series.base is deprecated and will be removed in a future version\n",
      "  if getattr(data, 'base', None) is not None and \\\n",
      "/home/ggparente95/.local/lib/python3.7/site-packages/xgboost/core.py:588: FutureWarning: Series.base is deprecated and will be removed in a future version\n",
      "  data.base is not None and isinstance(data, np.ndarray) \\\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "XGBRegressor(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
       "             colsample_bynode=1, colsample_bytree=0.8, gamma=0,\n",
       "             importance_type='gain', learning_rate=0.1, max_delta_step=0,\n",
       "             max_depth=10, min_child_weight=1, missing=None, n_estimators=1000,\n",
       "             n_jobs=1, nthread=None, objective='reg:squarederror',\n",
       "             random_state=0, reg_alpha=1.2, reg_lambda=1.4, scale_pos_weight=1,\n",
       "             seed=30, silent=None, subsample=0.8, verbosity=1)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "xgb.fit(train[every_column_except_y], train['logpreciomt2'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "pred = xgb.predict(test[every_column_except_y])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "final = pd.DataFrame(test.index)\n",
    "final.loc[:, 'target'] = np.array(test.metros) * np.expm1(pred)\n",
    "pred_final = final.set_index('id')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([14.115986, 11.898907, 12.54506 , ..., 12.299816, 11.771861,\n",
       "       12.604658], dtype=float32)"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "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>target</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>id</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>4941</th>\n",
       "      <td>7.707450e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51775</th>\n",
       "      <td>1.241419e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>115253</th>\n",
       "      <td>2.552291e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>299321</th>\n",
       "      <td>1.422726e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>173570</th>\n",
       "      <td>5.318680e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75094</th>\n",
       "      <td>3.527576e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>171847</th>\n",
       "      <td>7.737907e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>138313</th>\n",
       "      <td>1.018040e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>271268</th>\n",
       "      <td>1.276357e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>72612</th>\n",
       "      <td>3.048480e+06</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>60000 rows × 1 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "              target\n",
       "id                  \n",
       "4941    7.707450e+06\n",
       "51775   1.241419e+06\n",
       "115253  2.552291e+06\n",
       "299321  1.422726e+06\n",
       "173570  5.318680e+05\n",
       "...              ...\n",
       "75094   3.527576e+06\n",
       "171847  7.737907e+05\n",
       "138313  1.018040e+06\n",
       "271268  1.276357e+06\n",
       "72612   3.048480e+06\n",
       "\n",
       "[60000 rows x 1 columns]"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_final"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestRegressor(bootstrap=False, criterion='mse', max_depth=60,\n",
       "                      max_features='sqrt', max_leaf_nodes=None,\n",
       "                      min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "                      min_samples_leaf=2, min_samples_split=5,\n",
       "                      min_weight_fraction_leaf=0.0, n_estimators=1000,\n",
       "                      n_jobs=None, oob_score=False, random_state=None,\n",
       "                      verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rf = RandomForestRegressor(n_estimators=1000, min_samples_split=5, min_samples_leaf=2,\\\n",
    "                           max_features='sqrt', max_depth=60, bootstrap= False)\n",
    "\n",
    "# aca deberia armar las variables X_train e y_train, llenando los nans de df_train\n",
    "\n",
    "train.fillna(0, inplace=True)\n",
    "\n",
    "rf.fit(train[every_column_except_y], train['logpreciomt2'])\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "preds = rf.predict(test[every_column_except_y].fillna(0))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "#preds = rf.predict(test[every_column_except_y])\n",
    "final = pd.DataFrame(test.index)\n",
    "final.loc[:, 'target'] = np.array(test.metros) * np.expm1(pred)\n",
    "pred_final_rf = final.set_index('id')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "pred_total = (0.6*pred_final+0.4*pred_final_rf)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>target</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>id</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>4941</th>\n",
       "      <td>7.707450e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51775</th>\n",
       "      <td>1.241419e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>115253</th>\n",
       "      <td>2.552291e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>299321</th>\n",
       "      <td>1.422726e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>173570</th>\n",
       "      <td>5.318680e+05</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              target\n",
       "id                  \n",
       "4941    7.707450e+06\n",
       "51775   1.241419e+06\n",
       "115253  2.552291e+06\n",
       "299321  1.422726e+06\n",
       "173570  5.318680e+05"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "res = pd.DataFrame(pred_total, index=test.index, columns=['target'])\n",
    "display(res.head())\n",
    "res.to_csv(\"submission-rf-xgb-text.csv\", header=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# SCORE: 542"
   ]
  },
  {
   "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": "3.7.5rc1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}