Milestone3. Bernoulli Naive Bayes.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import math\n",
    "import re\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from scipy import stats\n",
    "from sklearn.naive_bayes import BernoulliNB\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.model_selection import StratifiedKFold\n",
    "from collections import Counter \n",
    "from sklearn.metrics import plot_confusion_matrix\n",
    "from sklearn.metrics import log_loss\n",
    "\n",
    "CONST_FEATURE = 5\n",
    "CONST_INTERVAL = 50"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "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>bathrooms</th>\n",
       "      <th>bedrooms</th>\n",
       "      <th>building_id</th>\n",
       "      <th>created</th>\n",
       "      <th>description</th>\n",
       "      <th>display_address</th>\n",
       "      <th>features</th>\n",
       "      <th>latitude</th>\n",
       "      <th>listing_id</th>\n",
       "      <th>longitude</th>\n",
       "      <th>manager_id</th>\n",
       "      <th>photos</th>\n",
       "      <th>price</th>\n",
       "      <th>street_address</th>\n",
       "      <th>interest_level</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1</td>\n",
       "      <td>8579a0b0d54db803821a35a4a615e97a</td>\n",
       "      <td>2016-06-16 05:55:27</td>\n",
       "      <td>Spacious 1 Bedroom 1 Bathroom in Williamsburg!...</td>\n",
       "      <td>145 Borinquen Place</td>\n",
       "      <td>[Dining Room, Pre-War, Laundry in Building, Di...</td>\n",
       "      <td>40.7108</td>\n",
       "      <td>7170325</td>\n",
       "      <td>-73.9539</td>\n",
       "      <td>a10db4590843d78c784171a107bdacb4</td>\n",
       "      <td>[https://photos.renthop.com/2/7170325_3bb5ac84...</td>\n",
       "      <td>2400</td>\n",
       "      <td>145 Borinquen Place</td>\n",
       "      <td>medium</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>1.0</td>\n",
       "      <td>2</td>\n",
       "      <td>b8e75fc949a6cd8225b455648a951712</td>\n",
       "      <td>2016-06-01 05:44:33</td>\n",
       "      <td>BRAND NEW GUT RENOVATED TRUE 2 BEDROOMFind you...</td>\n",
       "      <td>East 44th</td>\n",
       "      <td>[Doorman, Elevator, Laundry in Building, Dishw...</td>\n",
       "      <td>40.7513</td>\n",
       "      <td>7092344</td>\n",
       "      <td>-73.9722</td>\n",
       "      <td>955db33477af4f40004820b4aed804a0</td>\n",
       "      <td>[https://photos.renthop.com/2/7092344_7663c19a...</td>\n",
       "      <td>3800</td>\n",
       "      <td>230 East 44th</td>\n",
       "      <td>low</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   bathrooms  bedrooms                       building_id              created  \\\n",
       "4        1.0         1  8579a0b0d54db803821a35a4a615e97a  2016-06-16 05:55:27   \n",
       "6        1.0         2  b8e75fc949a6cd8225b455648a951712  2016-06-01 05:44:33   \n",
       "\n",
       "                                         description      display_address  \\\n",
       "4  Spacious 1 Bedroom 1 Bathroom in Williamsburg!...  145 Borinquen Place   \n",
       "6  BRAND NEW GUT RENOVATED TRUE 2 BEDROOMFind you...            East 44th   \n",
       "\n",
       "                                            features  latitude  listing_id  \\\n",
       "4  [Dining Room, Pre-War, Laundry in Building, Di...   40.7108     7170325   \n",
       "6  [Doorman, Elevator, Laundry in Building, Dishw...   40.7513     7092344   \n",
       "\n",
       "   longitude                        manager_id  \\\n",
       "4   -73.9539  a10db4590843d78c784171a107bdacb4   \n",
       "6   -73.9722  955db33477af4f40004820b4aed804a0   \n",
       "\n",
       "                                              photos  price  \\\n",
       "4  [https://photos.renthop.com/2/7170325_3bb5ac84...   2400   \n",
       "6  [https://photos.renthop.com/2/7092344_7663c19a...   3800   \n",
       "\n",
       "        street_address interest_level  \n",
       "4  145 Borinquen Place         medium  \n",
       "6        230 East 44th            low  "
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.read_json(\"./train.json\")\n",
    "df.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>bathrooms</th>\n",
       "      <th>bedrooms</th>\n",
       "      <th>building_id</th>\n",
       "      <th>created</th>\n",
       "      <th>description</th>\n",
       "      <th>display_address</th>\n",
       "      <th>features</th>\n",
       "      <th>latitude</th>\n",
       "      <th>listing_id</th>\n",
       "      <th>longitude</th>\n",
       "      <th>manager_id</th>\n",
       "      <th>photos</th>\n",
       "      <th>price</th>\n",
       "      <th>street_address</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1</td>\n",
       "      <td>79780be1514f645d7e6be99a3de696c5</td>\n",
       "      <td>2016-06-11 05:29:41</td>\n",
       "      <td>Large with awesome terrace--accessible via bed...</td>\n",
       "      <td>Suffolk Street</td>\n",
       "      <td>[Elevator, Laundry in Building, Laundry in Uni...</td>\n",
       "      <td>40.7185</td>\n",
       "      <td>7142618</td>\n",
       "      <td>-73.9865</td>\n",
       "      <td>b1b1852c416d78d7765d746cb1b8921f</td>\n",
       "      <td>[https://photos.renthop.com/2/7142618_1c45a2c8...</td>\n",
       "      <td>2950</td>\n",
       "      <td>99 Suffolk Street</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.0</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>2016-06-24 06:36:34</td>\n",
       "      <td>Prime Soho - between Bleecker and Houston - Ne...</td>\n",
       "      <td>Thompson Street</td>\n",
       "      <td>[Pre-War, Dogs Allowed, Cats Allowed]</td>\n",
       "      <td>40.7278</td>\n",
       "      <td>7210040</td>\n",
       "      <td>-74.0000</td>\n",
       "      <td>d0b5648017832b2427eeb9956d966a14</td>\n",
       "      <td>[https://photos.renthop.com/2/7210040_d824cc71...</td>\n",
       "      <td>2850</td>\n",
       "      <td>176 Thompson Street</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   bathrooms  bedrooms                       building_id              created  \\\n",
       "0        1.0         1  79780be1514f645d7e6be99a3de696c5  2016-06-11 05:29:41   \n",
       "1        1.0         2                                 0  2016-06-24 06:36:34   \n",
       "\n",
       "                                         description  display_address  \\\n",
       "0  Large with awesome terrace--accessible via bed...   Suffolk Street   \n",
       "1  Prime Soho - between Bleecker and Houston - Ne...  Thompson Street   \n",
       "\n",
       "                                            features  latitude  listing_id  \\\n",
       "0  [Elevator, Laundry in Building, Laundry in Uni...   40.7185     7142618   \n",
       "1              [Pre-War, Dogs Allowed, Cats Allowed]   40.7278     7210040   \n",
       "\n",
       "   longitude                        manager_id  \\\n",
       "0   -73.9865  b1b1852c416d78d7765d746cb1b8921f   \n",
       "1   -74.0000  d0b5648017832b2427eeb9956d966a14   \n",
       "\n",
       "                                              photos  price  \\\n",
       "0  [https://photos.renthop.com/2/7142618_1c45a2c8...   2950   \n",
       "1  [https://photos.renthop.com/2/7210040_d824cc71...   2850   \n",
       "\n",
       "        street_address  \n",
       "0    99 Suffolk Street  \n",
       "1  176 Thompson Street  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df2 = pd.read_json(\"./test.json\")\n",
    "df2.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(40729, 15)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#handle outliners\n",
    "#bathroom\n",
    "df = df.loc[ (df['bathrooms'] > 0) & (df['bathrooms'] != 10)]\n",
    "z_bath = np.abs(stats.zscore(df['bathrooms']))\n",
    "df = df[(z_bath < 3)]\n",
    "\n",
    "#bedroom\n",
    "z_bed = np.abs(stats.zscore(df['bedrooms']))\n",
    "df = df[(z_bed < 3)]\n",
    "\n",
    "#latitude\n",
    "df = df.loc[\n",
    "    (df['latitude'] >= 19.50139) & \n",
    "    (df['latitude'] <= 64.85694)\n",
    "    ]\n",
    "z_lati = np.abs(stats.zscore(df['latitude']))\n",
    "df = df[(z_lati < 3)]\n",
    "\n",
    "#longitude\n",
    "df = df.loc[\n",
    "    (df['longitude'] >= -161.75583) & \n",
    "    (df['longitude'] <= -68.01197)\n",
    "    ]\n",
    "z_long = np.abs(stats.zscore(df['longitude']))\n",
    "df = df[(z_long < 3)]\n",
    "\n",
    "#price\n",
    "z_price = np.abs(stats.zscore(df['price']))\n",
    "df = df[(z_price < 3)]\n",
    "\n",
    "df = df[(df['description'] != '') & (df['description'] != '<p><a  website_redacted')]\n",
    "df = df[df['display_address'] != '']\n",
    "df = df[df['features'].str.len() > 0]\n",
    "df = df[(df['interest_level'] == 'low') | (df['interest_level'] == 'medium') | (df['interest_level'] == 'high')]\n",
    "\n",
    "df = df[df['latitude'] != 0]\n",
    "df = df[df['longitude'] != 0]\n",
    "df = df[df['photos'].str.len() > 0]\n",
    "df = df[df['street_address'].str.len() > 0]\n",
    "\n",
    "df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "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>bathrooms</th>\n",
       "      <th>bedrooms</th>\n",
       "      <th>building_id</th>\n",
       "      <th>created</th>\n",
       "      <th>description</th>\n",
       "      <th>display_address</th>\n",
       "      <th>features</th>\n",
       "      <th>latitude</th>\n",
       "      <th>listing_id</th>\n",
       "      <th>longitude</th>\n",
       "      <th>manager_id</th>\n",
       "      <th>photos</th>\n",
       "      <th>price_category</th>\n",
       "      <th>price</th>\n",
       "      <th>street_address</th>\n",
       "      <th>interest_level</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1</td>\n",
       "      <td>8579a0b0d54db803821a35a4a615e97a</td>\n",
       "      <td>2016-06-16 05:55:27</td>\n",
       "      <td>Spacious 1 Bedroom 1 Bathroom in Williamsburg!...</td>\n",
       "      <td>145 Borinquen Place</td>\n",
       "      <td>[Dining Room, Pre-War, Laundry in Building, Di...</td>\n",
       "      <td>40.7108</td>\n",
       "      <td>7170325</td>\n",
       "      <td>-73.9539</td>\n",
       "      <td>a10db4590843d78c784171a107bdacb4</td>\n",
       "      <td>[https://photos.renthop.com/2/7170325_3bb5ac84...</td>\n",
       "      <td>6</td>\n",
       "      <td>2400</td>\n",
       "      <td>145 Borinquen Place</td>\n",
       "      <td>medium</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>1.0</td>\n",
       "      <td>2</td>\n",
       "      <td>b8e75fc949a6cd8225b455648a951712</td>\n",
       "      <td>2016-06-01 05:44:33</td>\n",
       "      <td>BRAND NEW GUT RENOVATED TRUE 2 BEDROOMFind you...</td>\n",
       "      <td>East 44th</td>\n",
       "      <td>[Doorman, Elevator, Laundry in Building, Dishw...</td>\n",
       "      <td>40.7513</td>\n",
       "      <td>7092344</td>\n",
       "      <td>-73.9722</td>\n",
       "      <td>955db33477af4f40004820b4aed804a0</td>\n",
       "      <td>[https://photos.renthop.com/2/7092344_7663c19a...</td>\n",
       "      <td>22</td>\n",
       "      <td>3800</td>\n",
       "      <td>230 East 44th</td>\n",
       "      <td>low</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>1.0</td>\n",
       "      <td>2</td>\n",
       "      <td>cd759a988b8f23924b5a2058d5ab2b49</td>\n",
       "      <td>2016-06-14 15:19:59</td>\n",
       "      <td>**FLEX 2 BEDROOM WITH FULL PRESSURIZED WALL**L...</td>\n",
       "      <td>East 56th Street</td>\n",
       "      <td>[Doorman, Elevator, Laundry in Building, Laund...</td>\n",
       "      <td>40.7575</td>\n",
       "      <td>7158677</td>\n",
       "      <td>-73.9625</td>\n",
       "      <td>c8b10a317b766204f08e613cef4ce7a0</td>\n",
       "      <td>[https://photos.renthop.com/2/7158677_c897a134...</td>\n",
       "      <td>19</td>\n",
       "      <td>3495</td>\n",
       "      <td>405 East 56th Street</td>\n",
       "      <td>medium</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>bfb9405149bfff42a92980b594c28234</td>\n",
       "      <td>2016-06-28 03:50:23</td>\n",
       "      <td>Over-sized Studio w abundant closets. Availabl...</td>\n",
       "      <td>East 34th Street</td>\n",
       "      <td>[Doorman, Elevator, Fitness Center, Laundry in...</td>\n",
       "      <td>40.7439</td>\n",
       "      <td>7225292</td>\n",
       "      <td>-73.9743</td>\n",
       "      <td>2c3b41f588fbb5234d8a1e885a436cfa</td>\n",
       "      <td>[https://photos.renthop.com/2/7225292_901f1984...</td>\n",
       "      <td>10</td>\n",
       "      <td>2795</td>\n",
       "      <td>340 East 34th Street</td>\n",
       "      <td>low</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>1.0</td>\n",
       "      <td>3</td>\n",
       "      <td>300d27d8ba2adbcbc8c6f2bcbc1c6f9d</td>\n",
       "      <td>2016-06-28 05:59:06</td>\n",
       "      <td>This spectacular converted 3 bed apartment all...</td>\n",
       "      <td>East 16th Street</td>\n",
       "      <td>[Doorman, Elevator, Loft, Dishwasher, Hardwood...</td>\n",
       "      <td>40.7348</td>\n",
       "      <td>7226687</td>\n",
       "      <td>-73.9865</td>\n",
       "      <td>30a7951cfc21240e8c46b2d079d292e5</td>\n",
       "      <td>[https://photos.renthop.com/2/7226687_4ea22bf1...</td>\n",
       "      <td>43</td>\n",
       "      <td>7200</td>\n",
       "      <td>145 East 16th Street</td>\n",
       "      <td>low</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    bathrooms  bedrooms                       building_id  \\\n",
       "4         1.0         1  8579a0b0d54db803821a35a4a615e97a   \n",
       "6         1.0         2  b8e75fc949a6cd8225b455648a951712   \n",
       "9         1.0         2  cd759a988b8f23924b5a2058d5ab2b49   \n",
       "15        1.0         0  bfb9405149bfff42a92980b594c28234   \n",
       "16        1.0         3  300d27d8ba2adbcbc8c6f2bcbc1c6f9d   \n",
       "\n",
       "                created                                        description  \\\n",
       "4   2016-06-16 05:55:27  Spacious 1 Bedroom 1 Bathroom in Williamsburg!...   \n",
       "6   2016-06-01 05:44:33  BRAND NEW GUT RENOVATED TRUE 2 BEDROOMFind you...   \n",
       "9   2016-06-14 15:19:59  **FLEX 2 BEDROOM WITH FULL PRESSURIZED WALL**L...   \n",
       "15  2016-06-28 03:50:23  Over-sized Studio w abundant closets. Availabl...   \n",
       "16  2016-06-28 05:59:06  This spectacular converted 3 bed apartment all...   \n",
       "\n",
       "        display_address                                           features  \\\n",
       "4   145 Borinquen Place  [Dining Room, Pre-War, Laundry in Building, Di...   \n",
       "6             East 44th  [Doorman, Elevator, Laundry in Building, Dishw...   \n",
       "9      East 56th Street  [Doorman, Elevator, Laundry in Building, Laund...   \n",
       "15     East 34th Street  [Doorman, Elevator, Fitness Center, Laundry in...   \n",
       "16     East 16th Street  [Doorman, Elevator, Loft, Dishwasher, Hardwood...   \n",
       "\n",
       "    latitude  listing_id  longitude                        manager_id  \\\n",
       "4    40.7108     7170325   -73.9539  a10db4590843d78c784171a107bdacb4   \n",
       "6    40.7513     7092344   -73.9722  955db33477af4f40004820b4aed804a0   \n",
       "9    40.7575     7158677   -73.9625  c8b10a317b766204f08e613cef4ce7a0   \n",
       "15   40.7439     7225292   -73.9743  2c3b41f588fbb5234d8a1e885a436cfa   \n",
       "16   40.7348     7226687   -73.9865  30a7951cfc21240e8c46b2d079d292e5   \n",
       "\n",
       "                                               photos  price_category  price  \\\n",
       "4   [https://photos.renthop.com/2/7170325_3bb5ac84...               6   2400   \n",
       "6   [https://photos.renthop.com/2/7092344_7663c19a...              22   3800   \n",
       "9   [https://photos.renthop.com/2/7158677_c897a134...              19   3495   \n",
       "15  [https://photos.renthop.com/2/7225292_901f1984...              10   2795   \n",
       "16  [https://photos.renthop.com/2/7226687_4ea22bf1...              43   7200   \n",
       "\n",
       "          street_address interest_level  \n",
       "4    145 Borinquen Place         medium  \n",
       "6          230 East 44th            low  \n",
       "9   405 East 56th Street         medium  \n",
       "15  340 East 34th Street            low  \n",
       "16  145 East 16th Street            low  "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#convert 'price' to categorical attributes\n",
    "indexes = list(df.index.values.tolist())\n",
    "price = list(df['price'][indexes])\n",
    "\n",
    "price_list = []\n",
    "for i in range(len(price)):\n",
    "    price_list.append(price[i])\n",
    "price_list.sort()\n",
    "price_list1 = list(dict.fromkeys(price_list))\n",
    "\n",
    "interval_num = 0\n",
    "if(len(price_list1) % CONST_INTERVAL ==0):\n",
    "    interval_num = len(price_list1) / CONST_INTERVAL\n",
    "else:\n",
    "    interval_num = math.floor(len(price_list1) / CONST_INTERVAL) + 1\n",
    "\n",
    "interval_num = int(interval_num)\n",
    "\n",
    "price_interval = [[0 for x in range(2)] for y in range(interval_num)] \n",
    "counter = 0\n",
    "label = 1\n",
    "price_interval[0][0] = price_list1[0]\n",
    "price_interval[0][1] = 0\n",
    "for i in range(len(price_list1)):\n",
    "    if(counter == CONST_INTERVAL):\n",
    "        counter = 0\n",
    "        price_interval[label][0] = price_list1[i]\n",
    "        price_interval[label][1] = label\n",
    "        label += 1\n",
    "    else:\n",
    "        counter += 1\n",
    "        \n",
    "#print(price_interval)\n",
    "price_category = []\n",
    "for i in range(len(price)):\n",
    "    for j in range(len(price_interval)):\n",
    "        if(price[i] < price_interval[j][0]):\n",
    "            price_category.append(price_interval[j][1]-1)\n",
    "            break\n",
    "        elif(j == (len(price_interval)-1)):\n",
    "            price_category.append(price_interval[j][1])\n",
    "            break\n",
    "    \n",
    "df.insert(12, \"price_category\", price_category, True) \n",
    "df.head(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>bathrooms</th>\n",
       "      <th>bedrooms</th>\n",
       "      <th>building_id</th>\n",
       "      <th>created</th>\n",
       "      <th>description</th>\n",
       "      <th>display_address</th>\n",
       "      <th>features</th>\n",
       "      <th>latitude</th>\n",
       "      <th>listing_id</th>\n",
       "      <th>longitude</th>\n",
       "      <th>manager_id</th>\n",
       "      <th>photos</th>\n",
       "      <th>price_category</th>\n",
       "      <th>price</th>\n",
       "      <th>street_address</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1</td>\n",
       "      <td>79780be1514f645d7e6be99a3de696c5</td>\n",
       "      <td>2016-06-11 05:29:41</td>\n",
       "      <td>Large with awesome terrace--accessible via bed...</td>\n",
       "      <td>Suffolk Street</td>\n",
       "      <td>[Elevator, Laundry in Building, Laundry in Uni...</td>\n",
       "      <td>40.7185</td>\n",
       "      <td>7142618</td>\n",
       "      <td>-73.9865</td>\n",
       "      <td>b1b1852c416d78d7765d746cb1b8921f</td>\n",
       "      <td>[https://photos.renthop.com/2/7142618_1c45a2c8...</td>\n",
       "      <td>16</td>\n",
       "      <td>2950</td>\n",
       "      <td>99 Suffolk Street</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.0</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>2016-06-24 06:36:34</td>\n",
       "      <td>Prime Soho - between Bleecker and Houston - Ne...</td>\n",
       "      <td>Thompson Street</td>\n",
       "      <td>[Pre-War, Dogs Allowed, Cats Allowed]</td>\n",
       "      <td>40.7278</td>\n",
       "      <td>7210040</td>\n",
       "      <td>-74.0000</td>\n",
       "      <td>d0b5648017832b2427eeb9956d966a14</td>\n",
       "      <td>[https://photos.renthop.com/2/7210040_d824cc71...</td>\n",
       "      <td>15</td>\n",
       "      <td>2850</td>\n",
       "      <td>176 Thompson Street</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2016-06-17 01:23:39</td>\n",
       "      <td>Spacious studio in Prime Location. Cleanbuildi...</td>\n",
       "      <td>Sullivan Street</td>\n",
       "      <td>[Pre-War, Dogs Allowed, Cats Allowed]</td>\n",
       "      <td>40.7260</td>\n",
       "      <td>7174566</td>\n",
       "      <td>-74.0026</td>\n",
       "      <td>e6472c7237327dd3903b3d6f6a94515a</td>\n",
       "      <td>[https://photos.renthop.com/2/7174566_ba3a35c5...</td>\n",
       "      <td>7</td>\n",
       "      <td>2295</td>\n",
       "      <td>115 Sullivan Street</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1.0</td>\n",
       "      <td>2</td>\n",
       "      <td>f9c826104b91d868e69bd25746448c0c</td>\n",
       "      <td>2016-06-21 05:06:02</td>\n",
       "      <td>For immediate access call Bryan.&lt;br /&gt;&lt;br /&gt;Bo...</td>\n",
       "      <td>Jones Street</td>\n",
       "      <td>[Hardwood Floors, Dogs Allowed, Cats Allowed]</td>\n",
       "      <td>40.7321</td>\n",
       "      <td>7191391</td>\n",
       "      <td>-74.0028</td>\n",
       "      <td>41735645e0f8f13993c42894023f8e58</td>\n",
       "      <td>[https://photos.renthop.com/2/7191391_8c2f2d49...</td>\n",
       "      <td>15</td>\n",
       "      <td>2900</td>\n",
       "      <td>23 Jones Street</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1</td>\n",
       "      <td>81062936e12ee5fa6cd2b965698e17d5</td>\n",
       "      <td>2016-06-16 07:24:27</td>\n",
       "      <td>Beautiful TRUE 1 bedroom in a luxury building ...</td>\n",
       "      <td>Exchange Place</td>\n",
       "      <td>[Roof Deck, Doorman, Elevator, Fitness Center,...</td>\n",
       "      <td>40.7054</td>\n",
       "      <td>7171695</td>\n",
       "      <td>-74.0095</td>\n",
       "      <td>a742cf7dd3b2627d83417bc3a1b3ec96</td>\n",
       "      <td>[https://photos.renthop.com/2/7171695_089ffee2...</td>\n",
       "      <td>21</td>\n",
       "      <td>3254</td>\n",
       "      <td>20 Exchange Place</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   bathrooms  bedrooms                       building_id              created  \\\n",
       "0        1.0         1  79780be1514f645d7e6be99a3de696c5  2016-06-11 05:29:41   \n",
       "1        1.0         2                                 0  2016-06-24 06:36:34   \n",
       "2        1.0         0                                 0  2016-06-17 01:23:39   \n",
       "3        1.0         2  f9c826104b91d868e69bd25746448c0c  2016-06-21 05:06:02   \n",
       "5        1.0         1  81062936e12ee5fa6cd2b965698e17d5  2016-06-16 07:24:27   \n",
       "\n",
       "                                         description  display_address  \\\n",
       "0  Large with awesome terrace--accessible via bed...   Suffolk Street   \n",
       "1  Prime Soho - between Bleecker and Houston - Ne...  Thompson Street   \n",
       "2  Spacious studio in Prime Location. Cleanbuildi...  Sullivan Street   \n",
       "3  For immediate access call Bryan.<br /><br />Bo...     Jones Street   \n",
       "5  Beautiful TRUE 1 bedroom in a luxury building ...   Exchange Place   \n",
       "\n",
       "                                            features  latitude  listing_id  \\\n",
       "0  [Elevator, Laundry in Building, Laundry in Uni...   40.7185     7142618   \n",
       "1              [Pre-War, Dogs Allowed, Cats Allowed]   40.7278     7210040   \n",
       "2              [Pre-War, Dogs Allowed, Cats Allowed]   40.7260     7174566   \n",
       "3      [Hardwood Floors, Dogs Allowed, Cats Allowed]   40.7321     7191391   \n",
       "5  [Roof Deck, Doorman, Elevator, Fitness Center,...   40.7054     7171695   \n",
       "\n",
       "   longitude                        manager_id  \\\n",
       "0   -73.9865  b1b1852c416d78d7765d746cb1b8921f   \n",
       "1   -74.0000  d0b5648017832b2427eeb9956d966a14   \n",
       "2   -74.0026  e6472c7237327dd3903b3d6f6a94515a   \n",
       "3   -74.0028  41735645e0f8f13993c42894023f8e58   \n",
       "5   -74.0095  a742cf7dd3b2627d83417bc3a1b3ec96   \n",
       "\n",
       "                                              photos  price_category  price  \\\n",
       "0  [https://photos.renthop.com/2/7142618_1c45a2c8...              16   2950   \n",
       "1  [https://photos.renthop.com/2/7210040_d824cc71...              15   2850   \n",
       "2  [https://photos.renthop.com/2/7174566_ba3a35c5...               7   2295   \n",
       "3  [https://photos.renthop.com/2/7191391_8c2f2d49...              15   2900   \n",
       "5  [https://photos.renthop.com/2/7171695_089ffee2...              21   3254   \n",
       "\n",
       "        street_address  \n",
       "0    99 Suffolk Street  \n",
       "1  176 Thompson Street  \n",
       "2  115 Sullivan Street  \n",
       "3      23 Jones Street  \n",
       "5    20 Exchange Place  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#convert 'price' to categorical attributes\n",
    "indexes_t = list(df2.index.values.tolist())\n",
    "price_t = list(df2['price'][indexes_t])\n",
    "\n",
    "price_list_t = []\n",
    "for i in range(len(price_t)):\n",
    "    price_list_t.append(price_t[i])\n",
    "price_list_t.sort()\n",
    "price_list1_t = list(dict.fromkeys(price_list_t))\n",
    "\n",
    "interval_num_t = 0\n",
    "if(len(price_list1_t) % CONST_INTERVAL ==0):\n",
    "    interval_num_t = len(price_list1_t) / CONST_INTERVAL\n",
    "else:\n",
    "    interval_num_t = math.floor(len(price_list1_t) / CONST_INTERVAL) + 1\n",
    "\n",
    "interval_num_t = int(interval_num_t)\n",
    "\n",
    "price_interval_t = [[0 for x in range(2)] for y in range(interval_num_t)] \n",
    "counter = 0\n",
    "label = 1\n",
    "price_interval_t[0][0] = price_list1_t[0]\n",
    "price_interval_t[0][1] = 0\n",
    "for i in range(len(price_list1_t)):\n",
    "    if(counter == CONST_INTERVAL):\n",
    "        counter = 0\n",
    "        price_interval_t[label][0] = price_list1_t[i]\n",
    "        price_interval_t[label][1] = label\n",
    "        label += 1\n",
    "    else:\n",
    "        counter += 1\n",
    "        \n",
    "#print(price_interval)\n",
    "price_category_t = []\n",
    "for i in range(len(price_t)):\n",
    "    for j in range(len(price_interval_t)):\n",
    "        if(price_t[i] < price_interval_t[j][0]):\n",
    "            price_category_t.append(price_interval_t[j][1]-1)\n",
    "            break\n",
    "        elif(j == (len(price_interval_t)-1)):\n",
    "            price_category_t.append(price_interval_t[j][1])\n",
    "            break\n",
    "    \n",
    "df2.insert(12, \"price_category\", price_category_t, True) \n",
    "df2.head(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>bathrooms</th>\n",
       "      <th>bedrooms</th>\n",
       "      <th>building_id</th>\n",
       "      <th>created</th>\n",
       "      <th>create time interval</th>\n",
       "      <th>description</th>\n",
       "      <th>display_address</th>\n",
       "      <th>features</th>\n",
       "      <th>latitude</th>\n",
       "      <th>listing_id</th>\n",
       "      <th>longitude</th>\n",
       "      <th>manager_id</th>\n",
       "      <th>photos</th>\n",
       "      <th>price_category</th>\n",
       "      <th>price</th>\n",
       "      <th>street_address</th>\n",
       "      <th>interest_level</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1</td>\n",
       "      <td>8579a0b0d54db803821a35a4a615e97a</td>\n",
       "      <td>2016-06-16 05:55:27</td>\n",
       "      <td>674</td>\n",
       "      <td>Spacious 1 Bedroom 1 Bathroom in Williamsburg!...</td>\n",
       "      <td>145 Borinquen Place</td>\n",
       "      <td>[Dining Room, Pre-War, Laundry in Building, Di...</td>\n",
       "      <td>40.7108</td>\n",
       "      <td>7170325</td>\n",
       "      <td>-73.9539</td>\n",
       "      <td>a10db4590843d78c784171a107bdacb4</td>\n",
       "      <td>[https://photos.renthop.com/2/7170325_3bb5ac84...</td>\n",
       "      <td>6</td>\n",
       "      <td>2400</td>\n",
       "      <td>145 Borinquen Place</td>\n",
       "      <td>medium</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>1.0</td>\n",
       "      <td>2</td>\n",
       "      <td>b8e75fc949a6cd8225b455648a951712</td>\n",
       "      <td>2016-06-01 05:44:33</td>\n",
       "      <td>528</td>\n",
       "      <td>BRAND NEW GUT RENOVATED TRUE 2 BEDROOMFind you...</td>\n",
       "      <td>East 44th</td>\n",
       "      <td>[Doorman, Elevator, Laundry in Building, Dishw...</td>\n",
       "      <td>40.7513</td>\n",
       "      <td>7092344</td>\n",
       "      <td>-73.9722</td>\n",
       "      <td>955db33477af4f40004820b4aed804a0</td>\n",
       "      <td>[https://photos.renthop.com/2/7092344_7663c19a...</td>\n",
       "      <td>22</td>\n",
       "      <td>3800</td>\n",
       "      <td>230 East 44th</td>\n",
       "      <td>low</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   bathrooms  bedrooms                       building_id              created  \\\n",
       "4        1.0         1  8579a0b0d54db803821a35a4a615e97a  2016-06-16 05:55:27   \n",
       "6        1.0         2  b8e75fc949a6cd8225b455648a951712  2016-06-01 05:44:33   \n",
       "\n",
       "   create time interval                                        description  \\\n",
       "4                   674  Spacious 1 Bedroom 1 Bathroom in Williamsburg!...   \n",
       "6                   528  BRAND NEW GUT RENOVATED TRUE 2 BEDROOMFind you...   \n",
       "\n",
       "       display_address                                           features  \\\n",
       "4  145 Borinquen Place  [Dining Room, Pre-War, Laundry in Building, Di...   \n",
       "6            East 44th  [Doorman, Elevator, Laundry in Building, Dishw...   \n",
       "\n",
       "   latitude  listing_id  longitude                        manager_id  \\\n",
       "4   40.7108     7170325   -73.9539  a10db4590843d78c784171a107bdacb4   \n",
       "6   40.7513     7092344   -73.9722  955db33477af4f40004820b4aed804a0   \n",
       "\n",
       "                                              photos  price_category  price  \\\n",
       "4  [https://photos.renthop.com/2/7170325_3bb5ac84...               6   2400   \n",
       "6  [https://photos.renthop.com/2/7092344_7663c19a...              22   3800   \n",
       "\n",
       "        street_address interest_level  \n",
       "4  145 Borinquen Place         medium  \n",
       "6        230 East 44th            low  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#convert 'created' to numerical attributes\n",
    "#this step will take about 1min to process\n",
    "indexes = list(df.index.values.tolist())\n",
    "time = list(df['created'][indexes])\n",
    "\n",
    "new_time = []\n",
    "for i in range(len(time)):\n",
    "    new_time.append(time[i])\n",
    "new_time.sort()\n",
    "new_time1 = list(dict.fromkeys(new_time))\n",
    "\n",
    "interval_leng = 0\n",
    "if(len(new_time1) % CONST_INTERVAL ==0):\n",
    "    interval_leng = len(new_time1) / CONST_INTERVAL\n",
    "else:\n",
    "    interval_leng = math.floor(len(new_time1) / CONST_INTERVAL) + 1\n",
    "    \n",
    "time_interval = [[0 for x in range(2)] for y in range(interval_leng)] \n",
    "counter = 0\n",
    "label = 1\n",
    "time_interval[0][0] = new_time1[0]\n",
    "time_interval[0][1] = 0\n",
    "for i in range(len(new_time1)):\n",
    "    if(counter == CONST_INTERVAL):\n",
    "        counter = 0\n",
    "        time_interval[label][0] = new_time1[i]\n",
    "        time_interval[label][1] = label\n",
    "        label += 1\n",
    "    else:\n",
    "        counter += 1\n",
    "        \n",
    "create_time = []\n",
    "for i in range(len(time)):\n",
    "    for j in range(len(time_interval)):\n",
    "        if(time[i] < str(time_interval[j][0])):\n",
    "            create_time.append(time_interval[j][1]-1)\n",
    "            break\n",
    "        elif(j == (len(time_interval)-1)):\n",
    "            create_time.append(time_interval[j][1])\n",
    "            break\n",
    "            \n",
    "df.insert(4, \"create time interval\", create_time, True) \n",
    "df.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "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>bathrooms</th>\n",
       "      <th>bedrooms</th>\n",
       "      <th>building_id</th>\n",
       "      <th>created</th>\n",
       "      <th>create time interval</th>\n",
       "      <th>description</th>\n",
       "      <th>display_address</th>\n",
       "      <th>features</th>\n",
       "      <th>latitude</th>\n",
       "      <th>listing_id</th>\n",
       "      <th>longitude</th>\n",
       "      <th>manager_id</th>\n",
       "      <th>photos</th>\n",
       "      <th>price_category</th>\n",
       "      <th>price</th>\n",
       "      <th>street_address</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1</td>\n",
       "      <td>79780be1514f645d7e6be99a3de696c5</td>\n",
       "      <td>2016-06-11 05:29:41</td>\n",
       "      <td>1130</td>\n",
       "      <td>Large with awesome terrace--accessible via bed...</td>\n",
       "      <td>Suffolk Street</td>\n",
       "      <td>[Elevator, Laundry in Building, Laundry in Uni...</td>\n",
       "      <td>40.7185</td>\n",
       "      <td>7142618</td>\n",
       "      <td>-73.9865</td>\n",
       "      <td>b1b1852c416d78d7765d746cb1b8921f</td>\n",
       "      <td>[https://photos.renthop.com/2/7142618_1c45a2c8...</td>\n",
       "      <td>16</td>\n",
       "      <td>2950</td>\n",
       "      <td>99 Suffolk Street</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.0</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>2016-06-24 06:36:34</td>\n",
       "      <td>1351</td>\n",
       "      <td>Prime Soho - between Bleecker and Houston - Ne...</td>\n",
       "      <td>Thompson Street</td>\n",
       "      <td>[Pre-War, Dogs Allowed, Cats Allowed]</td>\n",
       "      <td>40.7278</td>\n",
       "      <td>7210040</td>\n",
       "      <td>-74.0000</td>\n",
       "      <td>d0b5648017832b2427eeb9956d966a14</td>\n",
       "      <td>[https://photos.renthop.com/2/7210040_d824cc71...</td>\n",
       "      <td>15</td>\n",
       "      <td>2850</td>\n",
       "      <td>176 Thompson Street</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   bathrooms  bedrooms                       building_id              created  \\\n",
       "0        1.0         1  79780be1514f645d7e6be99a3de696c5  2016-06-11 05:29:41   \n",
       "1        1.0         2                                 0  2016-06-24 06:36:34   \n",
       "\n",
       "   create time interval                                        description  \\\n",
       "0                  1130  Large with awesome terrace--accessible via bed...   \n",
       "1                  1351  Prime Soho - between Bleecker and Houston - Ne...   \n",
       "\n",
       "   display_address                                           features  \\\n",
       "0   Suffolk Street  [Elevator, Laundry in Building, Laundry in Uni...   \n",
       "1  Thompson Street              [Pre-War, Dogs Allowed, Cats Allowed]   \n",
       "\n",
       "   latitude  listing_id  longitude                        manager_id  \\\n",
       "0   40.7185     7142618   -73.9865  b1b1852c416d78d7765d746cb1b8921f   \n",
       "1   40.7278     7210040   -74.0000  d0b5648017832b2427eeb9956d966a14   \n",
       "\n",
       "                                              photos  price_category  price  \\\n",
       "0  [https://photos.renthop.com/2/7142618_1c45a2c8...              16   2950   \n",
       "1  [https://photos.renthop.com/2/7210040_d824cc71...              15   2850   \n",
       "\n",
       "        street_address  \n",
       "0    99 Suffolk Street  \n",
       "1  176 Thompson Street  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#convert 'created' to categorical attributes(for test.json)\n",
    "#this step will take about 1min to process\n",
    "indexes_t = list(df2.index.values.tolist())\n",
    "time_t = list(df2['created'][indexes_t])\n",
    "\n",
    "new_time_t = []\n",
    "for i in range(len(time_t)):\n",
    "    new_time_t.append(time_t[i])\n",
    "new_time_t.sort()\n",
    "new_time_t1 = list(dict.fromkeys(new_time_t))\n",
    "\n",
    "interval_leng_t = 0\n",
    "if(len(new_time_t1) % CONST_INTERVAL ==0):\n",
    "    interval_leng_t = len(new_time_t1) / CONST_INTERVAL\n",
    "else:\n",
    "    interval_leng_t = math.floor(len(new_time_t1) / CONST_INTERVAL) + 1\n",
    "    \n",
    "time_interval_t = [[0 for x in range(2)] for y in range(interval_leng_t)] \n",
    "counter_t = 0\n",
    "label_t = 1\n",
    "time_interval_t[0][0] = new_time_t1[0]\n",
    "time_interval_t[0][1] = 0\n",
    "for i in range(len(new_time_t1)):\n",
    "    if(counter_t == CONST_INTERVAL):\n",
    "        counter_t = 0\n",
    "        time_interval_t[label_t][0] = new_time_t1[i]\n",
    "        time_interval_t[label_t][1] = label_t\n",
    "        label_t += 1\n",
    "    else:\n",
    "        counter_t += 1\n",
    "        \n",
    "create_time_t = []\n",
    "for i in range(len(time_t)):\n",
    "    for j in range(len(time_interval_t)):\n",
    "        if(time_t[i] < str(time_interval_t[j][0])):\n",
    "            create_time_t.append(time_interval_t[j][1]-1)\n",
    "            break\n",
    "        elif(j == (len(time_interval_t)-1)):\n",
    "            create_time_t.append(time_interval_t[j][1])\n",
    "            break\n",
    "            \n",
    "df2.insert(4, \"create time interval\", create_time_t, True) \n",
    "df2.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>bathrooms</th>\n",
       "      <th>bedrooms</th>\n",
       "      <th>building_id</th>\n",
       "      <th>created</th>\n",
       "      <th>create time interval</th>\n",
       "      <th>description</th>\n",
       "      <th>display_address</th>\n",
       "      <th>common word count</th>\n",
       "      <th>features</th>\n",
       "      <th>latitude</th>\n",
       "      <th>listing_id</th>\n",
       "      <th>longitude</th>\n",
       "      <th>manager_id</th>\n",
       "      <th>photos</th>\n",
       "      <th>price_category</th>\n",
       "      <th>price</th>\n",
       "      <th>street_address</th>\n",
       "      <th>interest_level</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1</td>\n",
       "      <td>8579a0b0d54db803821a35a4a615e97a</td>\n",
       "      <td>2016-06-16 05:55:27</td>\n",
       "      <td>674</td>\n",
       "      <td>Spacious 1 Bedroom 1 Bathroom in Williamsburg!...</td>\n",
       "      <td>145 Borinquen Place</td>\n",
       "      <td>4</td>\n",
       "      <td>[Dining Room, Pre-War, Laundry in Building, Di...</td>\n",
       "      <td>40.7108</td>\n",
       "      <td>7170325</td>\n",
       "      <td>-73.9539</td>\n",
       "      <td>a10db4590843d78c784171a107bdacb4</td>\n",
       "      <td>[https://photos.renthop.com/2/7170325_3bb5ac84...</td>\n",
       "      <td>6</td>\n",
       "      <td>2400</td>\n",
       "      <td>145 Borinquen Place</td>\n",
       "      <td>medium</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>1.0</td>\n",
       "      <td>2</td>\n",
       "      <td>b8e75fc949a6cd8225b455648a951712</td>\n",
       "      <td>2016-06-01 05:44:33</td>\n",
       "      <td>528</td>\n",
       "      <td>BRAND NEW GUT RENOVATED TRUE 2 BEDROOMFind you...</td>\n",
       "      <td>East 44th</td>\n",
       "      <td>3</td>\n",
       "      <td>[Doorman, Elevator, Laundry in Building, Dishw...</td>\n",
       "      <td>40.7513</td>\n",
       "      <td>7092344</td>\n",
       "      <td>-73.9722</td>\n",
       "      <td>955db33477af4f40004820b4aed804a0</td>\n",
       "      <td>[https://photos.renthop.com/2/7092344_7663c19a...</td>\n",
       "      <td>22</td>\n",
       "      <td>3800</td>\n",
       "      <td>230 East 44th</td>\n",
       "      <td>low</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   bathrooms  bedrooms                       building_id              created  \\\n",
       "4        1.0         1  8579a0b0d54db803821a35a4a615e97a  2016-06-16 05:55:27   \n",
       "6        1.0         2  b8e75fc949a6cd8225b455648a951712  2016-06-01 05:44:33   \n",
       "\n",
       "   create time interval                                        description  \\\n",
       "4                   674  Spacious 1 Bedroom 1 Bathroom in Williamsburg!...   \n",
       "6                   528  BRAND NEW GUT RENOVATED TRUE 2 BEDROOMFind you...   \n",
       "\n",
       "       display_address  common word count  \\\n",
       "4  145 Borinquen Place                  4   \n",
       "6            East 44th                  3   \n",
       "\n",
       "                                            features  latitude  listing_id  \\\n",
       "4  [Dining Room, Pre-War, Laundry in Building, Di...   40.7108     7170325   \n",
       "6  [Doorman, Elevator, Laundry in Building, Dishw...   40.7513     7092344   \n",
       "\n",
       "   longitude                        manager_id  \\\n",
       "4   -73.9539  a10db4590843d78c784171a107bdacb4   \n",
       "6   -73.9722  955db33477af4f40004820b4aed804a0   \n",
       "\n",
       "                                              photos  price_category  price  \\\n",
       "4  [https://photos.renthop.com/2/7170325_3bb5ac84...               6   2400   \n",
       "6  [https://photos.renthop.com/2/7092344_7663c19a...              22   3800   \n",
       "\n",
       "        street_address interest_level  \n",
       "4  145 Borinquen Place         medium  \n",
       "6        230 East 44th            low  "
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#tokenize 'features' then create vector\n",
    "#from milestone 1\n",
    "indexes = list(df.index.values.tolist())\n",
    "values = list(df['features'][indexes])\n",
    "text_list = []\n",
    "holder = 0\n",
    "for i in range(len(values)):\n",
    "    for ele in values[i]:\n",
    "        if(len(ele)>50):\n",
    "            temp = []\n",
    "            temp = re.split(r'[**]', ele)\n",
    "            for j in range(len(temp)):\n",
    "                text_list.append(temp[j])  \n",
    "        else:\n",
    "            text_list.append(ele)\n",
    "        holder += 1 \n",
    "        \n",
    "#milestone 2\n",
    "freq_word = Counter(text_list).most_common(CONST_FEATURE)\n",
    "freq_count = []\n",
    "\n",
    "for i in range(len(values)):\n",
    "    freq = 0\n",
    "    for ele in values[i]:\n",
    "        #print(ele)\n",
    "        for word in freq_word:\n",
    "            if(word[0] == ele):\n",
    "                #print('yes')\n",
    "                freq += 1\n",
    "    freq_count.append(freq)\n",
    "#print(freq_count)\n",
    "\n",
    "df.insert(7, \"common word count\", freq_count, True) \n",
    "df.head(2)"
   ]
  },
  {
   "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>bathrooms</th>\n",
       "      <th>bedrooms</th>\n",
       "      <th>building_id</th>\n",
       "      <th>created</th>\n",
       "      <th>create time interval</th>\n",
       "      <th>description</th>\n",
       "      <th>display_address</th>\n",
       "      <th>common word count</th>\n",
       "      <th>features</th>\n",
       "      <th>latitude</th>\n",
       "      <th>listing_id</th>\n",
       "      <th>longitude</th>\n",
       "      <th>manager_id</th>\n",
       "      <th>photos</th>\n",
       "      <th>price_category</th>\n",
       "      <th>price</th>\n",
       "      <th>street_address</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1</td>\n",
       "      <td>79780be1514f645d7e6be99a3de696c5</td>\n",
       "      <td>2016-06-11 05:29:41</td>\n",
       "      <td>1130</td>\n",
       "      <td>Large with awesome terrace--accessible via bed...</td>\n",
       "      <td>Suffolk Street</td>\n",
       "      <td>2</td>\n",
       "      <td>[Elevator, Laundry in Building, Laundry in Uni...</td>\n",
       "      <td>40.7185</td>\n",
       "      <td>7142618</td>\n",
       "      <td>-73.9865</td>\n",
       "      <td>b1b1852c416d78d7765d746cb1b8921f</td>\n",
       "      <td>[https://photos.renthop.com/2/7142618_1c45a2c8...</td>\n",
       "      <td>16</td>\n",
       "      <td>2950</td>\n",
       "      <td>99 Suffolk Street</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.0</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>2016-06-24 06:36:34</td>\n",
       "      <td>1351</td>\n",
       "      <td>Prime Soho - between Bleecker and Houston - Ne...</td>\n",
       "      <td>Thompson Street</td>\n",
       "      <td>2</td>\n",
       "      <td>[Pre-War, Dogs Allowed, Cats Allowed]</td>\n",
       "      <td>40.7278</td>\n",
       "      <td>7210040</td>\n",
       "      <td>-74.0000</td>\n",
       "      <td>d0b5648017832b2427eeb9956d966a14</td>\n",
       "      <td>[https://photos.renthop.com/2/7210040_d824cc71...</td>\n",
       "      <td>15</td>\n",
       "      <td>2850</td>\n",
       "      <td>176 Thompson Street</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   bathrooms  bedrooms                       building_id              created  \\\n",
       "0        1.0         1  79780be1514f645d7e6be99a3de696c5  2016-06-11 05:29:41   \n",
       "1        1.0         2                                 0  2016-06-24 06:36:34   \n",
       "\n",
       "   create time interval                                        description  \\\n",
       "0                  1130  Large with awesome terrace--accessible via bed...   \n",
       "1                  1351  Prime Soho - between Bleecker and Houston - Ne...   \n",
       "\n",
       "   display_address  common word count  \\\n",
       "0   Suffolk Street                  2   \n",
       "1  Thompson Street                  2   \n",
       "\n",
       "                                            features  latitude  listing_id  \\\n",
       "0  [Elevator, Laundry in Building, Laundry in Uni...   40.7185     7142618   \n",
       "1              [Pre-War, Dogs Allowed, Cats Allowed]   40.7278     7210040   \n",
       "\n",
       "   longitude                        manager_id  \\\n",
       "0   -73.9865  b1b1852c416d78d7765d746cb1b8921f   \n",
       "1   -74.0000  d0b5648017832b2427eeb9956d966a14   \n",
       "\n",
       "                                              photos  price_category  price  \\\n",
       "0  [https://photos.renthop.com/2/7142618_1c45a2c8...              16   2950   \n",
       "1  [https://photos.renthop.com/2/7210040_d824cc71...              15   2850   \n",
       "\n",
       "        street_address  \n",
       "0    99 Suffolk Street  \n",
       "1  176 Thompson Street  "
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#tokenize 'features' then create vector(for test.json)\n",
    "#from milestone 1\n",
    "indexes_t = list(df2.index.values.tolist())\n",
    "values_t = list(df2['features'][indexes_t])\n",
    "text_list_t = []\n",
    "holder_t = 0\n",
    "for i in range(len(values_t)):\n",
    "    for ele in values_t[i]:\n",
    "        if(len(ele)>50):\n",
    "            temp_t = []\n",
    "            temp_t = re.split(r'[**]', ele)\n",
    "            for j in range(len(temp_t)):\n",
    "                text_list_t.append(temp_t[j])  \n",
    "        else:\n",
    "            text_list_t.append(ele)\n",
    "        holder_t += 1 \n",
    "        \n",
    "#milestone 2\n",
    "freq_word_t = Counter(text_list_t).most_common(CONST_FEATURE)\n",
    "freq_count_t = []\n",
    "\n",
    "for i in range(len(values_t)):\n",
    "    freq_t = 0\n",
    "    for ele in values_t[i]:\n",
    "        #print(ele)\n",
    "        for word in freq_word_t:\n",
    "            if(word[0] == ele):\n",
    "                #print('yes')\n",
    "                freq_t += 1\n",
    "    freq_count_t.append(freq_t)\n",
    "#print(freq_count)\n",
    "\n",
    "df2.insert(7, \"common word count\", freq_count_t, True) \n",
    "df2.head(2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 1. First Classifier "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "use simple Cross-validation, and DecisionTreeClassifier without pruning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_x = df[['bathrooms',\n",
    "           'bedrooms',\n",
    "           'price_category',\n",
    "           'common word count',\n",
    "           'create time interval']]\n",
    "\n",
    "df_y = df[['interest_level']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "df_x_t = df2[['bathrooms',\n",
    "           'bedrooms',\n",
    "           'price_category',\n",
    "           'common word count',\n",
    "           'create time interval']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "x_train, x_test, y_train, y_test = train_test_split(\n",
    "    df_x, \n",
    "    df_y, \n",
    "    stratify = df_y, \n",
    "    test_size = 0.33, \n",
    "    random_state = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train_data_accuracy:0.6734095573145705\n",
      "Validation_data_accuracy:0.6736105944498177\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/zhilin/.local/lib/python3.6/site-packages/sklearn/utils/validation.py:760: 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",
      "  y = column_or_1d(y, warn=True)\n"
     ]
    }
   ],
   "source": [
    "clf = BernoulliNB()\n",
    "my_clf = clf.fit(x_train, y_train)\n",
    "\n",
    "print(f'Train_data_accuracy:{my_clf.score(x_train, y_train)}')\n",
    "print(f'Validation_data_accuracy:{my_clf.score(x_test, y_test)}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "proba_list = my_clf.predict_proba(df_x_t)\n",
    "\n",
    "high_p = []\n",
    "med_p = []\n",
    "low_p = []\n",
    "for i in range(len(proba_list)):\n",
    "    high_p = proba_list[i][0]\n",
    "    med_p = proba_list[i][1]\n",
    "    low_p = proba_list[i][2]\n",
    "    \n",
    "first_test_df = df2[['listing_id']]\n",
    "first_test_df.insert(1, 'high',high_p, True) \n",
    "first_test_df.insert(2, 'medium',med_p, True) \n",
    "first_test_df.insert(3, 'low',low_p, True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "compression_opts = dict(method='zip',\n",
    "                        archive_name='out1.csv')  \n",
    "first_test_df.to_csv('first_try.zip', index=False,\n",
    "          compression=compression_opts) "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 2. Improvment"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "new_df_x = df[['bathrooms', \n",
    "               'bedrooms', \n",
    "               'price_category', \n",
    "               'create time interval']]\n",
    "\n",
    "new_df_y = df[['interest_level']]\n",
    "\n",
    "new_df_x_t = df2[['bathrooms',\n",
    "                  'bedrooms',\n",
    "                  'price_category',\n",
    "                  'create time interval']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "#convert 'common word count' of train dataset \n",
    "#into several binary features\n",
    "\n",
    "word_matrix = [[0 for x in range(CONST_FEATURE)] for y in range(len(values))] \n",
    "\n",
    "for i in range(len(values)):\n",
    "    for j in range(CONST_FEATURE):\n",
    "        for ele in values[i]:\n",
    "            if(freq_word[j][0] == ele):\n",
    "                word_matrix[i][j] = 1\n",
    "\n",
    "for i in range(CONST_FEATURE):\n",
    "    word_list = []\n",
    "    for j in range(len(values)):  \n",
    "        word_list.append(word_matrix[j][i])\n",
    "    new_df_x.insert(i+4, i, word_list, True) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "#convert 'common word count' of test dataset \n",
    "#into several binary features\n",
    "\n",
    "word_matrix_t = [[0 for x in range(CONST_FEATURE)] for y in range(len(values_t))] \n",
    "\n",
    "for i in range(len(values_t)):\n",
    "    for j in range(CONST_FEATURE):\n",
    "        for ele in values_t[i]:\n",
    "            if(freq_word_t[j][0] == ele):\n",
    "                word_matrix_t[i][j] = 1\n",
    "\n",
    "for i in range(CONST_FEATURE):\n",
    "    word_list_t = []\n",
    "    for j in range(len(values_t)):  \n",
    "        word_list_t.append(word_matrix_t[j][i])\n",
    "    new_df_x_t.insert(i+4, i, word_list_t, True) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/zhilin/.local/lib/python3.6/site-packages/sklearn/utils/validation.py:760: 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",
      "  y = column_or_1d(y, warn=True)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "(40729, 5)"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#L1-based feature selection \n",
    "from sklearn.feature_selection import SelectFromModel\n",
    "clf2 = BernoulliNB().fit(new_df_x, new_df_y)\n",
    "model = SelectFromModel(clf2, prefit=True)\n",
    "final_x =  model.transform(new_df_x)\n",
    "final_x.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(32584, 1)"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#use Stratified k-fold for Cross-Validation\n",
    "skf = StratifiedKFold(n_splits = 5)\n",
    "\n",
    "for train_index, test_index in skf.split(final_x, new_df_y):\n",
    "    new_x_train, new_x_test =  new_df_x.iloc[train_index], new_df_x.iloc[test_index]\n",
    "    new_y_train, new_y_test = new_df_y.iloc[train_index], new_df_y.iloc[test_index]\n",
    "    \n",
    "new_y_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train_data_accuracy:0.6742266142892217\n",
      "Validation_data_accuracy:0.6747697974217312\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/zhilin/.local/lib/python3.6/site-packages/sklearn/utils/validation.py:760: 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",
      "  y = column_or_1d(y, warn=True)\n"
     ]
    }
   ],
   "source": [
    "#improved classifier\n",
    "clf = BernoulliNB(alpha=1.0,\n",
    "                  binarize = 0.0,\n",
    "                  class_prior=None, \n",
    "                  fit_prior=True)\n",
    "\n",
    "improved_clf = clf.fit(new_x_train, new_y_train)\n",
    "\n",
    "print(f'Train_data_accuracy:{improved_clf.score(new_x_train, new_y_train)}')\n",
    "print(f'Validation_data_accuracy:{improved_clf.score(new_x_test, new_y_test)}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion matrix with normalization\n",
      "[[0.02643857 0.97356143 0.        ]\n",
      " [0.00109389 0.99890611 0.        ]\n",
      " [0.00594943 0.99405057 0.        ]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#confusion matrix\n",
    "titles_options = [(\"Confusion matrix with normalization\", None)]\n",
    "for title, normalize in titles_options:\n",
    "    disp = plot_confusion_matrix(estimator = improved_clf, \n",
    "                                 X = new_x_test, \n",
    "                                 y_true = new_y_test,\n",
    "                                 display_labels=['high','medium','low'],\n",
    "                                 cmap=plt.cm.Blues,\n",
    "                                 normalize= 'true')\n",
    "    disp.ax_.set_title(title)\n",
    "\n",
    "    print(title)\n",
    "    print(disp.confusion_matrix)\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8078372061590404"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#logarithmic loss\n",
    "predict_list = improved_clf.predict_proba(new_x_test)\n",
    "log_loss(new_y_test, predict_list)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "proba_list = improved_clf.predict_proba(new_df_x_t)\n",
    "\n",
    "high_p = []\n",
    "med_p = []\n",
    "low_p = []\n",
    "for i in range(len(proba_list)):\n",
    "    high_p = proba_list[i][0]\n",
    "    med_p = proba_list[i][1]\n",
    "    low_p = proba_list[i][2]\n",
    "    \n",
    "final_test_df = df2[['listing_id']]\n",
    "final_test_df.insert(1, 'high',high_p, True) \n",
    "final_test_df.insert(2, 'medium',med_p, True) \n",
    "final_test_df.insert(3, 'low',low_p, True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "compression_opts = dict(method='zip',\n",
    "                        archive_name='out2.csv')  \n",
    "final_test_df.to_csv('prediction_improved_classifier.zip', index=False,\n",
    "          compression=compression_opts) "
   ]
  },
  {
   "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": "3.6.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}