Wine Prediction Kaggle Competition Solution Group 5.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "___\n",
    "\n",
    "<a href='https://aims.edu.gh/'> <img src='http://gh.nexteinstein.org/wp-content/uploads/sites/15/2016/09/logo.jpg' /></a>\n",
    "___\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Import the needed libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 146,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/lib/python3/dist-packages/ipykernel_launcher.py:9: FutureWarning: Sorting because non-concatenation axis is not aligned. A future version\n",
      "of pandas will change to not sort by default.\n",
      "\n",
      "To accept the future behavior, pass 'sort=False'.\n",
      "\n",
      "To retain the current behavior and silence the warning, pass 'sort=True'.\n",
      "\n",
      "  if __name__ == '__main__':\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "train_df=pd.read_csv('train.csv') #we use this dataset to train our model\n",
    "test_df=pd.read_csv('test.csv') #we will use this data set later to validate our model\n",
    "\n",
    "#Combined both training and test\n",
    "Combined_df=pd.concat([train_df,test_df])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 147,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The shape of the train dataset is: (175000, 14)\n",
      "The shape of the test dataset is: (83210, 15)\n",
      "The shape of the combined dataset is: (258210, 15)\n"
     ]
    }
   ],
   "source": [
    "# Confirm you have the right data and the combination is fine.\n",
    "print('The shape of the train dataset is:',train_df.shape)\n",
    "print('The shape of the test dataset is:',test_df.shape)\n",
    "print('The shape of the combined dataset is:',Combined_df.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 148,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Int64Index: 258210 entries, 0 to 83209\n",
      "Data columns (total 15 columns):\n",
      "country                  258146 non-null object\n",
      "description              258210 non-null object\n",
      "designation              181120 non-null object\n",
      "id                       258210 non-null int64\n",
      "index                    83210 non-null float64\n",
      "points                   258210 non-null float64\n",
      "price                    175000 non-null float64\n",
      "province                 258146 non-null object\n",
      "region_1                 215793 non-null object\n",
      "region_2                 110996 non-null object\n",
      "taster_name              96479 non-null object\n",
      "taster_twitter_handle    91559 non-null object\n",
      "title                    120975 non-null object\n",
      "variety                  258209 non-null object\n",
      "winery                   258210 non-null object\n",
      "dtypes: float64(3), int64(1), object(11)\n",
      "memory usage: 31.5+ MB\n"
     ]
    }
   ],
   "source": [
    "Combined_df.info() "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 149,
   "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>country</th>\n",
       "      <th>description</th>\n",
       "      <th>designation</th>\n",
       "      <th>id</th>\n",
       "      <th>index</th>\n",
       "      <th>points</th>\n",
       "      <th>price</th>\n",
       "      <th>province</th>\n",
       "      <th>region_1</th>\n",
       "      <th>region_2</th>\n",
       "      <th>taster_name</th>\n",
       "      <th>taster_twitter_handle</th>\n",
       "      <th>title</th>\n",
       "      <th>variety</th>\n",
       "      <th>winery</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Portugal</td>\n",
       "      <td>This is a fine rich balanced wine. It has ripe...</td>\n",
       "      <td>Vila Santa Reserva</td>\n",
       "      <td>32027</td>\n",
       "      <td>NaN</td>\n",
       "      <td>88.870874</td>\n",
       "      <td>20.0</td>\n",
       "      <td>Alentejano</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>PORTUGUESE RED</td>\n",
       "      <td>J. Portugal Ramos</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>France</td>\n",
       "      <td>A solid, chunky wine, with a structure that is...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>71079</td>\n",
       "      <td>NaN</td>\n",
       "      <td>88.041695</td>\n",
       "      <td>28.0</td>\n",
       "      <td>Bordeaux</td>\n",
       "      <td>Lalande de Pomerol</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BORDEAUX-STYLE RED BLEND</td>\n",
       "      <td>Château Tour Grand Colombier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>France</td>\n",
       "      <td>This is powerful and concentrated, with the hi...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>32440</td>\n",
       "      <td>NaN</td>\n",
       "      <td>94.085021</td>\n",
       "      <td>130.0</td>\n",
       "      <td>Bordeaux</td>\n",
       "      <td>Saint-Émilion</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BORDEAUX-STYLE RED BLEND</td>\n",
       "      <td>Château Figeac</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>US</td>\n",
       "      <td>Rich, ripe and oaky, this Petite Sirah charms ...</td>\n",
       "      <td>Thompson Vineyard</td>\n",
       "      <td>124405</td>\n",
       "      <td>NaN</td>\n",
       "      <td>89.869797</td>\n",
       "      <td>34.0</td>\n",
       "      <td>California</td>\n",
       "      <td>Santa Barbara County</td>\n",
       "      <td>Central Coast</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Jaffurs 2010 Thompson Vineyard Petite Sirah (S...</td>\n",
       "      <td>PETITE SIRAH</td>\n",
       "      <td>Jaffurs</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>US</td>\n",
       "      <td>This wine is a unique in the state blend and f...</td>\n",
       "      <td>McKinley Springs Vineyard</td>\n",
       "      <td>33649</td>\n",
       "      <td>NaN</td>\n",
       "      <td>89.017651</td>\n",
       "      <td>24.0</td>\n",
       "      <td>Washington</td>\n",
       "      <td>Horse Heaven Hills</td>\n",
       "      <td>Columbia Valley</td>\n",
       "      <td>Sean P. Sullivan</td>\n",
       "      <td>@wawinereport</td>\n",
       "      <td>Syncline 2016 McKinley Springs Vineyard Rosé (...</td>\n",
       "      <td>ROSé</td>\n",
       "      <td>Syncline</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    country                                        description  \\\n",
       "0  Portugal  This is a fine rich balanced wine. It has ripe...   \n",
       "1    France  A solid, chunky wine, with a structure that is...   \n",
       "2    France  This is powerful and concentrated, with the hi...   \n",
       "3        US  Rich, ripe and oaky, this Petite Sirah charms ...   \n",
       "4        US  This wine is a unique in the state blend and f...   \n",
       "\n",
       "                 designation      id  index     points  price    province  \\\n",
       "0         Vila Santa Reserva   32027    NaN  88.870874   20.0  Alentejano   \n",
       "1                        NaN   71079    NaN  88.041695   28.0    Bordeaux   \n",
       "2                        NaN   32440    NaN  94.085021  130.0    Bordeaux   \n",
       "3          Thompson Vineyard  124405    NaN  89.869797   34.0  California   \n",
       "4  McKinley Springs Vineyard   33649    NaN  89.017651   24.0  Washington   \n",
       "\n",
       "               region_1         region_2       taster_name  \\\n",
       "0                   NaN              NaN               NaN   \n",
       "1    Lalande de Pomerol              NaN               NaN   \n",
       "2         Saint-Émilion              NaN               NaN   \n",
       "3  Santa Barbara County    Central Coast               NaN   \n",
       "4    Horse Heaven Hills  Columbia Valley  Sean P. Sullivan   \n",
       "\n",
       "  taster_twitter_handle                                              title  \\\n",
       "0                   NaN                                                NaN   \n",
       "1                   NaN                                                NaN   \n",
       "2                   NaN                                                NaN   \n",
       "3                   NaN  Jaffurs 2010 Thompson Vineyard Petite Sirah (S...   \n",
       "4         @wawinereport  Syncline 2016 McKinley Springs Vineyard Rosé (...   \n",
       "\n",
       "                    variety                        winery  \n",
       "0            PORTUGUESE RED             J. Portugal Ramos  \n",
       "1  BORDEAUX-STYLE RED BLEND  Château Tour Grand Colombier  \n",
       "2  BORDEAUX-STYLE RED BLEND                Château Figeac  \n",
       "3              PETITE SIRAH                       Jaffurs  \n",
       "4                      ROSé                      Syncline  "
      ]
     },
     "execution_count": 149,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Combined_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 150,
   "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>points</th>\n",
       "      <th>price</th>\n",
       "      <th>id</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>175000.000000</td>\n",
       "      <td>175000.000000</td>\n",
       "      <td>175000.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>88.083987</td>\n",
       "      <td>34.304400</td>\n",
       "      <td>70684.047240</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>3.157001</td>\n",
       "      <td>38.398146</td>\n",
       "      <td>41341.638798</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>79.636128</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>85.971283</td>\n",
       "      <td>16.000000</td>\n",
       "      <td>35020.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>87.981631</td>\n",
       "      <td>25.000000</td>\n",
       "      <td>70256.500000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>90.085631</td>\n",
       "      <td>40.000000</td>\n",
       "      <td>105550.250000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>100.220603</td>\n",
       "      <td>2500.000000</td>\n",
       "      <td>150929.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              points          price             id\n",
       "count  175000.000000  175000.000000  175000.000000\n",
       "mean       88.083987      34.304400   70684.047240\n",
       "std         3.157001      38.398146   41341.638798\n",
       "min        79.636128       4.000000       1.000000\n",
       "25%        85.971283      16.000000   35020.000000\n",
       "50%        87.981631      25.000000   70256.500000\n",
       "75%        90.085631      40.000000  105550.250000\n",
       "max       100.220603    2500.000000  150929.000000"
      ]
     },
     "execution_count": 150,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_df.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Missing Values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 151,
   "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>Total</th>\n",
       "      <th>Percent</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>index</th>\n",
       "      <td>175000</td>\n",
       "      <td>67.774292</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>taster_twitter_handle</th>\n",
       "      <td>166651</td>\n",
       "      <td>64.540878</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>taster_name</th>\n",
       "      <td>161731</td>\n",
       "      <td>62.635452</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>region_2</th>\n",
       "      <td>147214</td>\n",
       "      <td>57.013284</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>title</th>\n",
       "      <td>137235</td>\n",
       "      <td>53.148600</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>price</th>\n",
       "      <td>83210</td>\n",
       "      <td>32.225708</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>designation</th>\n",
       "      <td>77090</td>\n",
       "      <td>29.855544</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>region_1</th>\n",
       "      <td>42417</td>\n",
       "      <td>16.427327</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>province</th>\n",
       "      <td>64</td>\n",
       "      <td>0.024786</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>country</th>\n",
       "      <td>64</td>\n",
       "      <td>0.024786</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>variety</th>\n",
       "      <td>1</td>\n",
       "      <td>0.000387</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>winery</th>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>points</th>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>id</th>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>description</th>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                        Total    Percent\n",
       "index                  175000  67.774292\n",
       "taster_twitter_handle  166651  64.540878\n",
       "taster_name            161731  62.635452\n",
       "region_2               147214  57.013284\n",
       "title                  137235  53.148600\n",
       "price                   83210  32.225708\n",
       "designation             77090  29.855544\n",
       "region_1                42417  16.427327\n",
       "province                   64   0.024786\n",
       "country                    64   0.024786\n",
       "variety                     1   0.000387\n",
       "winery                      0   0.000000\n",
       "points                      0   0.000000\n",
       "id                          0   0.000000\n",
       "description                 0   0.000000"
      ]
     },
     "execution_count": 151,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "total = Combined_df.isnull().sum().sort_values(ascending = False)\n",
    "percent = (Combined_df.isnull().sum()/Combined_df.isnull().count()*100).sort_values(ascending = False)\n",
    "missing_data  = pd.concat([total, percent], axis=1, keys=['Total', 'Percent'])\n",
    "missing_data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The most missing values are in region, destination, tester name and price columns.\n",
    "\n",
    "I'm worried the most about wines with NaN in price columns. We don't want to predict points for wines which price are undeclared. In this case, the prices with NAs are from the test data. Thus, we won't drop them.\n",
    "\n",
    "Usefulness of other columns will be investigate on the Feature extraction stage. Maybe that NaN values are meaningful for particular columns.."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## EDA"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "What country has the best wine? What country has the worst?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 209,
   "metadata": {},
   "outputs": [],
   "source": [
    "def category_plot(data, cat, col):\n",
    "    avg = data.groupby(cat, as_index = False)[col].mean()\n",
    "    avg.set_index(cat, inplace = True)\n",
    "    best = avg.nlargest(5, col)\n",
    "    worst = avg.nsmallest(5, col)\n",
    "    fig, ax = plt.subplots(1,2,figsize = (12,3))\n",
    "    # Reverse so the plot looks nicer; the \"best\" will be on top\n",
    "    best = best.iloc[::-1]\n",
    "    best.plot(kind = \"barh\", ax= ax[0], color=\"#3F5D7D\",\n",
    "              fontsize = 14, legend = False, title = \"Best \" + cat + \" by \" + col)\n",
    "    worst.plot(kind = \"barh\", ax=ax[1], color=\"#3F5D7D\",\n",
    "               fontsize = 14, legend = False, title = \"Worst \" + cat + \" by \" + col)\n",
    "    ax[0].title.set_size(16) # Set title size\n",
    "    ax[1].title.set_size(16)\n",
    "    plt.tight_layout()\n",
    "    plt.savefig('plot.png')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 198,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "from IPython.display import display_html\n",
    "def display_side_by_side(*args):\n",
    "    html_str=''\n",
    "    for df in args:\n",
    "        html_str+=df.to_html()\n",
    "    display_html(html_str.replace('table','table style=\"display:inline\"'),raw=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 199,
   "metadata": {},
   "outputs": [],
   "source": [
    "def display_dfs(data, cat, col):\n",
    "    avg_df = data.groupby(cat, as_index=False)[col].mean().sort_values(by = col, ascending = False)\n",
    "    avg_df = avg_df.dropna(axis = 0) # control for NaN\n",
    "    # top 5\n",
    "    df1 = avg_df.head()\n",
    "    # bottom 5\n",
    "    df2 = avg_df.tail()\n",
    "    display_side_by_side(df1,df2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Country with the best and worst wine, by points, on average.¶\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 200,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x216 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#display_dfs(train_df, \"country\", \"points\")\n",
    "category_plot(train_df, \"country\", \"points\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Country with the most expensive and cheapest wine, on average.¶\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 206,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x216 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#display_dfs(train_df, \"country\", \"price\")\n",
    "category_plot(train_df, \"country\", \"price\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Price by variety. Maybe some varities are more expensive? And, is the cheapest wine also the worst? Is the best wine also the most expensive?¶\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 210,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x216 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#display_dfs(train_df, \"variety\", \"price\")\n",
    "category_plot(train_df, \"variety\", \"price\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Feature Engineering"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 137,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Convert variables to categories with label encoding\n",
    "Combined_df[\"designation\"] = Combined_df[\"designation\"].astype('category').cat.codes\n",
    "Combined_df[\"province\"] = Combined_df[\"province\"].astype('category').cat.codes\n",
    "Combined_df[\"region_1\"] = Combined_df[\"region_1\"].astype('category').cat.codes\n",
    "Combined_df[\"region_2\"] = Combined_df[\"region_2\"].astype('category').cat.codes\n",
    "Combined_df[\"title\"] = Combined_df[\"title\"].astype('category').cat.codes\n",
    "Combined_df[\"variety\"] = Combined_df[\"variety\"].astype('category').cat.codes\n",
    "Combined_df[\"winery\"] = Combined_df[\"winery\"].astype('category').cat.codes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 138,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Applying one hot encoding in some variables\n",
    "Combined_df = pd.get_dummies(Combined_df, columns = ['country','taster_twitter_handle'], drop_first = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(258210, 819)"
      ]
     },
     "execution_count": 69,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Combined_df.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## TF- IDF"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.feature_extraction.text import TfidfTransformer\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "from sklearn.decomposition import TruncatedSVD"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Train tfidf and svd\n",
    "tf = TfidfVectorizer(stop_words='english')\n",
    "\n",
    "svd = TruncatedSVD(n_components=5)\n",
    "\n",
    "def transform_text(data, col):\n",
    "    #Fit tfidf and svd, and transform training data\n",
    "    tfidf_matrix = tf.fit_transform(data[col])\n",
    "    lsa_features = pd.DataFrame(svd.fit_transform(tfidf_matrix))\n",
    "\n",
    "    #Creat meaningful column names\n",
    "    collist = map(str, range(0, 5))\n",
    "    collist = [\"latent_\" + col + '_' + s for s in collist]\n",
    "    lsa_features.columns = collist\n",
    "    lsa_features = lsa_features.set_index(data.index)\n",
    "    return lsa_features\n",
    "\n",
    "lsa_features = transform_text(Combined_df, 'description') #this is the reduced features from the description column"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "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>latent_description_0</th>\n",
       "      <th>latent_description_1</th>\n",
       "      <th>latent_description_2</th>\n",
       "      <th>latent_description_3</th>\n",
       "      <th>latent_description_4</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.216267</td>\n",
       "      <td>0.009842</td>\n",
       "      <td>0.205410</td>\n",
       "      <td>-0.076283</td>\n",
       "      <td>-0.050708</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.226328</td>\n",
       "      <td>-0.045308</td>\n",
       "      <td>0.120272</td>\n",
       "      <td>-0.011149</td>\n",
       "      <td>-0.083990</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.218478</td>\n",
       "      <td>-0.128571</td>\n",
       "      <td>0.171616</td>\n",
       "      <td>-0.014873</td>\n",
       "      <td>-0.152442</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.149738</td>\n",
       "      <td>-0.088167</td>\n",
       "      <td>0.012866</td>\n",
       "      <td>0.085328</td>\n",
       "      <td>-0.069703</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.144091</td>\n",
       "      <td>0.082410</td>\n",
       "      <td>-0.032871</td>\n",
       "      <td>0.009524</td>\n",
       "      <td>0.007647</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   latent_description_0  latent_description_1  latent_description_2  \\\n",
       "0              0.216267              0.009842              0.205410   \n",
       "1              0.226328             -0.045308              0.120272   \n",
       "2              0.218478             -0.128571              0.171616   \n",
       "3              0.149738             -0.088167              0.012866   \n",
       "4              0.144091              0.082410             -0.032871   \n",
       "\n",
       "   latent_description_3  latent_description_4  \n",
       "0             -0.076283             -0.050708  \n",
       "1             -0.011149             -0.083990  \n",
       "2             -0.014873             -0.152442  \n",
       "3              0.085328             -0.069703  \n",
       "4              0.009524              0.007647  "
      ]
     },
     "execution_count": 72,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lsa_features.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Add description features to dataframe\n",
    "Combined_df_new=pd.concat([Combined_df,lsa_features],axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_df_new=Combined_df_new.iloc[0:len(train_df),]\n",
    "test_df_new=Combined_df_new.iloc[len(train_df):,]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "y=train_df_new[['price']]\n",
    "X=train_df_new.drop(['description', 'price','id','index' ],axis=1)\n",
    "x_test=test_df_new.drop(['description', 'price','id','index' ],axis=1) #xtest for test data---"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {
    "scrolled": false
   },
   "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>designation</th>\n",
       "      <th>points</th>\n",
       "      <th>province</th>\n",
       "      <th>region_1</th>\n",
       "      <th>region_2</th>\n",
       "      <th>taster_name</th>\n",
       "      <th>title</th>\n",
       "      <th>winery</th>\n",
       "      <th>text length</th>\n",
       "      <th>raw_word_len</th>\n",
       "      <th>...</th>\n",
       "      <th>taster_twitter_handle_@vossroger</th>\n",
       "      <th>taster_twitter_handle_@wawinereport</th>\n",
       "      <th>taster_twitter_handle_@wineschach</th>\n",
       "      <th>taster_twitter_handle_@winewchristina</th>\n",
       "      <th>taster_twitter_handle_@worldwineguys</th>\n",
       "      <th>latent_description_0</th>\n",
       "      <th>latent_description_1</th>\n",
       "      <th>latent_description_2</th>\n",
       "      <th>latent_description_3</th>\n",
       "      <th>latent_description_4</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>42041</td>\n",
       "      <td>88.870874</td>\n",
       "      <td>8</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>9470</td>\n",
       "      <td>267</td>\n",
       "      <td>46</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.216267</td>\n",
       "      <td>0.009842</td>\n",
       "      <td>0.205410</td>\n",
       "      <td>-0.076283</td>\n",
       "      <td>-0.050708</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>-1</td>\n",
       "      <td>88.041695</td>\n",
       "      <td>38</td>\n",
       "      <td>582</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>4345</td>\n",
       "      <td>225</td>\n",
       "      <td>41</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.226328</td>\n",
       "      <td>-0.045308</td>\n",
       "      <td>0.120272</td>\n",
       "      <td>-0.011149</td>\n",
       "      <td>-0.083990</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>-1</td>\n",
       "      <td>94.085021</td>\n",
       "      <td>38</td>\n",
       "      <td>983</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>3564</td>\n",
       "      <td>199</td>\n",
       "      <td>33</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.218478</td>\n",
       "      <td>-0.128571</td>\n",
       "      <td>0.171616</td>\n",
       "      <td>-0.014873</td>\n",
       "      <td>-0.152442</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>39632</td>\n",
       "      <td>89.869797</td>\n",
       "      <td>53</td>\n",
       "      <td>1014</td>\n",
       "      <td>1</td>\n",
       "      <td>-1</td>\n",
       "      <td>53573</td>\n",
       "      <td>9532</td>\n",
       "      <td>187</td>\n",
       "      <td>31</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.149738</td>\n",
       "      <td>-0.088167</td>\n",
       "      <td>0.012866</td>\n",
       "      <td>0.085328</td>\n",
       "      <td>-0.069703</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>25325</td>\n",
       "      <td>89.017651</td>\n",
       "      <td>468</td>\n",
       "      <td>536</td>\n",
       "      <td>3</td>\n",
       "      <td>16</td>\n",
       "      <td>95520</td>\n",
       "      <td>15765</td>\n",
       "      <td>290</td>\n",
       "      <td>51</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.144091</td>\n",
       "      <td>0.082410</td>\n",
       "      <td>-0.032871</td>\n",
       "      <td>0.009524</td>\n",
       "      <td>0.007647</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 820 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   designation     points  province  region_1  region_2  taster_name  title  \\\n",
       "0        42041  88.870874         8        -1        -1           -1     -1   \n",
       "1           -1  88.041695        38       582        -1           -1     -1   \n",
       "2           -1  94.085021        38       983        -1           -1     -1   \n",
       "3        39632  89.869797        53      1014         1           -1  53573   \n",
       "4        25325  89.017651       468       536         3           16  95520   \n",
       "\n",
       "   winery  text length  raw_word_len          ...           \\\n",
       "0    9470          267            46          ...            \n",
       "1    4345          225            41          ...            \n",
       "2    3564          199            33          ...            \n",
       "3    9532          187            31          ...            \n",
       "4   15765          290            51          ...            \n",
       "\n",
       "   taster_twitter_handle_@vossroger  taster_twitter_handle_@wawinereport  \\\n",
       "0                                 0                                    0   \n",
       "1                                 0                                    0   \n",
       "2                                 0                                    0   \n",
       "3                                 0                                    0   \n",
       "4                                 0                                    1   \n",
       "\n",
       "   taster_twitter_handle_@wineschach  taster_twitter_handle_@winewchristina  \\\n",
       "0                                  0                                      0   \n",
       "1                                  0                                      0   \n",
       "2                                  0                                      0   \n",
       "3                                  0                                      0   \n",
       "4                                  0                                      0   \n",
       "\n",
       "   taster_twitter_handle_@worldwineguys  latent_description_0  \\\n",
       "0                                     0              0.216267   \n",
       "1                                     0              0.226328   \n",
       "2                                     0              0.218478   \n",
       "3                                     0              0.149738   \n",
       "4                                     0              0.144091   \n",
       "\n",
       "   latent_description_1  latent_description_2  latent_description_3  \\\n",
       "0              0.009842              0.205410             -0.076283   \n",
       "1             -0.045308              0.120272             -0.011149   \n",
       "2             -0.128571              0.171616             -0.014873   \n",
       "3             -0.088167              0.012866              0.085328   \n",
       "4              0.082410             -0.032871              0.009524   \n",
       "\n",
       "   latent_description_4  \n",
       "0             -0.050708  \n",
       "1             -0.083990  \n",
       "2             -0.152442  \n",
       "3             -0.069703  \n",
       "4              0.007647  \n",
       "\n",
       "[5 rows x 820 columns]"
      ]
     },
     "execution_count": 76,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Feature Scaling\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/lib/python3/dist-packages/sklearn/preprocessing/data.py:625: DataConversionWarning: Data with input dtype int8, uint8, int16, int32, int64, float64 were all converted to float64 by StandardScaler.\n",
      "  return self.partial_fit(X, y)\n",
      "/usr/lib/python3/dist-packages/sklearn/base.py:462: DataConversionWarning: Data with input dtype int8, uint8, int16, int32, int64, float64 were all converted to float64 by StandardScaler.\n",
      "  return self.fit(X, **fit_params).transform(X)\n",
      "/usr/lib/python3/dist-packages/ipykernel_launcher.py:8: DataConversionWarning: Data with input dtype int8, uint8, int16, int32, int64, float64 were all converted to float64 by StandardScaler.\n",
      "  \n"
     ]
    }
   ],
   "source": [
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)\n",
    "\n",
    "sc = StandardScaler()\n",
    "X_train = pd.DataFrame(sc.fit_transform(X_train))\n",
    "X_test = pd.DataFrame(sc.transform(X_test))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Model Training "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 962,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/lib/python3/dist-packages/ipykernel_launcher.py:4: DataConversionWarning: Data with input dtype int8, uint8, int16, int32, int64, float64 were all converted to float64 by StandardScaler.\n",
      "  after removing the cwd from sys.path.\n"
     ]
    }
   ],
   "source": [
    "x_test=test_df_new.drop(['description', 'price','id','index' ],axis=1)\n",
    "\n",
    "#transform test column\n",
    "x_test = pd.DataFrame(sc.transform(x_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# RandomForest with K - FOLD"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "========Fold 0\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/lib/python3/dist-packages/ipykernel_launcher.py:12: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples,), for example using ravel().\n",
      "  if sys.path[0] == '':\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE : 20.11633145708801\n",
      "========Fold 1\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/lib/python3/dist-packages/ipykernel_launcher.py:12: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples,), for example using ravel().\n",
      "  if sys.path[0] == '':\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE : 26.721061601665458\n",
      "========Fold 2\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/lib/python3/dist-packages/ipykernel_launcher.py:12: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples,), for example using ravel().\n",
      "  if sys.path[0] == '':\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE : 21.448491114961243\n",
      "========Fold 3\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/lib/python3/dist-packages/ipykernel_launcher.py:12: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples,), for example using ravel().\n",
      "  if sys.path[0] == '':\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE : 26.935118397045464\n",
      "========Fold 4\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/lib/python3/dist-packages/ipykernel_launcher.py:12: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples,), for example using ravel().\n",
      "  if sys.path[0] == '':\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE : 18.80061817099184\n",
      "========Fold 5\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/lib/python3/dist-packages/ipykernel_launcher.py:12: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples,), for example using ravel().\n",
      "  if sys.path[0] == '':\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE : 24.708114266948247\n",
      "========Fold 6\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/lib/python3/dist-packages/ipykernel_launcher.py:12: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples,), for example using ravel().\n",
      "  if sys.path[0] == '':\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE : 25.752983401096326\n",
      "========Fold 7\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/lib/python3/dist-packages/ipykernel_launcher.py:12: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples,), for example using ravel().\n",
      "  if sys.path[0] == '':\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE : 21.326349878830314\n",
      "========Fold 8\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/lib/python3/dist-packages/ipykernel_launcher.py:12: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples,), for example using ravel().\n",
      "  if sys.path[0] == '':\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE : 21.584473173358404\n",
      "========Fold 9\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/lib/python3/dist-packages/ipykernel_launcher.py:12: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples,), for example using ravel().\n",
      "  if sys.path[0] == '':\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE : 20.452327618286244\n",
      "CPU times: user 18min 58s, sys: 132 ms, total: 18min 58s\n",
      "Wall time: 18min 58s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "from sklearn.model_selection import KFold\n",
    "from sklearn.ensemble import RandomForestRegressor\n",
    "\n",
    "KF = KFold(n_splits = 10, shuffle = True, random_state=1)\n",
    "preds_val = np.zeros([test_df_new.shape[0],])\n",
    "\n",
    "for i,(train_ind, test_ind) in enumerate(KF.split(X)):\n",
    "    print('========Fold',i)\n",
    "    Xtrain, XCV, ytrain, yCV = X.iloc[train_ind], X.iloc[test_ind], y.values[train_ind], y.values[test_ind]\n",
    "    \n",
    "    regressor = RandomForestRegressor(n_estimators=50, random_state=42)\n",
    "    regressor.fit(Xtrain, ytrain)\n",
    "    preds_val += regressor.predict(x_test)/10\n",
    "\n",
    "    pred = regressor.predict(XCV)\n",
    "    print('RMSE :', np.sqrt(metrics.mean_squared_error(pred, yCV)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "predictions_rfr=pd.DataFrame({'id':range(len(test_df)),'price':preds_val})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "34.68908568681648"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.mean(predictions_rfr['price'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# XGBoost with K - FOLD"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "========Fold 0\n",
      "[11:41:32] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "logloss : 22.001866077893844\n",
      "========Fold 1\n",
      "[11:50:47] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "logloss : 24.469343538760445\n",
      "========Fold 2\n",
      "[11:59:59] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "logloss : 22.48433378295003\n",
      "========Fold 3\n",
      "[12:08:49] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "logloss : 29.12200244903267\n",
      "========Fold 4\n",
      "[12:17:53] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "logloss : 21.551018475285865\n",
      "========Fold 5\n",
      "[12:26:59] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "logloss : 26.106869369719085\n",
      "========Fold 6\n",
      "[12:38:07] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "logloss : 28.291603913340076\n",
      "========Fold 7\n",
      "[12:47:11] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "logloss : 23.00199519060012\n",
      "========Fold 8\n",
      "[12:56:13] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "logloss : 22.95747816968494\n",
      "========Fold 9\n",
      "[13:04:57] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "logloss : 21.604995089036972\n",
      "CPU times: user 1h 29min 16s, sys: 23 s, total: 1h 29min 39s\n",
      "Wall time: 1h 32min 4s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "from sklearn.model_selection import KFold\n",
    "import xgboost as xgb\n",
    "from sklearn import metrics\n",
    "\n",
    "KF = KFold(n_splits = 10, shuffle = True, random_state=1)\n",
    "preds = np.zeros([test_df_new.shape[0],])\n",
    "#K-Fold\n",
    "params = {'eta': 0.4, 'max_depth':6 , 'objective': 'reg:linear',\n",
    "          'eval_metric': 'rmse', 'seed': 99, 'alpha':2}\n",
    "\n",
    "for i,(train_ind, test_ind) in enumerate(KF.split(X)):\n",
    "    print('========Fold',i)\n",
    "    Xtrain, XCV, ytrain, yCV = X.iloc[train_ind], X.iloc[test_ind], y.values[train_ind], y.values[test_ind]\n",
    "    \n",
    "    model = xgb.train(params, xgb.DMatrix(Xtrain, ytrain), 200, maximize=False)\n",
    "\n",
    "    preds += (model.predict(xgb.DMatrix(x_test)))/10\n",
    "\n",
    "    pred = model.predict(xgb.DMatrix(XCV))\n",
    "    print('logloss :', np.sqrt(metrics.mean_squared_error(pred, yCV)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "predictions_xgb=pd.DataFrame({'id':range(len(test_df)),'price':preds})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "34.160519793180775"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.mean(predictions_xgb['price'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "## Combined predictions\n",
    "predsfinal=(predictions_xgb['price']+predictions_rfr['price'])/2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "34.42480273999864"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.mean(predsfinal)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "predsfinal_df=pd.DataFrame({'id':range(len(test_df)),'price':predsfinal})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "predsfinal_df.to_csv('predictions.csv',index=False)"
   ]
  }
 ],
 "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}