MPS_Entity_Embedding_Data_Modeling.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "MPS: Entity Embedding Data Modeling.ipynb",
      "version": "0.3.2",
      "provenance": [],
      "collapsed_sections": [
        "aXZ-drvX4QrV"
      ],
      "machine_shape": "hm",
      "include_colab_link": true
    },
    "language_info": {
      "name": "python",
      "version": "3.6.4",
      "mimetype": "text/x-python",
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "pygments_lexer": "ipython3",
      "nbconvert_exporter": "python",
      "file_extension": ".py"
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/jasoriya/Mercari-Price-Suggestion-Challenge/blob/master/MPS_Entity_Embedding_Data_Modeling.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9YcLNOTSq2C5",
        "colab_type": "code",
        "outputId": "1e2883b3-c1f2-47df-feef-186f3a4e097a",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 122
        }
      },
      "source": [
        "# Don't run if not running in Google Colab\n",
        "from google.colab import drive\n",
        "drive.mount('/content/gdrive')"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Go to this URL in a browser: https://accounts.google.com/o/oauth2/auth?client_id=947318989803-6bn6qk8qdgf4n4g3pfee6491hc0brc4i.apps.googleusercontent.com&redirect_uri=urn%3Aietf%3Awg%3Aoauth%3A2.0%3Aoob&scope=email%20https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fdocs.test%20https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fdrive%20https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fdrive.photos.readonly%20https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fpeopleapi.readonly&response_type=code\n",
            "\n",
            "Enter your authorization code:\n",
            "··········\n",
            "Mounted at /content/gdrive\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "nZMPp5QmCPGS",
        "colab_type": "text"
      },
      "source": [
        "Input data files are available in the \"../mercari-price-suggestion-challenge/\" directory.\n",
        "**NOTE**: Please modify the absolute path before /mercari-price-suggestion-challenge accordingly\n",
        "\n",
        "For example, running this (by clicking run or pressing Shift+Enter) will list the files in the mercari-price-suggestion-challenge directory"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "_uuid": "8f2839f25d086af736a60e9eeb907d3b93b6e0e5",
        "_cell_guid": "b1076dfc-b9ad-4769-8c92-a6c4dae69d19",
        "trusted": true,
        "id": "bEC0wv0PqqK2",
        "colab_type": "code",
        "outputId": "52f27b79-4515-43cd-c634-cea3ca9cf09b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 71
        }
      },
      "source": [
        "import math\n",
        "import time\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n",
        "import dask.dataframe as dd\n",
        "import pydot_ng as pydot\n",
        "\n",
        "import gensim\n",
        "import string\n",
        "import re\n",
        "from gensim.models import KeyedVectors\n",
        "\n",
        "from sklearn import metrics, linear_model\n",
        "from sklearn.model_selection import train_test_split, cross_val_score, cross_val_predict\n",
        "from sklearn.preprocessing import StandardScaler\n",
        "\n",
        "from keras.models import Sequential, Model, load_model\n",
        "from keras.models import Model as KerasModel\n",
        "from keras.layers import Input, Dense, Activation, Reshape, Conv1D, Flatten, MaxPooling1D\n",
        "from keras.layers import Concatenate, Dropout\n",
        "from keras.layers.embeddings import Embedding\n",
        "from keras.callbacks import ModelCheckpoint, EarlyStopping, ReduceLROnPlateau, History\n",
        "from keras.utils import plot_model\n",
        "import keras\n",
        "import keras.backend as K\n",
        "\n",
        "# Input data files are available in the \"../mercari-price-suggestion-challenge/\" directory.\n",
        "# NOTE: Please modify the absolute path before /mercari-price-suggestion-challenge accordingly\n",
        "\n",
        "# For example, running this (by clicking run or pressing Shift+Enter) will list the files in the mercari-price-suggestion-challenge directory\n",
        "\n",
        "import os\n",
        "print(os.listdir(\"/content/gdrive/My Drive/DMT/mercari-price-suggestion-challenge\"))\n",
        "input_folder = '/content/gdrive/My Drive/DMT/mercari-price-suggestion-challenge/'\n",
        "\n",
        "# Any results you write to the current directory are saved as output."
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Using TensorFlow backend.\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "['sample_submission.csv.7z', 'sample_submission_stg2.csv', 'train.tsv', 'test.tsv', 'train.pkl', 'model.best.weights.hdf5', 'train_word2vec.pkl', 'model.png', 'train_xgboost.pkl', 'dumps', 'model_word2vec.best.weights.hdf5']\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "aXZ-drvX4QrV",
        "colab_type": "text"
      },
      "source": [
        "# 1. Load saved pickle and [skip till this cell](https://colab.research.google.com/drive/15-uCrPDef7kByQndWkC7JMH3-g3xOU_f?authuser=1#scrollTo=266Lf1n_6YMs)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BIM-xJgb4OcF",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "data = pd.read_pickle(input_folder + 'dumps/whole_data_word2vec.pkl')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "aEkiuDyH7uv3",
        "colab_type": "text"
      },
      "source": [
        "[Click here](https://colab.research.google.com/drive/15-uCrPDef7kByQndWkC7JMH3-g3xOU_f?authuser=1#scrollTo=266Lf1n_6YMs) to skip if you ran the above cell. **If not**, proceed to run normally and the program will compute the *whole_data_word2vec* pickle file"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "HPoAk9kPqqK6",
        "colab_type": "text"
      },
      "source": [
        "# 2. Data Preparation"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0YSQiGz6UkhR",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Load saved preprocessed train data and skip Data Preprocessing and Data Wrangling\n",
        "train_ids = pd.read_pickle(input_folder + 'dumps/train_data.pkl')['train_id'].tolist()\n",
        "val_ids = pd.read_pickle(input_folder + 'dumps/validation_data.pkl')['train_id'].tolist()\n",
        "test_ids = pd.read_pickle(input_folder + 'dumps/test_data.pkl')['train_id'].tolist()\n",
        "data = pd.read_pickle(input_folder + 'dumps/whole_data.pkl').drop(labels=['item_condition_id_1', 'item_condition_id_2', 'item_condition_id_3', 'item_condition_id_4', 'item_condition_id_5'], axis=1)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "dMqI8EGhIb-V",
        "colab_type": "text"
      },
      "source": [
        "### Creating word embeddings  using word2vec\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8RTicjY9KN4F",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Function to preprocess text for word2vec\n",
        "def textPreProcessing(allText): \n",
        "    #remove newlines with space\n",
        "    allText = allText.lower().replace(\"\\n\", \" \")\n",
        "    #remove punctuation\n",
        "    allText = allText.translate({ord(c): None for c in string.punctuation})\n",
        "    #remove multiple spaces\n",
        "    allText = re.sub(\" +\", \" \", allText)\n",
        "    #tokenize\n",
        "    tokens = allText.split(' ')\n",
        "    return tokens"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sBVStyDiJftj",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# train_data['name'] = train_data.name.str.translate({ord(c): ' ' for c in string.punctuation})\n",
        "data['name'] = data.name.apply(lambda x: textPreProcessing(x))\n",
        "data['item_description'] = data.item_description.apply(lambda x: textPreProcessing(x))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "u-vVQIzPHfNU",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def word_embeddings(word2VecModel, allTextList, dim=100):\n",
        "    word2Vec = word2VecModel.vocab.keys()\n",
        "    wordVector = [np.zeros(dim)]\n",
        "    for word in allTextList:\n",
        "      if word in word2Vec:\n",
        "        wordVector.append(word2VecModel[word][:dim])\n",
        "    wordVector = np.mean(wordVector, axis=0)\n",
        "    return wordVector"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "r4Cb0cqlH956",
        "colab_type": "code",
        "outputId": "b25b5d46-065a-4485-b662-f752152cb076",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 71
        }
      },
      "source": [
        "word2VecModel = gensim.models.KeyedVectors.load_word2vec_format(input_folder + 'wiki.en.bin.bin', binary = True )"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/smart_open/smart_open_lib.py:398: UserWarning: This function is deprecated, use smart_open.open instead. See the migration notes for details: https://github.com/RaRe-Technologies/smart_open/blob/master/README.rst#migrating-to-the-new-open-function\n",
            "  'See the migration notes for details: %s' % _MIGRATION_NOTES_URL\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "R0Q28dvZICYZ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Convert & expand columns 'name' & 'item_description' to 100 dim word vectore for each columns. This results in 200 extra columns\n",
        "data = pd.concat((data, data.apply(lambda x: word_embeddings(word2VecModel, x['name']), axis = 1, result_type='expand'), data.apply(lambda x: word_embeddings(word2VecModel, x['item_description']), axis = 1, result_type='expand')), axis=1)\n",
        "del word2VecModel\n",
        "data.to_pickle(input_folder + 'dumps/whole_data_word2vec.pkl')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "266Lf1n_6YMs",
        "colab_type": "text"
      },
      "source": [
        "<a id='restore_word2vec_data'></a>\n",
        "Now, removing 'name' & 'item_description' from our data"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LajcCW4EjPOp",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "data.drop(['name', 'item_description'], axis=1, inplace=True)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "tick7WgP6Tv0",
        "colab_type": "text"
      },
      "source": [
        "### Convert Categorical features into continuous features using Entity Embedding:  Embedding Preparation (Step 1)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lovcAT68A2jA",
        "colab_type": "text"
      },
      "source": [
        "### Sub-Feature Space"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1xzcR2fDn5Rv",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# rename duplicate columns of word2vec \n",
        "cols = pd.io.parsers.ParserBase({'names':data.columns})._maybe_dedup_names(data.columns)\n",
        "data.columns = cols"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_c-jyc71tk_a",
        "colab_type": "code",
        "outputId": "f2e41f99-aeff-4fb7-93e1-131925ad3f55",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 304
        }
      },
      "source": [
        "data.head()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "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>item_condition_id</th>\n",
              "      <th>category_name</th>\n",
              "      <th>price</th>\n",
              "      <th>shipping</th>\n",
              "      <th>item_description_readability_score</th>\n",
              "      <th>desc_len</th>\n",
              "      <th>name_len</th>\n",
              "      <th>cat_1</th>\n",
              "      <th>cat_2</th>\n",
              "      <th>cat_3</th>\n",
              "      <th>0</th>\n",
              "      <th>1</th>\n",
              "      <th>2</th>\n",
              "      <th>3</th>\n",
              "      <th>4</th>\n",
              "      <th>5</th>\n",
              "      <th>6</th>\n",
              "      <th>7</th>\n",
              "      <th>8</th>\n",
              "      <th>9</th>\n",
              "      <th>10</th>\n",
              "      <th>11</th>\n",
              "      <th>12</th>\n",
              "      <th>13</th>\n",
              "      <th>14</th>\n",
              "      <th>15</th>\n",
              "      <th>16</th>\n",
              "      <th>17</th>\n",
              "      <th>18</th>\n",
              "      <th>19</th>\n",
              "      <th>20</th>\n",
              "      <th>21</th>\n",
              "      <th>22</th>\n",
              "      <th>23</th>\n",
              "      <th>24</th>\n",
              "      <th>25</th>\n",
              "      <th>26</th>\n",
              "      <th>27</th>\n",
              "      <th>28</th>\n",
              "      <th>...</th>\n",
              "      <th>60.1</th>\n",
              "      <th>61.1</th>\n",
              "      <th>62.1</th>\n",
              "      <th>63.1</th>\n",
              "      <th>64.1</th>\n",
              "      <th>65.1</th>\n",
              "      <th>66.1</th>\n",
              "      <th>67.1</th>\n",
              "      <th>68.1</th>\n",
              "      <th>69.1</th>\n",
              "      <th>70.1</th>\n",
              "      <th>71.1</th>\n",
              "      <th>72.1</th>\n",
              "      <th>73.1</th>\n",
              "      <th>74.1</th>\n",
              "      <th>75.1</th>\n",
              "      <th>76.1</th>\n",
              "      <th>77.1</th>\n",
              "      <th>78.1</th>\n",
              "      <th>79.1</th>\n",
              "      <th>80.1</th>\n",
              "      <th>81.1</th>\n",
              "      <th>82.1</th>\n",
              "      <th>83.1</th>\n",
              "      <th>84.1</th>\n",
              "      <th>85.1</th>\n",
              "      <th>86.1</th>\n",
              "      <th>87.1</th>\n",
              "      <th>88.1</th>\n",
              "      <th>89.1</th>\n",
              "      <th>90.1</th>\n",
              "      <th>91.1</th>\n",
              "      <th>92.1</th>\n",
              "      <th>93.1</th>\n",
              "      <th>94.1</th>\n",
              "      <th>95.1</th>\n",
              "      <th>96.1</th>\n",
              "      <th>97.1</th>\n",
              "      <th>98.1</th>\n",
              "      <th>99.1</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>Men/Tops/T-shirts</td>\n",
              "      <td>-0.435170</td>\n",
              "      <td>1</td>\n",
              "      <td>99.00</td>\n",
              "      <td>1</td>\n",
              "      <td>7</td>\n",
              "      <td>Men</td>\n",
              "      <td>Tops</td>\n",
              "      <td>T-shirts</td>\n",
              "      <td>-0.047076</td>\n",
              "      <td>0.055123</td>\n",
              "      <td>0.096776</td>\n",
              "      <td>0.327872</td>\n",
              "      <td>-0.053400</td>\n",
              "      <td>-0.164231</td>\n",
              "      <td>0.123480</td>\n",
              "      <td>-0.097724</td>\n",
              "      <td>-0.025514</td>\n",
              "      <td>0.036671</td>\n",
              "      <td>0.278410</td>\n",
              "      <td>0.087304</td>\n",
              "      <td>-0.119336</td>\n",
              "      <td>0.021578</td>\n",
              "      <td>0.170031</td>\n",
              "      <td>-0.068347</td>\n",
              "      <td>-0.262792</td>\n",
              "      <td>0.085220</td>\n",
              "      <td>0.135382</td>\n",
              "      <td>0.257197</td>\n",
              "      <td>-0.033343</td>\n",
              "      <td>0.030893</td>\n",
              "      <td>-0.139916</td>\n",
              "      <td>-0.137116</td>\n",
              "      <td>0.101204</td>\n",
              "      <td>0.001052</td>\n",
              "      <td>-0.011945</td>\n",
              "      <td>-0.104977</td>\n",
              "      <td>-0.028738</td>\n",
              "      <td>...</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1</td>\n",
              "      <td>3</td>\n",
              "      <td>Electronics/Computers &amp; Tablets/Components &amp; P...</td>\n",
              "      <td>0.656049</td>\n",
              "      <td>0</td>\n",
              "      <td>8.13</td>\n",
              "      <td>36</td>\n",
              "      <td>5</td>\n",
              "      <td>Electronics</td>\n",
              "      <td>Computers &amp; Tablets</td>\n",
              "      <td>Components &amp; Parts</td>\n",
              "      <td>-0.120774</td>\n",
              "      <td>0.074269</td>\n",
              "      <td>-0.030055</td>\n",
              "      <td>0.025553</td>\n",
              "      <td>-0.201962</td>\n",
              "      <td>-0.062714</td>\n",
              "      <td>0.062537</td>\n",
              "      <td>-0.079967</td>\n",
              "      <td>0.040506</td>\n",
              "      <td>-0.000242</td>\n",
              "      <td>-0.121404</td>\n",
              "      <td>-0.110225</td>\n",
              "      <td>-0.192543</td>\n",
              "      <td>0.092441</td>\n",
              "      <td>0.236934</td>\n",
              "      <td>-0.005942</td>\n",
              "      <td>0.196020</td>\n",
              "      <td>-0.008345</td>\n",
              "      <td>0.047493</td>\n",
              "      <td>0.249163</td>\n",
              "      <td>-0.034059</td>\n",
              "      <td>0.198465</td>\n",
              "      <td>-0.045585</td>\n",
              "      <td>-0.253211</td>\n",
              "      <td>-0.335730</td>\n",
              "      <td>-0.180222</td>\n",
              "      <td>-0.164575</td>\n",
              "      <td>0.075460</td>\n",
              "      <td>0.113512</td>\n",
              "      <td>...</td>\n",
              "      <td>0.003353</td>\n",
              "      <td>-0.038637</td>\n",
              "      <td>0.157240</td>\n",
              "      <td>-0.036119</td>\n",
              "      <td>-0.045491</td>\n",
              "      <td>-0.108587</td>\n",
              "      <td>-0.072237</td>\n",
              "      <td>0.012078</td>\n",
              "      <td>-0.098535</td>\n",
              "      <td>-0.013373</td>\n",
              "      <td>-0.030974</td>\n",
              "      <td>0.013579</td>\n",
              "      <td>0.136929</td>\n",
              "      <td>-0.073364</td>\n",
              "      <td>-0.120016</td>\n",
              "      <td>0.062757</td>\n",
              "      <td>0.126728</td>\n",
              "      <td>0.174966</td>\n",
              "      <td>0.015804</td>\n",
              "      <td>-0.064356</td>\n",
              "      <td>0.096281</td>\n",
              "      <td>0.051810</td>\n",
              "      <td>-0.082231</td>\n",
              "      <td>0.058357</td>\n",
              "      <td>-0.034755</td>\n",
              "      <td>-0.016786</td>\n",
              "      <td>0.001092</td>\n",
              "      <td>0.065343</td>\n",
              "      <td>-0.037469</td>\n",
              "      <td>-0.078727</td>\n",
              "      <td>0.023271</td>\n",
              "      <td>0.045434</td>\n",
              "      <td>0.218250</td>\n",
              "      <td>-0.214543</td>\n",
              "      <td>0.151894</td>\n",
              "      <td>-0.005827</td>\n",
              "      <td>0.050133</td>\n",
              "      <td>0.017548</td>\n",
              "      <td>-0.006065</td>\n",
              "      <td>0.026098</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>Women/Tops &amp; Blouses/Blouse</td>\n",
              "      <td>-0.435170</td>\n",
              "      <td>1</td>\n",
              "      <td>5.80</td>\n",
              "      <td>29</td>\n",
              "      <td>3</td>\n",
              "      <td>Women</td>\n",
              "      <td>Tops &amp; Blouses</td>\n",
              "      <td>Blouse</td>\n",
              "      <td>-0.006465</td>\n",
              "      <td>0.141957</td>\n",
              "      <td>-0.238290</td>\n",
              "      <td>0.147485</td>\n",
              "      <td>0.055197</td>\n",
              "      <td>0.094854</td>\n",
              "      <td>-0.036636</td>\n",
              "      <td>0.084593</td>\n",
              "      <td>0.201683</td>\n",
              "      <td>0.078710</td>\n",
              "      <td>0.044293</td>\n",
              "      <td>-0.031158</td>\n",
              "      <td>-0.118623</td>\n",
              "      <td>0.040195</td>\n",
              "      <td>0.145833</td>\n",
              "      <td>-0.082050</td>\n",
              "      <td>0.004320</td>\n",
              "      <td>-0.073832</td>\n",
              "      <td>0.243750</td>\n",
              "      <td>0.285230</td>\n",
              "      <td>0.056097</td>\n",
              "      <td>0.049053</td>\n",
              "      <td>-0.012986</td>\n",
              "      <td>0.044872</td>\n",
              "      <td>0.093857</td>\n",
              "      <td>0.025824</td>\n",
              "      <td>0.053050</td>\n",
              "      <td>0.129570</td>\n",
              "      <td>0.017823</td>\n",
              "      <td>...</td>\n",
              "      <td>-0.003588</td>\n",
              "      <td>-0.092107</td>\n",
              "      <td>0.160697</td>\n",
              "      <td>-0.082413</td>\n",
              "      <td>-0.039508</td>\n",
              "      <td>-0.140503</td>\n",
              "      <td>-0.165885</td>\n",
              "      <td>0.025680</td>\n",
              "      <td>-0.086208</td>\n",
              "      <td>0.045485</td>\n",
              "      <td>-0.145674</td>\n",
              "      <td>-0.069647</td>\n",
              "      <td>0.032995</td>\n",
              "      <td>-0.136541</td>\n",
              "      <td>-0.073738</td>\n",
              "      <td>-0.010755</td>\n",
              "      <td>0.111959</td>\n",
              "      <td>0.116104</td>\n",
              "      <td>-0.021502</td>\n",
              "      <td>0.002786</td>\n",
              "      <td>0.025296</td>\n",
              "      <td>-0.020523</td>\n",
              "      <td>-0.088607</td>\n",
              "      <td>-0.049129</td>\n",
              "      <td>-0.074163</td>\n",
              "      <td>-0.042309</td>\n",
              "      <td>-0.040612</td>\n",
              "      <td>0.000737</td>\n",
              "      <td>-0.017215</td>\n",
              "      <td>-0.093759</td>\n",
              "      <td>-0.038437</td>\n",
              "      <td>-0.044795</td>\n",
              "      <td>0.142766</td>\n",
              "      <td>-0.126110</td>\n",
              "      <td>0.110071</td>\n",
              "      <td>0.016851</td>\n",
              "      <td>0.059525</td>\n",
              "      <td>0.116074</td>\n",
              "      <td>-0.080349</td>\n",
              "      <td>0.021306</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>3</td>\n",
              "      <td>1</td>\n",
              "      <td>Home/Home Décor/Home Décor Accents</td>\n",
              "      <td>0.214365</td>\n",
              "      <td>1</td>\n",
              "      <td>2.12</td>\n",
              "      <td>32</td>\n",
              "      <td>3</td>\n",
              "      <td>Home</td>\n",
              "      <td>Home Décor</td>\n",
              "      <td>Home Décor Accents</td>\n",
              "      <td>0.040486</td>\n",
              "      <td>0.072910</td>\n",
              "      <td>-0.064110</td>\n",
              "      <td>0.221228</td>\n",
              "      <td>-0.012104</td>\n",
              "      <td>0.212159</td>\n",
              "      <td>-0.067450</td>\n",
              "      <td>-0.134002</td>\n",
              "      <td>0.118982</td>\n",
              "      <td>0.148352</td>\n",
              "      <td>0.106880</td>\n",
              "      <td>0.146617</td>\n",
              "      <td>0.014503</td>\n",
              "      <td>0.045757</td>\n",
              "      <td>0.174585</td>\n",
              "      <td>0.002362</td>\n",
              "      <td>-0.070346</td>\n",
              "      <td>0.041782</td>\n",
              "      <td>-0.009823</td>\n",
              "      <td>0.125930</td>\n",
              "      <td>0.006232</td>\n",
              "      <td>-0.130985</td>\n",
              "      <td>-0.276471</td>\n",
              "      <td>0.055807</td>\n",
              "      <td>0.003177</td>\n",
              "      <td>-0.086490</td>\n",
              "      <td>0.036153</td>\n",
              "      <td>-0.011490</td>\n",
              "      <td>0.081834</td>\n",
              "      <td>...</td>\n",
              "      <td>-0.055100</td>\n",
              "      <td>-0.103526</td>\n",
              "      <td>0.173241</td>\n",
              "      <td>-0.062109</td>\n",
              "      <td>0.028187</td>\n",
              "      <td>-0.071492</td>\n",
              "      <td>-0.145623</td>\n",
              "      <td>0.103997</td>\n",
              "      <td>-0.098056</td>\n",
              "      <td>-0.034283</td>\n",
              "      <td>-0.030356</td>\n",
              "      <td>0.036376</td>\n",
              "      <td>0.160841</td>\n",
              "      <td>-0.136764</td>\n",
              "      <td>-0.049553</td>\n",
              "      <td>0.040392</td>\n",
              "      <td>0.189657</td>\n",
              "      <td>0.155428</td>\n",
              "      <td>0.025785</td>\n",
              "      <td>-0.072665</td>\n",
              "      <td>0.114834</td>\n",
              "      <td>0.054884</td>\n",
              "      <td>-0.076146</td>\n",
              "      <td>-0.028510</td>\n",
              "      <td>-0.047560</td>\n",
              "      <td>-0.018541</td>\n",
              "      <td>-0.021588</td>\n",
              "      <td>0.053808</td>\n",
              "      <td>0.006598</td>\n",
              "      <td>-0.070150</td>\n",
              "      <td>0.029136</td>\n",
              "      <td>-0.057160</td>\n",
              "      <td>0.148940</td>\n",
              "      <td>-0.159945</td>\n",
              "      <td>0.097450</td>\n",
              "      <td>0.151162</td>\n",
              "      <td>0.020571</td>\n",
              "      <td>-0.019095</td>\n",
              "      <td>-0.086801</td>\n",
              "      <td>0.045617</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>Women/Jewelry/Necklaces</td>\n",
              "      <td>0.448198</td>\n",
              "      <td>0</td>\n",
              "      <td>18.00</td>\n",
              "      <td>5</td>\n",
              "      <td>4</td>\n",
              "      <td>Women</td>\n",
              "      <td>Jewelry</td>\n",
              "      <td>Necklaces</td>\n",
              "      <td>0.112307</td>\n",
              "      <td>0.004079</td>\n",
              "      <td>-0.044416</td>\n",
              "      <td>0.120497</td>\n",
              "      <td>-0.011662</td>\n",
              "      <td>0.185782</td>\n",
              "      <td>0.026220</td>\n",
              "      <td>-0.208632</td>\n",
              "      <td>0.140207</td>\n",
              "      <td>0.068560</td>\n",
              "      <td>0.238925</td>\n",
              "      <td>0.084456</td>\n",
              "      <td>0.058310</td>\n",
              "      <td>-0.058165</td>\n",
              "      <td>0.189186</td>\n",
              "      <td>-0.193500</td>\n",
              "      <td>-0.243460</td>\n",
              "      <td>-0.195550</td>\n",
              "      <td>-0.266418</td>\n",
              "      <td>0.109663</td>\n",
              "      <td>0.006964</td>\n",
              "      <td>-0.048250</td>\n",
              "      <td>-0.108760</td>\n",
              "      <td>0.087225</td>\n",
              "      <td>-0.212959</td>\n",
              "      <td>-0.034300</td>\n",
              "      <td>0.039582</td>\n",
              "      <td>0.084704</td>\n",
              "      <td>0.114839</td>\n",
              "      <td>...</td>\n",
              "      <td>-0.073604</td>\n",
              "      <td>-0.096410</td>\n",
              "      <td>0.099351</td>\n",
              "      <td>-0.043521</td>\n",
              "      <td>-0.047153</td>\n",
              "      <td>-0.090752</td>\n",
              "      <td>-0.106993</td>\n",
              "      <td>0.039182</td>\n",
              "      <td>-0.105936</td>\n",
              "      <td>-0.084915</td>\n",
              "      <td>-0.081389</td>\n",
              "      <td>-0.009586</td>\n",
              "      <td>0.087161</td>\n",
              "      <td>-0.088486</td>\n",
              "      <td>0.013255</td>\n",
              "      <td>-0.058588</td>\n",
              "      <td>0.253675</td>\n",
              "      <td>0.121763</td>\n",
              "      <td>0.009180</td>\n",
              "      <td>-0.043005</td>\n",
              "      <td>0.204753</td>\n",
              "      <td>0.033616</td>\n",
              "      <td>-0.061259</td>\n",
              "      <td>-0.168861</td>\n",
              "      <td>-0.098418</td>\n",
              "      <td>-0.011999</td>\n",
              "      <td>0.015224</td>\n",
              "      <td>0.073108</td>\n",
              "      <td>-0.060641</td>\n",
              "      <td>-0.056723</td>\n",
              "      <td>-0.083615</td>\n",
              "      <td>0.040498</td>\n",
              "      <td>0.152760</td>\n",
              "      <td>-0.201911</td>\n",
              "      <td>0.039790</td>\n",
              "      <td>0.087265</td>\n",
              "      <td>-0.018196</td>\n",
              "      <td>0.043004</td>\n",
              "      <td>0.026020</td>\n",
              "      <td>0.029785</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>5 rows × 211 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "   train_id  item_condition_id  ...      98.1      99.1\n",
              "0         0                  3  ...  0.000000  0.000000\n",
              "1         1                  3  ... -0.006065  0.026098\n",
              "2         2                  1  ... -0.080349  0.021306\n",
              "3         3                  1  ... -0.086801  0.045617\n",
              "4         4                  1  ...  0.026020  0.029785\n",
              "\n",
              "[5 rows x 211 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "s80Ug79-b9LU",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "train_data = data.loc[train_ids]\n",
        "val_data = data.loc[val_ids]\n",
        "test_data = data.loc[test_ids]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VDd15j3BAvDu",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "target = ['price']\n",
        "\n",
        "features = data.columns.tolist()\n",
        "features.remove(target[0])\n",
        "features.remove('train_id')\n",
        "features.remove('category_name')\n",
        "features.remove('name_len')\n",
        "features.remove('desc_len')\n",
        "\n",
        "embed_cols = ['cat_1', 'cat_2', 'cat_3', 'item_condition_id']\n",
        "contin_vars = features.copy()\n",
        "for feat in embed_cols:\n",
        "  contin_vars.remove(feat)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KcnWoaz6K4sy",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "X_train, y_train, X_val, y_val, X_test, y_test = train_data[features], train_data[target], val_data[features], val_data[target], test_data[features], test_data[target]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "M_IxpHPPMI1x",
        "colab_type": "code",
        "outputId": "0fe78ee1-88bd-4456-eace-80a3add01c53",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 338
        }
      },
      "source": [
        "X_val.head()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "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>item_condition_id</th>\n",
              "      <th>shipping</th>\n",
              "      <th>item_description_readability_score</th>\n",
              "      <th>cat_1</th>\n",
              "      <th>cat_2</th>\n",
              "      <th>cat_3</th>\n",
              "      <th>0</th>\n",
              "      <th>1</th>\n",
              "      <th>2</th>\n",
              "      <th>3</th>\n",
              "      <th>4</th>\n",
              "      <th>5</th>\n",
              "      <th>6</th>\n",
              "      <th>7</th>\n",
              "      <th>8</th>\n",
              "      <th>9</th>\n",
              "      <th>10</th>\n",
              "      <th>11</th>\n",
              "      <th>12</th>\n",
              "      <th>13</th>\n",
              "      <th>14</th>\n",
              "      <th>15</th>\n",
              "      <th>16</th>\n",
              "      <th>17</th>\n",
              "      <th>18</th>\n",
              "      <th>19</th>\n",
              "      <th>20</th>\n",
              "      <th>21</th>\n",
              "      <th>22</th>\n",
              "      <th>23</th>\n",
              "      <th>24</th>\n",
              "      <th>25</th>\n",
              "      <th>26</th>\n",
              "      <th>27</th>\n",
              "      <th>28</th>\n",
              "      <th>29</th>\n",
              "      <th>30</th>\n",
              "      <th>31</th>\n",
              "      <th>32</th>\n",
              "      <th>33</th>\n",
              "      <th>...</th>\n",
              "      <th>60.1</th>\n",
              "      <th>61.1</th>\n",
              "      <th>62.1</th>\n",
              "      <th>63.1</th>\n",
              "      <th>64.1</th>\n",
              "      <th>65.1</th>\n",
              "      <th>66.1</th>\n",
              "      <th>67.1</th>\n",
              "      <th>68.1</th>\n",
              "      <th>69.1</th>\n",
              "      <th>70.1</th>\n",
              "      <th>71.1</th>\n",
              "      <th>72.1</th>\n",
              "      <th>73.1</th>\n",
              "      <th>74.1</th>\n",
              "      <th>75.1</th>\n",
              "      <th>76.1</th>\n",
              "      <th>77.1</th>\n",
              "      <th>78.1</th>\n",
              "      <th>79.1</th>\n",
              "      <th>80.1</th>\n",
              "      <th>81.1</th>\n",
              "      <th>82.1</th>\n",
              "      <th>83.1</th>\n",
              "      <th>84.1</th>\n",
              "      <th>85.1</th>\n",
              "      <th>86.1</th>\n",
              "      <th>87.1</th>\n",
              "      <th>88.1</th>\n",
              "      <th>89.1</th>\n",
              "      <th>90.1</th>\n",
              "      <th>91.1</th>\n",
              "      <th>92.1</th>\n",
              "      <th>93.1</th>\n",
              "      <th>94.1</th>\n",
              "      <th>95.1</th>\n",
              "      <th>96.1</th>\n",
              "      <th>97.1</th>\n",
              "      <th>98.1</th>\n",
              "      <th>99.1</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>1365791</th>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>4.00</td>\n",
              "      <td>Beauty</td>\n",
              "      <td>Fragrance</td>\n",
              "      <td>Women</td>\n",
              "      <td>-0.012190</td>\n",
              "      <td>0.029873</td>\n",
              "      <td>-0.240805</td>\n",
              "      <td>0.198942</td>\n",
              "      <td>-0.100929</td>\n",
              "      <td>-0.035169</td>\n",
              "      <td>-0.015586</td>\n",
              "      <td>-0.027224</td>\n",
              "      <td>-0.007503</td>\n",
              "      <td>0.270154</td>\n",
              "      <td>0.051441</td>\n",
              "      <td>-0.042243</td>\n",
              "      <td>0.178979</td>\n",
              "      <td>-0.011001</td>\n",
              "      <td>-0.025115</td>\n",
              "      <td>-0.178549</td>\n",
              "      <td>-0.112389</td>\n",
              "      <td>-0.080231</td>\n",
              "      <td>-0.061040</td>\n",
              "      <td>0.201338</td>\n",
              "      <td>-0.082679</td>\n",
              "      <td>0.153556</td>\n",
              "      <td>-0.232386</td>\n",
              "      <td>-0.220772</td>\n",
              "      <td>-0.093753</td>\n",
              "      <td>-0.017486</td>\n",
              "      <td>-0.054351</td>\n",
              "      <td>-0.124890</td>\n",
              "      <td>0.066865</td>\n",
              "      <td>0.181532</td>\n",
              "      <td>-0.024232</td>\n",
              "      <td>0.282230</td>\n",
              "      <td>-0.114021</td>\n",
              "      <td>0.122455</td>\n",
              "      <td>...</td>\n",
              "      <td>-0.067474</td>\n",
              "      <td>-0.047046</td>\n",
              "      <td>0.133540</td>\n",
              "      <td>-0.046317</td>\n",
              "      <td>-0.104770</td>\n",
              "      <td>-0.165336</td>\n",
              "      <td>-0.093953</td>\n",
              "      <td>0.015939</td>\n",
              "      <td>0.069700</td>\n",
              "      <td>0.060737</td>\n",
              "      <td>-0.092608</td>\n",
              "      <td>0.097383</td>\n",
              "      <td>0.039260</td>\n",
              "      <td>-0.126100</td>\n",
              "      <td>-0.108281</td>\n",
              "      <td>0.132013</td>\n",
              "      <td>0.147092</td>\n",
              "      <td>0.157484</td>\n",
              "      <td>-0.076523</td>\n",
              "      <td>-0.001736</td>\n",
              "      <td>0.014080</td>\n",
              "      <td>-0.001209</td>\n",
              "      <td>-0.042704</td>\n",
              "      <td>-0.071986</td>\n",
              "      <td>-0.007445</td>\n",
              "      <td>0.029889</td>\n",
              "      <td>0.006454</td>\n",
              "      <td>0.065194</td>\n",
              "      <td>0.019889</td>\n",
              "      <td>-0.019623</td>\n",
              "      <td>0.030608</td>\n",
              "      <td>0.029440</td>\n",
              "      <td>0.123867</td>\n",
              "      <td>-0.188649</td>\n",
              "      <td>0.130663</td>\n",
              "      <td>0.084876</td>\n",
              "      <td>0.111980</td>\n",
              "      <td>0.117493</td>\n",
              "      <td>-0.094717</td>\n",
              "      <td>0.083751</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>986670</th>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>9.29</td>\n",
              "      <td>Beauty</td>\n",
              "      <td>Tools &amp; Accessories</td>\n",
              "      <td>Makeup Brushes &amp; Tools</td>\n",
              "      <td>0.064924</td>\n",
              "      <td>0.100857</td>\n",
              "      <td>-0.177311</td>\n",
              "      <td>0.222203</td>\n",
              "      <td>-0.055053</td>\n",
              "      <td>-0.097597</td>\n",
              "      <td>0.078476</td>\n",
              "      <td>-0.084850</td>\n",
              "      <td>0.107998</td>\n",
              "      <td>0.159537</td>\n",
              "      <td>0.098485</td>\n",
              "      <td>0.075785</td>\n",
              "      <td>0.021189</td>\n",
              "      <td>0.051966</td>\n",
              "      <td>0.096036</td>\n",
              "      <td>-0.245698</td>\n",
              "      <td>-0.206459</td>\n",
              "      <td>-0.098937</td>\n",
              "      <td>-0.032707</td>\n",
              "      <td>0.222372</td>\n",
              "      <td>-0.153566</td>\n",
              "      <td>-0.043081</td>\n",
              "      <td>-0.131456</td>\n",
              "      <td>-0.111103</td>\n",
              "      <td>-0.081334</td>\n",
              "      <td>-0.111455</td>\n",
              "      <td>0.037815</td>\n",
              "      <td>0.040076</td>\n",
              "      <td>0.121637</td>\n",
              "      <td>0.013383</td>\n",
              "      <td>-0.113488</td>\n",
              "      <td>0.094832</td>\n",
              "      <td>-0.099189</td>\n",
              "      <td>0.182174</td>\n",
              "      <td>...</td>\n",
              "      <td>-0.043322</td>\n",
              "      <td>-0.050127</td>\n",
              "      <td>0.146413</td>\n",
              "      <td>-0.066345</td>\n",
              "      <td>0.032073</td>\n",
              "      <td>-0.097000</td>\n",
              "      <td>-0.138161</td>\n",
              "      <td>0.033366</td>\n",
              "      <td>-0.047446</td>\n",
              "      <td>0.026055</td>\n",
              "      <td>-0.101641</td>\n",
              "      <td>0.033396</td>\n",
              "      <td>0.032060</td>\n",
              "      <td>-0.095924</td>\n",
              "      <td>-0.094244</td>\n",
              "      <td>0.040442</td>\n",
              "      <td>0.174707</td>\n",
              "      <td>0.138396</td>\n",
              "      <td>0.016974</td>\n",
              "      <td>-0.031742</td>\n",
              "      <td>0.053740</td>\n",
              "      <td>0.071648</td>\n",
              "      <td>-0.053110</td>\n",
              "      <td>-0.040805</td>\n",
              "      <td>-0.089486</td>\n",
              "      <td>0.014546</td>\n",
              "      <td>-0.008924</td>\n",
              "      <td>0.080460</td>\n",
              "      <td>-0.018821</td>\n",
              "      <td>-0.064583</td>\n",
              "      <td>0.037078</td>\n",
              "      <td>-0.016789</td>\n",
              "      <td>0.227826</td>\n",
              "      <td>-0.188903</td>\n",
              "      <td>0.101062</td>\n",
              "      <td>0.059297</td>\n",
              "      <td>0.020905</td>\n",
              "      <td>0.020958</td>\n",
              "      <td>-0.061580</td>\n",
              "      <td>0.016407</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>651498</th>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>0.80</td>\n",
              "      <td>Women</td>\n",
              "      <td>Tops &amp; Blouses</td>\n",
              "      <td>T-Shirts</td>\n",
              "      <td>-0.050139</td>\n",
              "      <td>0.085613</td>\n",
              "      <td>0.009561</td>\n",
              "      <td>0.074084</td>\n",
              "      <td>0.000826</td>\n",
              "      <td>0.170170</td>\n",
              "      <td>0.069811</td>\n",
              "      <td>0.060261</td>\n",
              "      <td>0.022933</td>\n",
              "      <td>-0.039378</td>\n",
              "      <td>0.169179</td>\n",
              "      <td>0.086883</td>\n",
              "      <td>-0.132656</td>\n",
              "      <td>-0.030997</td>\n",
              "      <td>-0.019939</td>\n",
              "      <td>-0.086942</td>\n",
              "      <td>0.044034</td>\n",
              "      <td>0.028354</td>\n",
              "      <td>0.061850</td>\n",
              "      <td>0.139474</td>\n",
              "      <td>-0.048270</td>\n",
              "      <td>0.084365</td>\n",
              "      <td>-0.000242</td>\n",
              "      <td>-0.051999</td>\n",
              "      <td>0.229487</td>\n",
              "      <td>-0.127276</td>\n",
              "      <td>-0.135697</td>\n",
              "      <td>-0.034684</td>\n",
              "      <td>0.043894</td>\n",
              "      <td>0.070231</td>\n",
              "      <td>0.031445</td>\n",
              "      <td>0.146981</td>\n",
              "      <td>-0.009615</td>\n",
              "      <td>-0.054560</td>\n",
              "      <td>...</td>\n",
              "      <td>-0.020236</td>\n",
              "      <td>-0.173833</td>\n",
              "      <td>0.164927</td>\n",
              "      <td>-0.029310</td>\n",
              "      <td>-0.106280</td>\n",
              "      <td>-0.104965</td>\n",
              "      <td>-0.048157</td>\n",
              "      <td>-0.083455</td>\n",
              "      <td>-0.086498</td>\n",
              "      <td>0.104310</td>\n",
              "      <td>0.063494</td>\n",
              "      <td>-0.011256</td>\n",
              "      <td>-0.115367</td>\n",
              "      <td>-0.069666</td>\n",
              "      <td>0.027439</td>\n",
              "      <td>-0.081740</td>\n",
              "      <td>0.198063</td>\n",
              "      <td>0.158500</td>\n",
              "      <td>0.036404</td>\n",
              "      <td>0.102568</td>\n",
              "      <td>0.166170</td>\n",
              "      <td>0.016854</td>\n",
              "      <td>0.058331</td>\n",
              "      <td>0.009282</td>\n",
              "      <td>0.001924</td>\n",
              "      <td>-0.046797</td>\n",
              "      <td>0.033320</td>\n",
              "      <td>-0.128883</td>\n",
              "      <td>0.069148</td>\n",
              "      <td>-0.129010</td>\n",
              "      <td>-0.023203</td>\n",
              "      <td>0.214427</td>\n",
              "      <td>0.198050</td>\n",
              "      <td>-0.187947</td>\n",
              "      <td>0.154407</td>\n",
              "      <td>0.106637</td>\n",
              "      <td>-0.016431</td>\n",
              "      <td>-0.056499</td>\n",
              "      <td>-0.058958</td>\n",
              "      <td>-0.025788</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>107499</th>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>8.20</td>\n",
              "      <td>Electronics</td>\n",
              "      <td>Cell Phones &amp; Accessories</td>\n",
              "      <td>Cables &amp; Adapters</td>\n",
              "      <td>-0.068943</td>\n",
              "      <td>0.154125</td>\n",
              "      <td>0.067599</td>\n",
              "      <td>-0.091210</td>\n",
              "      <td>0.085836</td>\n",
              "      <td>0.085842</td>\n",
              "      <td>0.025464</td>\n",
              "      <td>0.079881</td>\n",
              "      <td>0.019828</td>\n",
              "      <td>0.112787</td>\n",
              "      <td>0.132463</td>\n",
              "      <td>-0.077508</td>\n",
              "      <td>0.093937</td>\n",
              "      <td>0.059388</td>\n",
              "      <td>0.125673</td>\n",
              "      <td>-0.034645</td>\n",
              "      <td>0.111879</td>\n",
              "      <td>0.075468</td>\n",
              "      <td>0.144868</td>\n",
              "      <td>0.074442</td>\n",
              "      <td>-0.202212</td>\n",
              "      <td>0.030239</td>\n",
              "      <td>-0.138610</td>\n",
              "      <td>-0.264715</td>\n",
              "      <td>-0.192195</td>\n",
              "      <td>-0.140730</td>\n",
              "      <td>-0.034809</td>\n",
              "      <td>-0.066247</td>\n",
              "      <td>0.054475</td>\n",
              "      <td>0.036618</td>\n",
              "      <td>-0.128621</td>\n",
              "      <td>0.186472</td>\n",
              "      <td>-0.124602</td>\n",
              "      <td>0.160424</td>\n",
              "      <td>...</td>\n",
              "      <td>0.055676</td>\n",
              "      <td>0.015008</td>\n",
              "      <td>0.005646</td>\n",
              "      <td>-0.015713</td>\n",
              "      <td>0.025200</td>\n",
              "      <td>-0.151504</td>\n",
              "      <td>-0.012315</td>\n",
              "      <td>-0.081559</td>\n",
              "      <td>-0.186804</td>\n",
              "      <td>-0.043387</td>\n",
              "      <td>0.050397</td>\n",
              "      <td>0.134087</td>\n",
              "      <td>0.248749</td>\n",
              "      <td>-0.119213</td>\n",
              "      <td>-0.089777</td>\n",
              "      <td>0.117574</td>\n",
              "      <td>0.106202</td>\n",
              "      <td>0.121184</td>\n",
              "      <td>-0.041585</td>\n",
              "      <td>-0.110279</td>\n",
              "      <td>0.067597</td>\n",
              "      <td>0.084913</td>\n",
              "      <td>0.034701</td>\n",
              "      <td>-0.120192</td>\n",
              "      <td>-0.093992</td>\n",
              "      <td>-0.101698</td>\n",
              "      <td>0.075702</td>\n",
              "      <td>0.157492</td>\n",
              "      <td>0.092456</td>\n",
              "      <td>0.062870</td>\n",
              "      <td>-0.032619</td>\n",
              "      <td>-0.051356</td>\n",
              "      <td>0.127510</td>\n",
              "      <td>0.025421</td>\n",
              "      <td>0.100677</td>\n",
              "      <td>0.027878</td>\n",
              "      <td>-0.118652</td>\n",
              "      <td>-0.045458</td>\n",
              "      <td>0.036423</td>\n",
              "      <td>0.011246</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>380562</th>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>3.60</td>\n",
              "      <td>Beauty</td>\n",
              "      <td>Makeup</td>\n",
              "      <td>Face</td>\n",
              "      <td>-0.042173</td>\n",
              "      <td>0.064546</td>\n",
              "      <td>-0.080930</td>\n",
              "      <td>0.173368</td>\n",
              "      <td>-0.184879</td>\n",
              "      <td>0.071995</td>\n",
              "      <td>0.120838</td>\n",
              "      <td>0.167566</td>\n",
              "      <td>0.057058</td>\n",
              "      <td>0.098058</td>\n",
              "      <td>0.383890</td>\n",
              "      <td>-0.008603</td>\n",
              "      <td>-0.024244</td>\n",
              "      <td>-0.081049</td>\n",
              "      <td>-0.110556</td>\n",
              "      <td>-0.053211</td>\n",
              "      <td>0.087859</td>\n",
              "      <td>-0.134017</td>\n",
              "      <td>0.025501</td>\n",
              "      <td>0.258850</td>\n",
              "      <td>-0.029576</td>\n",
              "      <td>0.195395</td>\n",
              "      <td>0.027596</td>\n",
              "      <td>-0.047979</td>\n",
              "      <td>-0.045614</td>\n",
              "      <td>-0.165561</td>\n",
              "      <td>-0.078852</td>\n",
              "      <td>0.013541</td>\n",
              "      <td>0.023611</td>\n",
              "      <td>0.085217</td>\n",
              "      <td>0.021601</td>\n",
              "      <td>0.263731</td>\n",
              "      <td>-0.131372</td>\n",
              "      <td>-0.021078</td>\n",
              "      <td>...</td>\n",
              "      <td>-0.035426</td>\n",
              "      <td>0.074640</td>\n",
              "      <td>0.217703</td>\n",
              "      <td>-0.116637</td>\n",
              "      <td>-0.024329</td>\n",
              "      <td>0.014202</td>\n",
              "      <td>-0.129988</td>\n",
              "      <td>0.066009</td>\n",
              "      <td>-0.128031</td>\n",
              "      <td>-0.068399</td>\n",
              "      <td>-0.096803</td>\n",
              "      <td>-0.036968</td>\n",
              "      <td>-0.014862</td>\n",
              "      <td>-0.125123</td>\n",
              "      <td>-0.009954</td>\n",
              "      <td>-0.093841</td>\n",
              "      <td>0.141546</td>\n",
              "      <td>-0.051876</td>\n",
              "      <td>-0.102803</td>\n",
              "      <td>-0.039546</td>\n",
              "      <td>-0.007021</td>\n",
              "      <td>-0.003102</td>\n",
              "      <td>-0.125072</td>\n",
              "      <td>0.042220</td>\n",
              "      <td>-0.092647</td>\n",
              "      <td>0.020989</td>\n",
              "      <td>-0.079747</td>\n",
              "      <td>-0.014150</td>\n",
              "      <td>0.147574</td>\n",
              "      <td>-0.046121</td>\n",
              "      <td>0.023539</td>\n",
              "      <td>-0.115821</td>\n",
              "      <td>0.071051</td>\n",
              "      <td>-0.086312</td>\n",
              "      <td>-0.003830</td>\n",
              "      <td>0.074819</td>\n",
              "      <td>0.030872</td>\n",
              "      <td>-0.046160</td>\n",
              "      <td>-0.103159</td>\n",
              "      <td>0.033319</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>5 rows × 206 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "         item_condition_id  shipping  ...      98.1      99.1\n",
              "1365791                  3         0  ... -0.094717  0.083751\n",
              "986670                   1         1  ... -0.061580  0.016407\n",
              "651498                   3         0  ... -0.058958 -0.025788\n",
              "107499                   1         1  ...  0.036423  0.011246\n",
              "380562                   1         1  ... -0.103159  0.033319\n",
              "\n",
              "[5 rows x 206 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 18
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vldjdLAcXRHb",
        "colab_type": "code",
        "outputId": "3a7c7981-6fcf-454d-fb48-5df3f4c957ae",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        }
      },
      "source": [
        "y_train.head()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "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>price</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>958469</th>\n",
              "      <td>-0.305263</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>693778</th>\n",
              "      <td>-0.227319</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>210050</th>\n",
              "      <td>-0.045449</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>447294</th>\n",
              "      <td>-0.071430</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1480041</th>\n",
              "      <td>-0.357226</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "            price\n",
              "958469  -0.305263\n",
              "693778  -0.227319\n",
              "210050  -0.045449\n",
              "447294  -0.071430\n",
              "1480041 -0.357226"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 19
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rsY9wMKc6Mvx",
        "colab_type": "text"
      },
      "source": [
        "### Categorical Features To List Format"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Vvg0amWP6F_k",
        "colab_type": "text"
      },
      "source": [
        "we need to convert data to list format to match the NN network structure.\n",
        "\n",
        "The following function takes the list of categorical features, and prepare such lists for the NN input."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rLdSZ8bGOp4C",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Ref: https://github.com/mmortazavi/EntityEmbedding-Working_Example/blob/master/EntityEmbedding.ipynb\n",
        "\n",
        "embed_val_map = {}\n",
        "for c in embed_cols:\n",
        "        raw_vals = np.unique(train_data[c])\n",
        "        val_map = {}\n",
        "        for i in range(len(raw_vals)):\n",
        "            val_map[raw_vals[i]] = i\n",
        "        embed_val_map[c] = val_map\n",
        "\n",
        "def preproc_generator(X, embed_val_map):\n",
        "    input_list = []\n",
        "    unk_in_embed = {'cat_1': 'unk_cat1', 'cat_2': 'unk_cat2', 'cat_3': 'unk_cat3', 'item_condition_id': 1}\n",
        "    \n",
        "    #the cols to be embedded: rescaling to range [0, # values)\n",
        "    for c in embed_cols:\n",
        "        val_map = embed_val_map[c]\n",
        "        input_list.append(X[c].map(val_map).fillna(val_map[unk_in_embed[c]]).values)\n",
        "     \n",
        "    #the rest of the columns\n",
        "    other_cols = [c for c in X.columns if (not c in embed_cols)]\n",
        "    input_list.append(X[other_cols].values)\n",
        "    \n",
        "    return input_list"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Jrqf0by__pO_",
        "colab_type": "text"
      },
      "source": [
        "### Embedding Dimension - Hyperparamter \n",
        "\n",
        "The choice of embedding dimension is optional, essentially it is a hyperparamter that one need to choose beforehand and reseach. \n",
        "\n",
        "##### One rule of thumb is to choose half of the cardinality of the categorical feature if that is is up to 50 in length."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TrKTNRk4O75M",
        "colab_type": "code",
        "outputId": "02476bf9-efcc-474c-eb7e-60f39131a410",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 85
        }
      },
      "source": [
        "for categorical_var in embed_cols:\n",
        "    \n",
        "    cat_emb_name= categorical_var.replace(\" \", \"\")+'_Embedding'\n",
        "  \n",
        "    no_of_unique_cat  = train_data[categorical_var].nunique()\n",
        "    embedding_size = int(min(np.ceil((no_of_unique_cat)/2), 50 ))\n",
        "  \n",
        "    print('Categorical Variable:', categorical_var,\n",
        "        'Unique Categories:', no_of_unique_cat,\n",
        "        'Embedding Size:', embedding_size)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Categorical Variable: cat_1 Unique Categories: 11 Embedding Size: 6\n",
            "Categorical Variable: cat_2 Unique Categories: 114 Embedding Size: 50\n",
            "Categorical Variable: cat_3 Unique Categories: 871 Embedding Size: 50\n",
            "Categorical Variable: item_condition_id Unique Categories: 5 Embedding Size: 3\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0ALAyYalAuzT",
        "colab_type": "text"
      },
      "source": [
        "#### Proper Naming of Categorical Features for Labelling NN Layers"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "outputId": "ca170823-38cd-4538-b14b-e766d0985b1d",
        "id": "3Gpl2LVrJ6UG",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 85
        }
      },
      "source": [
        "for categorical_var in embed_cols:\n",
        "    \n",
        "    input_name= 'Input_' + categorical_var.replace(\" \", \"\")\n",
        "    print(input_name)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Input_cat_1\n",
            "Input_cat_2\n",
            "Input_cat_3\n",
            "Input_item_condition_id\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "t7IEzr52ACdl",
        "colab_type": "text"
      },
      "source": [
        "# 3. Data Modelling: DNN with Entity Embedding "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OgWh8sA4Oy0M",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# RMSLE Custom Loss\n",
        "def rmsle(y_true, y_pred):\n",
        "    assert K.int_shape(y_true)[0] == K.int_shape(y_pred)[0]\n",
        "    return K.sqrt(K.mean(K.square(K.log(y_pred+1) - K.log(y_true+1))))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hLG0ohZ9A6uA",
        "colab_type": "text"
      },
      "source": [
        "###  Building Neural Network\n",
        "\n",
        "Here we basically make the embeding layers one at a time and append, and at the end we concatenate it together with the numerical features."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "saxiHCvPXMCJ",
        "colab_type": "code",
        "outputId": "76263d35-2736-42a0-de38-33e96a8f7884",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 309
        }
      },
      "source": [
        "input_models=[]\n",
        "output_embeddings=[]\n",
        "numerics = ['uint8', 'int16', 'int32', 'int64', 'float16', 'float32', 'float64']\n",
        "\n",
        "for categorical_var in embed_cols:\n",
        "    \n",
        "    #Name of the categorical variable that will be used in the Keras Embedding layer\n",
        "    cat_emb_name= categorical_var.replace(\" \", \"\")+'_Embedding'\n",
        "  \n",
        "    # Define the embedding_size\n",
        "    no_of_unique_cat  = train_data[categorical_var].nunique()\n",
        "    embedding_size = int(min(np.ceil((no_of_unique_cat)/2), 50 ))\n",
        "  \n",
        "    #One Embedding Layer for each categorical variable\n",
        "    input_model = Input(shape=(1,))\n",
        "    output_model = Embedding(no_of_unique_cat, embedding_size, name=cat_emb_name)(input_model)\n",
        "    output_model = Reshape(target_shape=(embedding_size,))(output_model)    \n",
        "  \n",
        "    #Appending all the categorical inputs\n",
        "    input_models.append(input_model)\n",
        "  \n",
        "    #Appending all the embeddings\n",
        "    output_embeddings.append(output_model)\n",
        "  \n",
        "#Other non-categorical data columns (numerical). \n",
        "#We define another network for the other columns and add them to our models list.\n",
        "input_numeric = Input(shape=(len(contin_vars),)) #-1 for train_id\n",
        "embedding_numeric = Dense(128)(input_numeric) \n",
        "input_models.append(input_numeric)\n",
        "output_embeddings.append(embedding_numeric)\n",
        "\n",
        "#At the end we concatenate altogther and add other Dense layers\n",
        "output = Concatenate()(output_embeddings)\n",
        "output = Dense(1024, kernel_initializer=\"glorot_uniform\")(output)\n",
        "output = Activation('relu')(output)\n",
        "output= Dropout(0.4)(output)\n",
        "output = Dense(512, kernel_initializer=\"glorot_uniform\")(output)\n",
        "output = Activation('relu')(output)\n",
        "output= Dropout(0.3)(output)\n",
        "output = Dense(1, activation='sigmoid')(output)\n",
        "\n",
        "model = Model(inputs=input_models, outputs=output)\n",
        "# opt = keras.optimizers.Adadelta(lr=1.0, rho=0.95, epsilon=None, decay=0.0)\n",
        "model.compile(loss='mse', optimizer='adam', metrics=[rmsle])"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "WARNING: Logging before flag parsing goes to stderr.\n",
            "W0807 12:58:10.271957 139716535281536 deprecation_wrapper.py:119] From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:74: The name tf.get_default_graph is deprecated. Please use tf.compat.v1.get_default_graph instead.\n",
            "\n",
            "W0807 12:58:10.328999 139716535281536 deprecation_wrapper.py:119] From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:517: The name tf.placeholder is deprecated. Please use tf.compat.v1.placeholder instead.\n",
            "\n",
            "W0807 12:58:10.360865 139716535281536 deprecation_wrapper.py:119] From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:4138: The name tf.random_uniform is deprecated. Please use tf.random.uniform instead.\n",
            "\n",
            "W0807 12:58:11.199496 139716535281536 deprecation_wrapper.py:119] From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:133: The name tf.placeholder_with_default is deprecated. Please use tf.compat.v1.placeholder_with_default instead.\n",
            "\n",
            "W0807 12:58:11.208280 139716535281536 deprecation.py:506] From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:3445: calling dropout (from tensorflow.python.ops.nn_ops) with keep_prob is deprecated and will be removed in a future version.\n",
            "Instructions for updating:\n",
            "Please use `rate` instead of `keep_prob`. Rate should be set to `rate = 1 - keep_prob`.\n",
            "W0807 12:58:11.271213 139716535281536 deprecation_wrapper.py:119] From /usr/local/lib/python3.6/dist-packages/keras/optimizers.py:790: The name tf.train.Optimizer is deprecated. Please use tf.compat.v1.train.Optimizer instead.\n",
            "\n",
            "W0807 12:58:11.308475 139716535281536 deprecation_wrapper.py:119] From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:1521: The name tf.log is deprecated. Please use tf.math.log instead.\n",
            "\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yonTOg8ElKHg",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# K.clear_session()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Smsmi0rJZL9j",
        "colab_type": "code",
        "outputId": "56cdaf4b-f174-4394-bbef-b304fb46300f",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 607
        }
      },
      "source": [
        "plot_model(model, show_shapes=True, show_layer_names=True, to_file= input_folder + 'dumps/ee_dnn_model.png')\n",
        "from IPython.display import Image\n",
        "Image(retina=True, filename=input_folder + 'dumps/ee_dnn_model.png')"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAACXgAAAScCAIAAAC00eKHAAAABmJLR0QA/wD/AP+gvaeTAAAgAElE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\nAAAAAAAAAIDR/g99q+5Dv/6w+wAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {
            "tags": [],
            "image/png": {
              "width": 1212,
              "height": 590
            }
          },
          "execution_count": 40
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kDjHo-rAZOI4",
        "colab_type": "code",
        "outputId": "1891c5b9-4ca8-4946-d8b3-bb0e7cfe4bc8",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 952
        }
      },
      "source": [
        "model.summary()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "input_1 (InputLayer)            (None, 1)            0                                            \n",
            "__________________________________________________________________________________________________\n",
            "input_2 (InputLayer)            (None, 1)            0                                            \n",
            "__________________________________________________________________________________________________\n",
            "input_3 (InputLayer)            (None, 1)            0                                            \n",
            "__________________________________________________________________________________________________\n",
            "input_4 (InputLayer)            (None, 1)            0                                            \n",
            "__________________________________________________________________________________________________\n",
            "cat_1_Embedding (Embedding)     (None, 1, 6)         66          input_1[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "cat_2_Embedding (Embedding)     (None, 1, 50)        5700        input_2[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "cat_3_Embedding (Embedding)     (None, 1, 50)        43550       input_3[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "item_condition_id_Embedding (Em (None, 1, 3)         15          input_4[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "input_5 (InputLayer)            (None, 202)          0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape_1 (Reshape)             (None, 6)            0           cat_1_Embedding[0][0]            \n",
            "__________________________________________________________________________________________________\n",
            "reshape_2 (Reshape)             (None, 50)           0           cat_2_Embedding[0][0]            \n",
            "__________________________________________________________________________________________________\n",
            "reshape_3 (Reshape)             (None, 50)           0           cat_3_Embedding[0][0]            \n",
            "__________________________________________________________________________________________________\n",
            "reshape_4 (Reshape)             (None, 3)            0           item_condition_id_Embedding[0][0]\n",
            "__________________________________________________________________________________________________\n",
            "dense_1 (Dense)                 (None, 128)          25984       input_5[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate_1 (Concatenate)     (None, 237)          0           reshape_1[0][0]                  \n",
            "                                                                 reshape_2[0][0]                  \n",
            "                                                                 reshape_3[0][0]                  \n",
            "                                                                 reshape_4[0][0]                  \n",
            "                                                                 dense_1[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "dense_2 (Dense)                 (None, 1024)         243712      concatenate_1[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "activation_1 (Activation)       (None, 1024)         0           dense_2[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "dropout_1 (Dropout)             (None, 1024)         0           activation_1[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "dense_3 (Dense)                 (None, 512)          524800      dropout_1[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_2 (Activation)       (None, 512)          0           dense_3[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "dropout_2 (Dropout)             (None, 512)          0           activation_2[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "dense_4 (Dense)                 (None, 1)            513         dropout_2[0][0]                  \n",
            "==================================================================================================\n",
            "Total params: 844,340\n",
            "Trainable params: 844,340\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "teUQBpnUGAcZ",
        "colab_type": "text"
      },
      "source": [
        "### Training neural network"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ti45wskuGVwT",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "history = History()\n",
        "\n",
        "callbacks = [history, \n",
        "             EarlyStopping(monitor='val_rmsle', patience=10, verbose=1, min_delta=1e-4, restore_best_weights=True, mode='min'),\n",
        "             ReduceLROnPlateau(monitor='val_rmsle', factor=0.1, patience=7, cooldown=10, min_lr=1e-7, verbose=1, mode='min'),\n",
        "             ModelCheckpoint(filepath=input_folder + 'model_word2vec.best.weights.hdf5', \n",
        "             monitor='val_rmsle', verbose=1, save_weights_only=True,  \n",
        "             save_best_only=True, mode='min')]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7rmnAFtErpTK",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "X_train_list = preproc_generator(X_train, embed_val_map)\n",
        "X_val_list = preproc_generator(X_val, embed_val_map)\n",
        "X_test_list = preproc_generator(X_test, embed_val_map)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0MiDao4GZcTe",
        "colab_type": "code",
        "outputId": "bf2b51af-8f59-4e58-ce93-401d77321f8b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "history  =  model.fit(X_train_list,y_train,validation_data=(X_val_list,y_val) , epochs = 1200 , batch_size = 512, verbose= 1, callbacks=callbacks)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Train on 1333492 samples, validate on 74083 samples\n",
            "Epoch 1/1200\n",
            "1333492/1333492 [==============================] - 41s 31us/step - loss: 0.8995 - rmsle: 0.5230 - val_loss: 0.7688 - val_rmsle: 0.5071\n",
            "\n",
            "Epoch 00001: val_rmsle improved from inf to 0.50708, saving model to /content/gdrive/My Drive/DMT/mercari-price-suggestion-challenge/model_word2vec.best.weights.hdf5\n",
            "Epoch 2/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.8278 - rmsle: 0.5061 - val_loss: 0.7593 - val_rmsle: 0.5025\n",
            "\n",
            "Epoch 00002: val_rmsle improved from 0.50708 to 0.50253, saving model to /content/gdrive/My Drive/DMT/mercari-price-suggestion-challenge/model_word2vec.best.weights.hdf5\n",
            "Epoch 3/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.8205 - rmsle: 0.5031 - val_loss: 0.7495 - val_rmsle: 0.4976\n",
            "\n",
            "Epoch 00003: val_rmsle improved from 0.50253 to 0.49758, saving model to /content/gdrive/My Drive/DMT/mercari-price-suggestion-challenge/model_word2vec.best.weights.hdf5\n",
            "Epoch 4/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.8147 - rmsle: 0.5008 - val_loss: 0.7470 - val_rmsle: 0.4977\n",
            "\n",
            "Epoch 00004: val_rmsle did not improve from 0.49758\n",
            "Epoch 5/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.8115 - rmsle: 0.4996 - val_loss: 0.7487 - val_rmsle: 0.4970\n",
            "\n",
            "Epoch 00005: val_rmsle improved from 0.49758 to 0.49695, saving model to /content/gdrive/My Drive/DMT/mercari-price-suggestion-challenge/model_word2vec.best.weights.hdf5\n",
            "Epoch 6/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.8088 - rmsle: 0.4985 - val_loss: 0.7518 - val_rmsle: 0.4953\n",
            "\n",
            "Epoch 00006: val_rmsle improved from 0.49695 to 0.49535, saving model to /content/gdrive/My Drive/DMT/mercari-price-suggestion-challenge/model_word2vec.best.weights.hdf5\n",
            "Epoch 7/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.8063 - rmsle: 0.4974 - val_loss: 0.7429 - val_rmsle: 0.4967\n",
            "\n",
            "Epoch 00007: val_rmsle did not improve from 0.49535\n",
            "Epoch 8/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.8049 - rmsle: 0.4969 - val_loss: 0.7415 - val_rmsle: 0.4949\n",
            "\n",
            "Epoch 00008: val_rmsle improved from 0.49535 to 0.49492, saving model to /content/gdrive/My Drive/DMT/mercari-price-suggestion-challenge/model_word2vec.best.weights.hdf5\n",
            "Epoch 9/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.8033 - rmsle: 0.4962 - val_loss: 0.7393 - val_rmsle: 0.4951\n",
            "\n",
            "Epoch 00009: val_rmsle did not improve from 0.49492\n",
            "Epoch 10/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.8016 - rmsle: 0.4955 - val_loss: 0.7383 - val_rmsle: 0.4936\n",
            "\n",
            "Epoch 00010: val_rmsle improved from 0.49492 to 0.49361, saving model to /content/gdrive/My Drive/DMT/mercari-price-suggestion-challenge/model_word2vec.best.weights.hdf5\n",
            "Epoch 11/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.8012 - rmsle: 0.4953 - val_loss: 0.7364 - val_rmsle: 0.4941\n",
            "\n",
            "Epoch 00011: val_rmsle did not improve from 0.49361\n",
            "Epoch 12/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.7990 - rmsle: 0.4944 - val_loss: 0.7380 - val_rmsle: 0.4938\n",
            "\n",
            "Epoch 00012: val_rmsle did not improve from 0.49361\n",
            "Epoch 13/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.7980 - rmsle: 0.4940 - val_loss: 0.7364 - val_rmsle: 0.4926\n",
            "\n",
            "Epoch 00013: val_rmsle improved from 0.49361 to 0.49263, saving model to /content/gdrive/My Drive/DMT/mercari-price-suggestion-challenge/model_word2vec.best.weights.hdf5\n",
            "Epoch 14/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.7973 - rmsle: 0.4936 - val_loss: 0.7364 - val_rmsle: 0.4924\n",
            "\n",
            "Epoch 00014: val_rmsle improved from 0.49263 to 0.49240, saving model to /content/gdrive/My Drive/DMT/mercari-price-suggestion-challenge/model_word2vec.best.weights.hdf5\n",
            "Epoch 15/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.7973 - rmsle: 0.4937 - val_loss: 0.7376 - val_rmsle: 0.4923\n",
            "\n",
            "Epoch 00015: val_rmsle improved from 0.49240 to 0.49235, saving model to /content/gdrive/My Drive/DMT/mercari-price-suggestion-challenge/model_word2vec.best.weights.hdf5\n",
            "Epoch 16/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.7959 - rmsle: 0.4931 - val_loss: 0.7351 - val_rmsle: 0.4932\n",
            "\n",
            "Epoch 00016: val_rmsle did not improve from 0.49235\n",
            "Epoch 17/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.7956 - rmsle: 0.4929 - val_loss: 0.7352 - val_rmsle: 0.4923\n",
            "\n",
            "Epoch 00017: val_rmsle improved from 0.49235 to 0.49234, saving model to /content/gdrive/My Drive/DMT/mercari-price-suggestion-challenge/model_word2vec.best.weights.hdf5\n",
            "Epoch 18/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.7950 - rmsle: 0.4927 - val_loss: 0.7335 - val_rmsle: 0.4922\n",
            "\n",
            "Epoch 00018: val_rmsle improved from 0.49234 to 0.49223, saving model to /content/gdrive/My Drive/DMT/mercari-price-suggestion-challenge/model_word2vec.best.weights.hdf5\n",
            "Epoch 19/1200\n",
            "1333492/1333492 [==============================] - 38s 28us/step - loss: 0.7934 - rmsle: 0.4921 - val_loss: 0.7357 - val_rmsle: 0.4913\n",
            "\n",
            "Epoch 00019: val_rmsle improved from 0.49223 to 0.49127, saving model to /content/gdrive/My Drive/DMT/mercari-price-suggestion-challenge/model_word2vec.best.weights.hdf5\n",
            "Epoch 20/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.7938 - rmsle: 0.4921 - val_loss: 0.7334 - val_rmsle: 0.4916\n",
            "\n",
            "Epoch 00020: val_rmsle did not improve from 0.49127\n",
            "Epoch 21/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.7927 - rmsle: 0.4917 - val_loss: 0.7364 - val_rmsle: 0.4927\n",
            "\n",
            "Epoch 00021: val_rmsle did not improve from 0.49127\n",
            "Epoch 22/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.7922 - rmsle: 0.4915 - val_loss: 0.7337 - val_rmsle: 0.4909\n",
            "\n",
            "Epoch 00022: val_rmsle improved from 0.49127 to 0.49092, saving model to /content/gdrive/My Drive/DMT/mercari-price-suggestion-challenge/model_word2vec.best.weights.hdf5\n",
            "Epoch 23/1200\n",
            "1333492/1333492 [==============================] - 38s 28us/step - loss: 0.7913 - rmsle: 0.4912 - val_loss: 0.7336 - val_rmsle: 0.4918\n",
            "\n",
            "Epoch 00023: val_rmsle did not improve from 0.49092\n",
            "Epoch 24/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.7908 - rmsle: 0.4909 - val_loss: 0.7340 - val_rmsle: 0.4923\n",
            "\n",
            "Epoch 00024: val_rmsle did not improve from 0.49092\n",
            "Epoch 25/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.7906 - rmsle: 0.4909 - val_loss: 0.7338 - val_rmsle: 0.4929\n",
            "\n",
            "Epoch 00025: val_rmsle did not improve from 0.49092\n",
            "Epoch 26/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.7909 - rmsle: 0.4911 - val_loss: 0.7334 - val_rmsle: 0.4906\n",
            "\n",
            "Epoch 00026: val_rmsle improved from 0.49092 to 0.49056, saving model to /content/gdrive/My Drive/DMT/mercari-price-suggestion-challenge/model_word2vec.best.weights.hdf5\n",
            "Epoch 27/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.7898 - rmsle: 0.4905 - val_loss: 0.7319 - val_rmsle: 0.4901\n",
            "\n",
            "Epoch 00027: val_rmsle improved from 0.49056 to 0.49006, saving model to /content/gdrive/My Drive/DMT/mercari-price-suggestion-challenge/model_word2vec.best.weights.hdf5\n",
            "Epoch 28/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.7889 - rmsle: 0.4900 - val_loss: 0.7319 - val_rmsle: 0.4907\n",
            "\n",
            "Epoch 00028: val_rmsle did not improve from 0.49006\n",
            "Epoch 29/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.7893 - rmsle: 0.4902 - val_loss: 0.7323 - val_rmsle: 0.4908\n",
            "\n",
            "Epoch 00029: val_rmsle did not improve from 0.49006\n",
            "Epoch 30/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.7887 - rmsle: 0.4900 - val_loss: 0.7320 - val_rmsle: 0.4905\n",
            "\n",
            "Epoch 00030: val_rmsle did not improve from 0.49006\n",
            "Epoch 31/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.7885 - rmsle: 0.4899 - val_loss: 0.7321 - val_rmsle: 0.4910\n",
            "\n",
            "Epoch 00031: val_rmsle did not improve from 0.49006\n",
            "Epoch 32/1200\n",
            "1333492/1333492 [==============================] - 38s 28us/step - loss: 0.7880 - rmsle: 0.4897 - val_loss: 0.7334 - val_rmsle: 0.4925\n",
            "\n",
            "Epoch 00032: val_rmsle did not improve from 0.49006\n",
            "Epoch 33/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.7883 - rmsle: 0.4898 - val_loss: 0.7318 - val_rmsle: 0.4910\n",
            "\n",
            "Epoch 00033: val_rmsle did not improve from 0.49006\n",
            "Epoch 34/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.7873 - rmsle: 0.4894 - val_loss: 0.7312 - val_rmsle: 0.4908\n",
            "\n",
            "Epoch 00034: ReduceLROnPlateau reducing learning rate to 0.00010000000474974513.\n",
            "\n",
            "Epoch 00034: val_rmsle did not improve from 0.49006\n",
            "Epoch 35/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.7822 - rmsle: 0.4873 - val_loss: 0.7293 - val_rmsle: 0.4889\n",
            "\n",
            "Epoch 00035: val_rmsle improved from 0.49006 to 0.48893, saving model to /content/gdrive/My Drive/DMT/mercari-price-suggestion-challenge/model_word2vec.best.weights.hdf5\n",
            "Epoch 36/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.7805 - rmsle: 0.4865 - val_loss: 0.7292 - val_rmsle: 0.4884\n",
            "\n",
            "Epoch 00036: val_rmsle improved from 0.48893 to 0.48845, saving model to /content/gdrive/My Drive/DMT/mercari-price-suggestion-challenge/model_word2vec.best.weights.hdf5\n",
            "Epoch 37/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.7798 - rmsle: 0.4862 - val_loss: 0.7291 - val_rmsle: 0.4883\n",
            "\n",
            "Epoch 00037: val_rmsle improved from 0.48845 to 0.48832, saving model to /content/gdrive/My Drive/DMT/mercari-price-suggestion-challenge/model_word2vec.best.weights.hdf5\n",
            "Epoch 38/1200\n",
            "1333492/1333492 [==============================] - 38s 28us/step - loss: 0.7792 - rmsle: 0.4859 - val_loss: 0.7286 - val_rmsle: 0.4888\n",
            "\n",
            "Epoch 00038: val_rmsle did not improve from 0.48832\n",
            "Epoch 39/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.7788 - rmsle: 0.4857 - val_loss: 0.7287 - val_rmsle: 0.4887\n",
            "\n",
            "Epoch 00039: val_rmsle did not improve from 0.48832\n",
            "Epoch 40/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.7787 - rmsle: 0.4856 - val_loss: 0.7286 - val_rmsle: 0.4889\n",
            "\n",
            "Epoch 00040: val_rmsle did not improve from 0.48832\n",
            "Epoch 41/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.7781 - rmsle: 0.4855 - val_loss: 0.7285 - val_rmsle: 0.4884\n",
            "\n",
            "Epoch 00041: val_rmsle did not improve from 0.48832\n",
            "Epoch 42/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.7781 - rmsle: 0.4853 - val_loss: 0.7282 - val_rmsle: 0.4887\n",
            "\n",
            "Epoch 00042: val_rmsle did not improve from 0.48832\n",
            "Epoch 43/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.7776 - rmsle: 0.4852 - val_loss: 0.7284 - val_rmsle: 0.4888\n",
            "\n",
            "Epoch 00043: val_rmsle did not improve from 0.48832\n",
            "Epoch 44/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.7774 - rmsle: 0.4851 - val_loss: 0.7287 - val_rmsle: 0.4886\n",
            "\n",
            "Epoch 00044: val_rmsle did not improve from 0.48832\n",
            "Epoch 45/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.7771 - rmsle: 0.4850 - val_loss: 0.7285 - val_rmsle: 0.4885\n",
            "\n",
            "Epoch 00045: val_rmsle did not improve from 0.48832\n",
            "Epoch 46/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.7769 - rmsle: 0.4849 - val_loss: 0.7284 - val_rmsle: 0.4883\n",
            "\n",
            "Epoch 00046: val_rmsle improved from 0.48832 to 0.48831, saving model to /content/gdrive/My Drive/DMT/mercari-price-suggestion-challenge/model_word2vec.best.weights.hdf5\n",
            "Epoch 47/1200\n",
            "1333492/1333492 [==============================] - 37s 28us/step - loss: 0.7768 - rmsle: 0.4848 - val_loss: 0.7284 - val_rmsle: 0.4885\n",
            "Restoring model weights from the end of the best epoch\n",
            "\n",
            "Epoch 00047: val_rmsle did not improve from 0.48831\n",
            "Epoch 00047: early stopping\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1bMt1gaQZisy",
        "colab_type": "code",
        "outputId": "b02b29a3-f2ed-4f86-e8b8-5acc0eec1387",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        }
      },
      "source": [
        "# summarize history for accuracy\n",
        "plt.plot(history.history['rmsle'])\n",
        "plt.plot(history.history['val_rmsle'])\n",
        "plt.title('DNN with Entity Embedding')\n",
        "plt.ylabel('rmsle')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['train', 'val'], loc='upper right')\n",
        "plt.show()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZIAAAEWCAYAAABMoxE0AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nWAAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3jl0nX1Zcwh-",
        "colab_type": "code",
        "outputId": "87f2c4de-ddce-4dda-81ca-4a3b71ca362d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "#List out Hyperparameters ???\n",
        "history.history['val_rmsle'][-1]"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.48851542824000127"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 33
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "miBTE0fGL5Qb",
        "colab_type": "text"
      },
      "source": [
        "# 4. Data Modelling: XgBoost with Entity Embedding "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PFt4nccaN0lj",
        "colab_type": "text"
      },
      "source": [
        "Run the above code atleast till the DNN model compilation cell."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pS7LNHo6oZgb",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "model.load_weights(filepath=input_folder + 'model_word2vec.best.weights.hdf5')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hIG3SpD1VDAH",
        "colab_type": "text"
      },
      "source": [
        "After training of our model, a new category can come up in test set that was not previously in the train dataset. We need to handle that for XgBoost. In the DNN, we handled the same in the prepoc_generator function."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yY5Cyv0MOJZu",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# This function compares the unique values in test data with that of train, and replaces any new unique value in the test set with UNK token.  \n",
        "def mark_unseen_cat_as_unk(X_train, X_test):\n",
        "  unk_in_embed = {'cat_1': 'unk_cat1', 'cat_2': 'unk_cat2', 'cat_3': 'unk_cat3', 'item_condition_id': 1}\n",
        "  for c in embed_cols:\n",
        "    for unk_cat in set(X_test[c].unique()).difference(set(X_train[c].unique())):\n",
        "      X_test.loc[X_test[c] == unk_cat, c] = unk_in_embed[c]  \n",
        "  return X_test\n",
        "\n",
        "val_data = mark_unseen_cat_as_unk(train_data, val_data)\n",
        "test_data = mark_unseen_cat_as_unk(train_data, test_data)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LJhXsPX6VUJO",
        "colab_type": "text"
      },
      "source": [
        "Extract the embedding layer weights from the trained DNN model. These embedding layer weights replace the categorical attribute with N-dim vectors that are able to capture richer & complex relationships between those categories that OHE can not. "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5EFDUW1Kdrx-",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "cat_1_Embedding = model.get_layer('cat_1_Embedding').get_weights()[0]\n",
        "cat_2_Embedding = model.get_layer('cat_2_Embedding').get_weights()[0]\n",
        "cat_3_Embedding = model.get_layer('cat_3_Embedding').get_weights()[0]\n",
        "item_condition_id_Embedding  = model.get_layer('item_condition_id_Embedding').get_weights()[0]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "EQyIBQ1dV4hN",
        "colab_type": "text"
      },
      "source": [
        "**Now, we will replace each value of categorical attribute columns with their learnt embedding vectors.**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZzIHiGkQmuSc",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "embed_map = {'cat_1': cat_1_Embedding, 'cat_2': cat_2_Embedding, 'cat_3': cat_3_Embedding, 'item_condition_id': item_condition_id_Embedding}"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GcyVNAJRh8IH",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "embed_val_map = {}\n",
        "for c in embed_cols:\n",
        "        raw_vals = np.unique(train_data[c])\n",
        "        val_map = {}\n",
        "        for i in range(len(raw_vals)):\n",
        "            val_map[raw_vals[i]] = embed_map[c][i]\n",
        "        embed_val_map[c] = val_map"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lX7ZG90OLbk5",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "train_data = pd.concat((train_data, \n",
        "                        train_data.apply(lambda x: embed_val_map['cat_1'][x['cat_1']], axis=1, result_type='expand'), \n",
        "                        train_data.apply(lambda x: embed_val_map['cat_2'][x['cat_2']], axis=1, result_type='expand'),\n",
        "                        train_data.apply(lambda x: embed_val_map['cat_3'][x['cat_3']], axis=1, result_type='expand'),\n",
        "                        train_data.apply(lambda x: embed_val_map['item_condition_id'][x['item_condition_id']], axis=1, result_type='expand')),\n",
        "                        axis=1)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "18ZaOivXT2b3",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "val_data = pd.concat((val_data, \n",
        "                        val_data.apply(lambda x: embed_val_map['cat_1'][x['cat_1']], axis=1, result_type='expand'), \n",
        "                        val_data.apply(lambda x: embed_val_map['cat_2'][x['cat_2']], axis=1, result_type='expand'),\n",
        "                        val_data.apply(lambda x: embed_val_map['cat_3'][x['cat_3']], axis=1, result_type='expand'),\n",
        "                        val_data.apply(lambda x: embed_val_map['item_condition_id'][x['item_condition_id']], axis=1, result_type='expand')),\n",
        "                        axis=1)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lq_0dJMSltQn",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "test_data = pd.concat((test_data, \n",
        "                        test_data.apply(lambda x: embed_val_map['cat_1'][x['cat_1']], axis=1, result_type='expand'), \n",
        "                        test_data.apply(lambda x: embed_val_map['cat_2'][x['cat_2']], axis=1, result_type='expand'),\n",
        "                        test_data.apply(lambda x: embed_val_map['cat_3'][x['cat_3']], axis=1, result_type='expand'),\n",
        "                        test_data.apply(lambda x: embed_val_map['item_condition_id'][x['item_condition_id']], axis=1, result_type='expand')),\n",
        "                        axis=1)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "M6LTYWm0WMT9",
        "colab_type": "text"
      },
      "source": [
        "Now, the embedding vectors have been added to the dataset. We rename some of the duplicate columns that were formed in the previous steps."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "oaTE3GvBhFTI",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import xgboost as xgb"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "V-f5YK0z16Mc",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "cols = pd.io.parsers.ParserBase({'names':train_data.columns})._maybe_dedup_names(train_data.columns)\n",
        "train_data.columns = cols\n",
        "val_data.columns = cols\n",
        "test_data.columns = cols\n",
        "train_data.drop(labels=['train_id', 'cat_1', 'cat_2', 'cat_3', 'item_condition_id', 'category_name'], axis=1, inplace=True)\n",
        "val_data.drop(labels=['train_id', 'cat_1', 'cat_2', 'cat_3', 'item_condition_id', 'category_name'], axis=1, inplace=True)\n",
        "test_data.drop(labels=['train_id', 'cat_1', 'cat_2', 'cat_3', 'item_condition_id', 'category_name'], axis=1, inplace=True)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qhc-JD1x4djK",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "train_data.to_pickle(input_folder + 'train_xgboost.pkl')\n",
        "val_data.to_pickle(input_folder + 'val_xgboost.pkl')\n",
        "test_data.to_pickle(input_folder + 'test_xgboost.pkl')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "oBR-iuYb2a54",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Selection of features\n",
        "xg_train_ft = train_data.columns.to_list()\n",
        "xg_train_ft.remove('price')\n",
        "\n",
        "xg_target = 'price'"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rLdiXnJ5tJsg",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "X_train, X_val, X_test, y_train, y_val, y_test = train_data[xg_train_ft], val_data[xg_train_ft], test_data[xg_train_ft], train_data[xg_target], val_data[xg_target], test_data[xg_target]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bFo9U1UhBhPQ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Redefine RMSLE function in NumPy\n",
        "def rmsle(y_true, y_pred):\n",
        "    return np.sqrt(np.mean(np.square(np.log1p(y_pred) - np.log1p(y_true))))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sHFYk5MeyqCs",
        "colab_type": "code",
        "outputId": "2ee76372-35b4-4c0d-d05b-19463f2a14ca",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 258
        }
      },
      "source": [
        "# Train the XgBoost Regressor using GPU\n",
        "regressor = xgb.XGBRegressor(**{\"updater\": \"grow_gpu_hist\"})\n",
        "regressor.fit(X_train, y_train)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/xgboost/core.py:587: FutureWarning: Series.base is deprecated and will be removed in a future version\n",
            "  if getattr(data, 'base', None) is not None and \\\n",
            "/usr/local/lib/python3.6/dist-packages/xgboost/core.py:588: FutureWarning: Series.base is deprecated and will be removed in a future version\n",
            "  data.base is not None and isinstance(data, np.ndarray) \\\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "[14:05:51] WARNING: /workspace/src/learner.cc:194: DANGER AHEAD: You have manually specified `updater` parameter. The `tree_method` parameter will be ignored. Incorrect sequence of updaters will produce undefined behavior. For common uses, we recommend using `tree_method` parameter instead.\n",
            "[14:05:51] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "XGBRegressor(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
              "             colsample_bynode=1, colsample_bytree=1, gamma=0,\n",
              "             importance_type='gain', learning_rate=0.1, max_delta_step=0,\n",
              "             max_depth=3, min_child_weight=1, missing=None, n_estimators=100,\n",
              "             n_jobs=1, nthread=None, objective='reg:linear', random_state=0,\n",
              "             reg_alpha=0, reg_lambda=1, scale_pos_weight=1, seed=None,\n",
              "             silent=None, subsample=1, updater='grow_gpu_hist', verbosity=1)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 52
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "moWZKpFN3BKS",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "y_pred = regressor.predict(X_test)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IthOAjTtA6_V",
        "colab_type": "code",
        "outputId": "60ce6834-355b-4ab5-e5f2-b16d743c8c90",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "print('The RMSL of XgBoost on test set is:', rmsle(y_test.values, y_pred))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "The RMSL of XgBoost on test set is: 0.4303965569919772\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "MtipOCc3dtOS",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}