GeneralML/Others/Retail Price Recommendation model with Gradient Boosting Tree.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Retail Price Recommendation model with Gradient Boosting Tree"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Background\n",
    "Mercari (Japan’s biggest shopping app) would like to offer pricing suggestions to sellers, but this is not easy because their sellers are enabled to put just about anything, or any bundle of things, on Mercari’s marketplace.\n",
    "In this machine learning project, we will build a model that automatically suggests the right product prices. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1. Load Data \n",
    "The dataset is from Kaggle [Link](https://www.kaggle.com/saitosean/mercari?select=train.tsv)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import gc\n",
    "import os\n",
    "import time\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "\n",
    "import lightgbm as lgb\n",
    "\n",
    "from sklearn.metrics import mean_squared_error\n",
    "from scipy.sparse import csr_matrix, hstack\n",
    "from sklearn.preprocessing import LabelBinarizer\n",
    "from sklearn.feature_extraction.text import CountVectorizer, TfidfVectorizer\n",
    "from sklearn.model_selection import train_test_split, cross_val_score\n",
    "\n",
    "import sweetviz as sv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "path = 'C:\\\\Users\\\\bokhy\\\\Desktop\\\\Python\\\\github\\\\Python-Projects'  \n",
    "\n",
    "df = pd.read_csv(os.path.join(path, 'train.tsv'), sep = '\\t')\n",
    "# test = pd.read_csv(os.path.join(path, 'test.tsv'), sep = '\\t')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {},
   "outputs": [],
   "source": [
    "#train, test = train_test_split(train, test_size=0.2) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "msk = np.random.rand(len(df)) < 0.8\n",
    "train = df[msk]\n",
    "test = df[~msk]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((1185737, 8), (296798, 8))"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.shape, test.shape"
   ]
  },
  {
   "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>train_id</th>\n",
       "      <th>name</th>\n",
       "      <th>item_condition_id</th>\n",
       "      <th>category_name</th>\n",
       "      <th>brand_name</th>\n",
       "      <th>price</th>\n",
       "      <th>shipping</th>\n",
       "      <th>item_description</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>Razer BlackWidow Chroma Keyboard</td>\n",
       "      <td>3</td>\n",
       "      <td>Electronics/Computers &amp; Tablets/Components &amp; P...</td>\n",
       "      <td>Razer</td>\n",
       "      <td>52.0</td>\n",
       "      <td>0</td>\n",
       "      <td>This keyboard is in great condition and works ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>AVA-VIV Blouse</td>\n",
       "      <td>1</td>\n",
       "      <td>Women/Tops &amp; Blouses/Blouse</td>\n",
       "      <td>Target</td>\n",
       "      <td>10.0</td>\n",
       "      <td>1</td>\n",
       "      <td>Adorable top with a hint of lace and a key hol...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>Leather Horse Statues</td>\n",
       "      <td>1</td>\n",
       "      <td>Home/Home Décor/Home Décor Accents</td>\n",
       "      <td>NaN</td>\n",
       "      <td>35.0</td>\n",
       "      <td>1</td>\n",
       "      <td>New with tags. Leather horses. Retail for [rm]...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>24K GOLD plated rose</td>\n",
       "      <td>1</td>\n",
       "      <td>Women/Jewelry/Necklaces</td>\n",
       "      <td>NaN</td>\n",
       "      <td>44.0</td>\n",
       "      <td>0</td>\n",
       "      <td>Complete with certificate of authenticity</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>5</td>\n",
       "      <td>Bundled items requested for Ruie</td>\n",
       "      <td>3</td>\n",
       "      <td>Women/Other/Other</td>\n",
       "      <td>NaN</td>\n",
       "      <td>59.0</td>\n",
       "      <td>0</td>\n",
       "      <td>Banana republic bottoms, Candies skirt with ma...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   train_id                              name  item_condition_id  \\\n",
       "1         1  Razer BlackWidow Chroma Keyboard                  3   \n",
       "2         2                    AVA-VIV Blouse                  1   \n",
       "3         3             Leather Horse Statues                  1   \n",
       "4         4              24K GOLD plated rose                  1   \n",
       "5         5  Bundled items requested for Ruie                  3   \n",
       "\n",
       "                                       category_name brand_name  price  \\\n",
       "1  Electronics/Computers & Tablets/Components & P...      Razer   52.0   \n",
       "2                        Women/Tops & Blouses/Blouse     Target   10.0   \n",
       "3                 Home/Home Décor/Home Décor Accents        NaN   35.0   \n",
       "4                            Women/Jewelry/Necklaces        NaN   44.0   \n",
       "5                                  Women/Other/Other        NaN   59.0   \n",
       "\n",
       "   shipping                                   item_description  \n",
       "1         0  This keyboard is in great condition and works ...  \n",
       "2         1  Adorable top with a hint of lace and a key hol...  \n",
       "3         1  New with tags. Leather horses. Retail for [rm]...  \n",
       "4         0          Complete with certificate of authenticity  \n",
       "5         0  Banana republic bottoms, Candies skirt with ma...  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "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>train_id</th>\n",
       "      <th>name</th>\n",
       "      <th>item_condition_id</th>\n",
       "      <th>category_name</th>\n",
       "      <th>brand_name</th>\n",
       "      <th>price</th>\n",
       "      <th>shipping</th>\n",
       "      <th>item_description</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>MLB Cincinnati Reds T Shirt Size XL</td>\n",
       "      <td>3</td>\n",
       "      <td>Men/Tops/T-shirts</td>\n",
       "      <td>NaN</td>\n",
       "      <td>10.0</td>\n",
       "      <td>1</td>\n",
       "      <td>No description yet</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>6</td>\n",
       "      <td>Acacia pacific tides santorini top</td>\n",
       "      <td>3</td>\n",
       "      <td>Women/Swimwear/Two-Piece</td>\n",
       "      <td>Acacia Swimwear</td>\n",
       "      <td>64.0</td>\n",
       "      <td>0</td>\n",
       "      <td>Size small but straps slightly shortened to fi...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>12</td>\n",
       "      <td>Black Skater dress</td>\n",
       "      <td>2</td>\n",
       "      <td>Women/Dresses/Above Knee, Mini</td>\n",
       "      <td>rue</td>\n",
       "      <td>16.0</td>\n",
       "      <td>0</td>\n",
       "      <td>Xl, great condition</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>23</td>\n",
       "      <td>Triple car charger</td>\n",
       "      <td>1</td>\n",
       "      <td>Electronics/Cell Phones &amp; Accessories/Chargers...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>8.0</td>\n",
       "      <td>1</td>\n",
       "      <td>Brand new never used All colors are available ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>41</th>\n",
       "      <td>41</td>\n",
       "      <td>tropical flower print bra from pink</td>\n",
       "      <td>3</td>\n",
       "      <td>Women/Underwear/Bras</td>\n",
       "      <td>PINK</td>\n",
       "      <td>13.0</td>\n",
       "      <td>1</td>\n",
       "      <td>Used buh looks brand new and is washed and rea...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    train_id                                 name  item_condition_id  \\\n",
       "0          0  MLB Cincinnati Reds T Shirt Size XL                  3   \n",
       "6          6   Acacia pacific tides santorini top                  3   \n",
       "12        12                   Black Skater dress                  2   \n",
       "23        23                   Triple car charger                  1   \n",
       "41        41  tropical flower print bra from pink                  3   \n",
       "\n",
       "                                        category_name       brand_name  price  \\\n",
       "0                                   Men/Tops/T-shirts              NaN   10.0   \n",
       "6                            Women/Swimwear/Two-Piece  Acacia Swimwear   64.0   \n",
       "12                     Women/Dresses/Above Knee, Mini              rue   16.0   \n",
       "23  Electronics/Cell Phones & Accessories/Chargers...              NaN    8.0   \n",
       "41                               Women/Underwear/Bras             PINK   13.0   \n",
       "\n",
       "    shipping                                   item_description  \n",
       "0          1                                 No description yet  \n",
       "6          0  Size small but straps slightly shortened to fi...  \n",
       "12         0                                Xl, great condition  \n",
       "23         1  Brand new never used All colors are available ...  \n",
       "41         1  Used buh looks brand new and is washed and rea...  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "1. train_id — the id of the listing\n",
    "2. name — the title of the listing\n",
    "3. item_condition_id — the condition of the items provided by the sellers\n",
    "4. category_name — category of the listing \n",
    "5. brand_name — the name of the brand\n",
    "6. price — the price that the item was sold for. **Target Variable**\n",
    "7. shipping — 1 if shipping fee is paid by seller and 0 if it's paid by buyer\n",
    "8. item_description — the full description of the item"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      ":FEATURES DONE:                    |█████████████████████| [100%]   00:23  -> (00:00 left)\n",
      ":PAIRWISE DONE:                    |█████████████████████| [100%]   00:05  -> (00:00 left)\n",
      "c:\\users\\bokhy\\appdata\\local\\programs\\python\\python37\\lib\\site-packages\\matplotlib\\cbook\\__init__.py:424: MatplotlibDeprecationWarning: \n",
      "Passing one of 'on', 'true', 'off', 'false' as a boolean is deprecated; use an actual boolean (True/False) instead.\n",
      "  \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Creating Associations graph... DONE!\n"
     ]
    }
   ],
   "source": [
    "my_report = sv.analyze(train,\n",
    "                      target_feat = 'price',\n",
    "                      pairwise_analysis= \"on\")\n",
    "my_report.show_html()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      ":FEATURES DONE:                    |█████████████████████| [100%]   00:34  -> (00:00 left)\n",
      ":PAIRWISE DONE:                    |█████████████████████| [100%]   00:06  -> (00:00 left)\n",
      "c:\\users\\bokhy\\appdata\\local\\programs\\python\\python37\\lib\\site-packages\\matplotlib\\cbook\\__init__.py:424: MatplotlibDeprecationWarning: \n",
      "Passing one of 'on', 'true', 'off', 'false' as a boolean is deprecated; use an actual boolean (True/False) instead.\n",
      "  \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Creating Associations graph... "
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\users\\bokhy\\appdata\\local\\programs\\python\\python37\\lib\\site-packages\\matplotlib\\cbook\\__init__.py:424: MatplotlibDeprecationWarning: \n",
      "Passing one of 'on', 'true', 'off', 'false' as a boolean is deprecated; use an actual boolean (True/False) instead.\n",
      "  \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DONE!\n"
     ]
    }
   ],
   "source": [
    "# Compare Train and Test set\n",
    "my_report = sv.compare([train, \"Train Data\"], [test, \"Test Data\"], \n",
    "                       target_feat = \"price\",\n",
    "                       pairwise_analysis= \"on\")\n",
    "my_report.show_html()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2.EDA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Examine Target Variable"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'Price Distribution')"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1200x600 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.subplot(1, 2, 1)\n",
    "(train['price']).plot.hist(bins=50, figsize=(12, 6), edgecolor = 'white', range = [0, 250])\n",
    "plt.xlabel('price', fontsize=12)\n",
    "plt.title('Price Distribution', fontsize=12)\n",
    "\n",
    "plt.subplot(1, 2, 2)\n",
    "np.log(train['price']+1).plot.hist(bins=50, figsize=(12,6), edgecolor='white')\n",
    "plt.xlabel('log(price+1)', fontsize=12)\n",
    "plt.title('Price Distribution', fontsize=12)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 'shipping variable'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    0.552642\n",
       "1    0.447358\n",
       "Name: shipping, dtype: float64"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train['shipping'].value_counts() / len(train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    0.55306\n",
       "1    0.44694\n",
       "Name: shipping, dtype: float64"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test['shipping'].value_counts() / len(test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In both train and test set, 55% of items has a value of '0', which means that shipping fee were paid by the buyers. \\\n",
    "How shipping related to price?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1800x800 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "shipping_fee_by_buyer = train.loc[train['shipping'] == 0, 'price']\n",
    "shipping_fee_by_seller = train.loc[train['shipping'] == 1, 'price']\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(18,8))\n",
    "ax.hist(shipping_fee_by_seller, color='#10B9C5', alpha=1.0, bins=50, range = [0, 100],\n",
    "       label='Price when Seller pays Shipping')\n",
    "ax.hist(shipping_fee_by_buyer, color='#C52F10', alpha=0.7, bins=50, range = [0, 100],\n",
    "       label='Price when Buyer pays Shipping')\n",
    "\n",
    "plt.xlabel('price', fontsize=12)\n",
    "plt.ylabel('frequency', fontsize=12)\n",
    "plt.title('Price Distribution by Shipping Type', fontsize=15)\n",
    "plt.tick_params(labelsize=12)\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The average price is 22.6 if seller pays shipping\n",
      "The average price is 30.13 if buyer pays shipping\n"
     ]
    }
   ],
   "source": [
    "print('The average price is {}'.format(round(shipping_fee_by_seller.mean(), 2)), 'if seller pays shipping');\n",
    "print('The average price is {}'.format(round(shipping_fee_by_buyer.mean(), 2)), 'if buyer pays shipping')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1800x800 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Log transformed version\n",
    "fig, ax = plt.subplots(figsize=(18,8))\n",
    "ax.hist(np.log(shipping_fee_by_seller+1), color='#8CB4E1', alpha=1.0, bins=50,\n",
    "       label='Price when Seller pays Shipping')\n",
    "ax.hist(np.log(shipping_fee_by_buyer+1), color='#007D00', alpha=0.7, bins=50,\n",
    "       label='Price when Buyer pays Shipping')\n",
    "plt.xlabel('log(price+1)', fontsize=12)\n",
    "plt.ylabel('frequency', fontsize=12)\n",
    "plt.title('Price Distribution by Shipping Type', fontsize=15)\n",
    "plt.tick_params(labelsize=12)\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can see that the average price is **higher** when buyer pays shipping."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Category Names"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "There are 1268 unique values in category name column\n"
     ]
    }
   ],
   "source": [
    "print('There are', train['category_name'].nunique(), 'unique values in category name column')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Women/Athletic Apparel/Pants, Tights, Leggings                 48093\n",
       "Women/Tops & Blouses/T-Shirts                                  37133\n",
       "Beauty/Makeup/Face                                             27548\n",
       "Beauty/Makeup/Lips                                             23908\n",
       "Electronics/Video Games & Consoles/Games                       21249\n",
       "Beauty/Makeup/Eyes                                             20241\n",
       "Electronics/Cell Phones & Accessories/Cases, Covers & Skins    19709\n",
       "Women/Underwear/Bras                                           16945\n",
       "Women/Tops & Blouses/Blouse                                    16288\n",
       "Women/Tops & Blouses/Tank, Cami                                16188\n",
       "Name: category_name, dtype: int64"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train['category_name'].value_counts()[:10]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Item condition vs. Price"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x2bb2fa21508>"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 600x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.boxplot(x = 'item_condition_id', y = np.log(train['price']+1), data = train, palette = sns.color_palette('RdBu',5))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Therefore, we are going to use all the features to build our model."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3. Create a Model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### (1) Hyperparameter setting"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "NUM_BRANDS = 4000\n",
    "NUM_CATEGORIES = 1000\n",
    "NAME_MIN_DF = 10\n",
    "MAX_FEATURES_ITEM_DESCRIPTION = 50000"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "There are 5069 items where category name is missing .\n"
     ]
    }
   ],
   "source": [
    "print('There are %d items where category name is missing .' %train['category_name'].isnull().sum())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "There are 506181 items where brand name is missing.\n"
     ]
    }
   ],
   "source": [
    "print('There are %d items where brand name is missing.' %train['brand_name'].isnull().sum())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "There are 4 items where description is missing.\n"
     ]
    }
   ],
   "source": [
    "print('There are %d items where description is missing.' %train['item_description'].isnull().sum())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### (2) Data Pre-processing\n",
    "###### Helper functions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "def handle_missing_inplace(dataset): \n",
    "    dataset['category_name'].fillna(value='missing', inplace=True) \n",
    "    dataset['brand_name'].fillna(value='missing', inplace=True) \n",
    "    dataset['item_description'].replace('No description yet,''missing', inplace=True) \n",
    "    dataset['item_description'].fillna(value='missing', inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "def cutting(dataset):\n",
    "    pop_brand = dataset['brand_name'].value_counts().loc[lambda x: x.index != 'missing'].index[:NUM_BRANDS]\n",
    "    dataset.loc[~dataset['brand_name'].isin(pop_brand), 'brand_name'] = 'missing'\n",
    "    pop_category = dataset['category_name'].value_counts().loc[lambda x: x.index != 'missing'].index[:NUM_CATEGORIES]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "def to_categorical(dataset):\n",
    "    dataset['category_name'] = dataset['category_name'].astype('category')\n",
    "    dataset['brand_name'] = dataset['brand_name'].astype('category')\n",
    "    dataset['item_condition_id'] = dataset['item_condition_id'].astype('category')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "# drop target column\n",
    "test_new = test.drop('price', axis=1)\n",
    "y_test = np.log1p(test[\"price\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Delete rows where price = 0\n",
    "train = train[train.price != 0].reset_index(drop=True)\n",
    "\n",
    "nrow_train = train.shape[0]\n",
    "y = np.log1p(train[\"price\"])\n",
    "merge: pd.DataFrame = pd.concat([train, test_new])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "handle_missing_inplace(merge)\n",
    "cutting(merge)\n",
    "to_categorical(merge)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "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>train_id</th>\n",
       "      <th>name</th>\n",
       "      <th>item_condition_id</th>\n",
       "      <th>category_name</th>\n",
       "      <th>brand_name</th>\n",
       "      <th>price</th>\n",
       "      <th>shipping</th>\n",
       "      <th>item_description</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>Razer BlackWidow Chroma Keyboard</td>\n",
       "      <td>3</td>\n",
       "      <td>Electronics/Computers &amp; Tablets/Components &amp; P...</td>\n",
       "      <td>Razer</td>\n",
       "      <td>52.0</td>\n",
       "      <td>0</td>\n",
       "      <td>This keyboard is in great condition and works ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>AVA-VIV Blouse</td>\n",
       "      <td>1</td>\n",
       "      <td>Women/Tops &amp; Blouses/Blouse</td>\n",
       "      <td>Target</td>\n",
       "      <td>10.0</td>\n",
       "      <td>1</td>\n",
       "      <td>Adorable top with a hint of lace and a key hol...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>Leather Horse Statues</td>\n",
       "      <td>1</td>\n",
       "      <td>Home/Home Décor/Home Décor Accents</td>\n",
       "      <td>missing</td>\n",
       "      <td>35.0</td>\n",
       "      <td>1</td>\n",
       "      <td>New with tags. Leather horses. Retail for [rm]...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>24K GOLD plated rose</td>\n",
       "      <td>1</td>\n",
       "      <td>Women/Jewelry/Necklaces</td>\n",
       "      <td>missing</td>\n",
       "      <td>44.0</td>\n",
       "      <td>0</td>\n",
       "      <td>Complete with certificate of authenticity</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>Bundled items requested for Ruie</td>\n",
       "      <td>3</td>\n",
       "      <td>Women/Other/Other</td>\n",
       "      <td>missing</td>\n",
       "      <td>59.0</td>\n",
       "      <td>0</td>\n",
       "      <td>Banana republic bottoms, Candies skirt with ma...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   train_id                              name item_condition_id  \\\n",
       "0         1  Razer BlackWidow Chroma Keyboard                 3   \n",
       "1         2                    AVA-VIV Blouse                 1   \n",
       "2         3             Leather Horse Statues                 1   \n",
       "3         4              24K GOLD plated rose                 1   \n",
       "4         5  Bundled items requested for Ruie                 3   \n",
       "\n",
       "                                       category_name brand_name  price  \\\n",
       "0  Electronics/Computers & Tablets/Components & P...      Razer   52.0   \n",
       "1                        Women/Tops & Blouses/Blouse     Target   10.0   \n",
       "2                 Home/Home Décor/Home Décor Accents    missing   35.0   \n",
       "3                            Women/Jewelry/Necklaces    missing   44.0   \n",
       "4                                  Women/Other/Other    missing   59.0   \n",
       "\n",
       "   shipping                                   item_description  \n",
       "0         0  This keyboard is in great condition and works ...  \n",
       "1         1  Adorable top with a hint of lace and a key hol...  \n",
       "2         1  New with tags. Leather horses. Retail for [rm]...  \n",
       "3         0          Complete with certificate of authenticity  \n",
       "4         0  Banana republic bottoms, Candies skirt with ma...  "
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# merge = merge.drop('test_id', axis=1)\n",
    "merge.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1481835, 8)"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "merge.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Count vectorize name and category name columns."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "cv = CountVectorizer(min_df=NAME_MIN_DF)\n",
    "X_name = cv.fit_transform(merge['name'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<1481835x17732 sparse matrix of type '<class 'numpy.int64'>'\n",
       "\twith 6066993 stored elements in Compressed Sparse Row format>"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_name"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "cv = CountVectorizer()\n",
    "X_category = cv.fit_transform(merge['category_name'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<1481835x1012 sparse matrix of type '<class 'numpy.int64'>'\n",
       "\twith 5933344 stored elements in Compressed Sparse Row format>"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_category"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### TFIDF Vectorize item_description column."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "tv = TfidfVectorizer(max_features=MAX_FEATURES_ITEM_DESCRIPTION, ngram_range=(1, 3), stop_words='english')\n",
    "X_description = tv.fit_transform(merge['item_description'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<1481835x50000 sparse matrix of type '<class 'numpy.float64'>'\n",
       "\twith 33047118 stored elements in Compressed Sparse Row format>"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_description"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Label binarize brand_name column."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "lb = LabelBinarizer(sparse_output=True)\n",
    "X_brand = lb.fit_transform(merge['brand_name'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<1481835x4001 sparse matrix of type '<class 'numpy.int32'>'\n",
       "\twith 1481835 stored elements in Compressed Sparse Row format>"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_brand"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Create dummy variables for item_condition_id and shipping columns."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "aa = pd.get_dummies(merge[['item_condition_id', 'shipping']], sparse=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "aa = np.array(aa, dtype=float)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_dummies = csr_matrix(aa)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<1481835x6 sparse matrix of type '<class 'numpy.float64'>'\n",
       "\twith 2144677 stored elements in Compressed Sparse Row format>"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_dummies"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Create sparse merge"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "sparse_merge = hstack((X_dummies, X_description, X_brand, X_category, X_name)).tocsr()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1481835, 72751)"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sparse_merge.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Remove features with document frequency <=1."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "mask = np.array(np.clip(sparse_merge.getnnz(axis=0) - 1, 0, 1), dtype=bool)\n",
    "sparse_merge = sparse_merge[:, mask]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1481835, 72285)"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sparse_merge.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Separate train and test data from sparse merge."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = sparse_merge[:nrow_train]\n",
    "X_test = sparse_merge[nrow_train:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_X = lgb.Dataset(X, label=y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [],
   "source": [
    "params = {\n",
    "        'learning_rate': 0.02,\n",
    "        'application': 'regression',\n",
    "        'max_depth': 9,\n",
    "        'num_leaves': 30,\n",
    "        'verbosity': -1,\n",
    "        'metric': 'RMSE',\n",
    "    }"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Start Training\n",
    "gbm = lgb.train(params, train_X, num_boost_round=1000, verbose_eval=100)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 4. Inference using Test set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred = gbm.predict(X_test, num_iteration=gbm.best_iteration)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([2.39003494, 3.75873855, 2.85574048, ..., 2.63401498, 2.7621529 ,\n",
       "       2.49262279])"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pred"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The rmse of prediction is: 0.5309652205674069\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import mean_squared_error\n",
    "print('The rmse of prediction is:', mean_squared_error(y_test, y_pred) ** 0.5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 5. Improvements"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "1. Fine-tuning hyperparameters\n",
    "2. Pre-processing"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Reference\n",
    ">https://www.kaggle.com/tunguz/more-effective-ridge-lgbm-script-lb-0-44823\n",
    ">https://towardsdatascience.com/machine-learning-for-retail-price-suggestion-with-python-64531e64186d"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}