Trip_Advisor_Hotel_Reviews.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Trip_Advisor_Hotel_Reviews.ipynb",
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "dZPP2CXCf2hh"
      },
      "source": [
        "# Submission Dicoding NLP"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hJ-LTwvtgGOn"
      },
      "source": [
        "## About this dataset\n",
        "\n",
        "Hotels play a crucial role in traveling and with the increased access to information new pathways of selecting the best ones emerged.\n",
        "With this dataset, consisting of 20k reviews crawled from Tripadvisor, you can explore what makes a great hotel and maybe even use this model in your travels!\n",
        "\n",
        "Feel free to [download dataset]( https://www.kaggle.com/andrewmvd/trip-advisor-hotel-reviews) from Kaggle\n",
        "\n",
        "### Citation\n",
        "\n",
        "Alam, M. H., Ryu, W.-J., Lee, S., 2016. Joint multi-grain topic sentiment: modeling semantic aspects for online reviews. Information Sciences 339, 206–223."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "C1s4rPB5f-XV"
      },
      "source": [
        "# Get the dataset\n",
        "\n",
        "We'll use the opendatasets library to download the data from Kaggle directly within Jupyter. Let's install and import opendatasets."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DzHcc5ihfs02"
      },
      "source": [
        "!pip install opendatasets --upgrade --quiet"
      ],
      "execution_count": 1,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VNb5txSyjlAr"
      },
      "source": [
        "import opendatasets as od"
      ],
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "id": "sMiBADE2jsXs",
        "outputId": "c7ad866e-3d3f-4c69-91ed-4bd61e701414"
      },
      "source": [
        "od.version()"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            },
            "text/plain": [
              "'0.1.20'"
            ]
          },
          "metadata": {},
          "execution_count": 3
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OWh6oNiUkQbI"
      },
      "source": [
        "The dataset can now be downloaded using od.download. When you execute od.download, you will be asked to provide your Kaggle username and API key. Follow these instructions to create an API key: http://bit.ly/kaggle-creds"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Cob4t8o3juAP"
      },
      "source": [
        "dataset_url = 'https://www.kaggle.com/andrewmvd/trip-advisor-hotel-reviews'"
      ],
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "nxlOr7NwkFTE",
        "outputId": "046cdce7-6b3b-481b-ac7b-a3513b9c0a7e"
      },
      "source": [
        "od.download(dataset_url)"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Please provide your Kaggle credentials to download this dataset. Learn more: http://bit.ly/kaggle-creds\n",
            "Your Kaggle username: salmanalshahab\n",
            "Your Kaggle Key: ··········\n",
            "Downloading trip-advisor-hotel-reviews.zip to ./trip-advisor-hotel-reviews\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100%|██████████| 5.14M/5.14M [00:00<00:00, 187MB/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ii0GuNHekp8p"
      },
      "source": [
        "Once the above command is executed, the dataset is downloaded and extracted to the directory trip-advisor-hotel-reviews"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Bwj7g4GKkIpb"
      },
      "source": [
        "import os"
      ],
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_Y7MTdtXlMP9"
      },
      "source": [
        "data_dir = './trip-advisor-hotel-reviews'"
      ],
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "FLcLAaOilYUQ",
        "outputId": "98d14a97-313a-4f7d-e869-0a43995386fb"
      },
      "source": [
        "os.listdir(data_dir)"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['tripadvisor_hotel_reviews.csv']"
            ]
          },
          "metadata": {},
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kJ8wuAy7lcQS"
      },
      "source": [
        "train_csv = data_dir + '/tripadvisor_hotel_reviews.csv'"
      ],
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "iqotX_QRmGDX"
      },
      "source": [
        "Let's load the data from *tripadvisor_hotel_reviews.csv* using Pandas."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 419
        },
        "id": "iWLIuqc7l-Og",
        "outputId": "c143458d-366a-4640-a559-64731f8ef06c"
      },
      "source": [
        "import pandas as pd\n",
        "\n",
        "dataset = pd.read_csv(train_csv)\n",
        "\n",
        "dataset"
      ],
      "execution_count": 10,
      "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>Review</th>\n",
              "      <th>Rating</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>nice hotel expensive parking got good deal sta...</td>\n",
              "      <td>4</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>ok nothing special charge diamond member hilto...</td>\n",
              "      <td>2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>nice rooms not 4* experience hotel monaco seat...</td>\n",
              "      <td>3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>unique, great stay, wonderful time hotel monac...</td>\n",
              "      <td>5</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>great stay great stay, went seahawk game aweso...</td>\n",
              "      <td>5</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>20486</th>\n",
              "      <td>best kept secret 3rd time staying charm, not 5...</td>\n",
              "      <td>5</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>20487</th>\n",
              "      <td>great location price view hotel great quick pl...</td>\n",
              "      <td>4</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>20488</th>\n",
              "      <td>ok just looks nice modern outside, desk staff ...</td>\n",
              "      <td>2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>20489</th>\n",
              "      <td>hotel theft ruined vacation hotel opened sept ...</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>20490</th>\n",
              "      <td>people talking, ca n't believe excellent ratin...</td>\n",
              "      <td>2</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>20491 rows × 2 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "                                                  Review  Rating\n",
              "0      nice hotel expensive parking got good deal sta...       4\n",
              "1      ok nothing special charge diamond member hilto...       2\n",
              "2      nice rooms not 4* experience hotel monaco seat...       3\n",
              "3      unique, great stay, wonderful time hotel monac...       5\n",
              "4      great stay great stay, went seahawk game aweso...       5\n",
              "...                                                  ...     ...\n",
              "20486  best kept secret 3rd time staying charm, not 5...       5\n",
              "20487  great location price view hotel great quick pl...       4\n",
              "20488  ok just looks nice modern outside, desk staff ...       2\n",
              "20489  hotel theft ruined vacation hotel opened sept ...       1\n",
              "20490  people talking, ca n't believe excellent ratin...       2\n",
              "\n",
              "[20491 rows x 2 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XnBNLZGgm0Bs"
      },
      "source": [
        "The dataset contains over 20,491 rows and 2 columns. Our Objectives is to create model to predict the value in the colums Rating."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "yEHqz459np6K"
      },
      "source": [
        "Let's check the data types and missing values in the various columns."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "aN361bC4mY2F",
        "outputId": "b622d8cd-d799-484b-c1b5-9070ac777e50"
      },
      "source": [
        "dataset.info()"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "RangeIndex: 20491 entries, 0 to 20490\n",
            "Data columns (total 2 columns):\n",
            " #   Column  Non-Null Count  Dtype \n",
            "---  ------  --------------  ----- \n",
            " 0   Review  20491 non-null  object\n",
            " 1   Rating  20491 non-null  int64 \n",
            "dtypes: int64(1), object(1)\n",
            "memory usage: 320.3+ KB\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hV0CSere5xXN"
      },
      "source": [
        "# Plot Hotel Review Ratings"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "EaAfoeqP6AxS",
        "outputId": "ab9f05d5-2cb8-447f-9145-9f4ef7258db0"
      },
      "source": [
        "dataset.Rating.value_counts()"
      ],
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "5    9054\n",
              "4    6039\n",
              "3    2184\n",
              "2    1793\n",
              "1    1421\n",
              "Name: Rating, dtype: int64"
            ]
          },
          "metadata": {},
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 475
        },
        "id": "qJJbqwIn6Eyk",
        "outputId": "7da1e2e4-00d7-47ff-aec0-b9032d5cf388"
      },
      "source": [
        "import seaborn as sns\n",
        "import matplotlib.pyplot as plt\n",
        "plt.figure(figsize=(8,6))\n",
        "sns.countplot(dataset[\"Rating\"])\n",
        "plt.title(\"Ratings of Hotels\")"
      ],
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/seaborn/_decorators.py:43: FutureWarning: Pass the following variable as a keyword arg: x. From version 0.12, the only valid positional argument will be `data`, and passing other arguments without an explicit keyword will result in an error or misinterpretation.\n",
            "  FutureWarning\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0.5, 1.0, 'Ratings of Hotels')"
            ]
          },
          "metadata": {},
          "execution_count": 13
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x432 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "N7cgIGxoP9qs"
      },
      "source": [
        "# One Hot Encoding Scores Column\n",
        "\n",
        "For multi-class classifcation using the neural network model, I needed the target variable to be one hot encoded, as the output layer would have 5 nodes, 1 for each score."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        },
        "id": "_bW3FIXHPtDA",
        "outputId": "a4ac7eeb-2304-4c5c-a450-7d069c233a8e"
      },
      "source": [
        "from sklearn.preprocessing import OneHotEncoder\n",
        "import pandas as pd\n",
        "enc = OneHotEncoder(handle_unknown='ignore')\n",
        "\n",
        "enc_df = pd.DataFrame(enc.fit_transform(dataset[['Rating']]).toarray())\n",
        "\n",
        "dataset = dataset.join(enc_df)\n",
        "dataset.head()"
      ],
      "execution_count": 14,
      "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>Review</th>\n",
              "      <th>Rating</th>\n",
              "      <th>0</th>\n",
              "      <th>1</th>\n",
              "      <th>2</th>\n",
              "      <th>3</th>\n",
              "      <th>4</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>nice hotel expensive parking got good deal sta...</td>\n",
              "      <td>4</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>ok nothing special charge diamond member hilto...</td>\n",
              "      <td>2</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>nice rooms not 4* experience hotel monaco seat...</td>\n",
              "      <td>3</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>unique, great stay, wonderful time hotel monac...</td>\n",
              "      <td>5</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>great stay great stay, went seahawk game aweso...</td>\n",
              "      <td>5</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                                              Review  Rating  ...    3    4\n",
              "0  nice hotel expensive parking got good deal sta...       4  ...  1.0  0.0\n",
              "1  ok nothing special charge diamond member hilto...       2  ...  0.0  0.0\n",
              "2  nice rooms not 4* experience hotel monaco seat...       3  ...  0.0  0.0\n",
              "3  unique, great stay, wonderful time hotel monac...       5  ...  0.0  1.0\n",
              "4  great stay great stay, went seahawk game aweso...       5  ...  0.0  1.0\n",
              "\n",
              "[5 rows x 7 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        },
        "id": "RAGdEpnARkzd",
        "outputId": "61a543c5-dc28-46b1-898e-0a844ef9ef21"
      },
      "source": [
        "dataset_new = dataset.drop(columns='Rating')\n",
        "dataset_new.head()"
      ],
      "execution_count": 15,
      "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>Review</th>\n",
              "      <th>0</th>\n",
              "      <th>1</th>\n",
              "      <th>2</th>\n",
              "      <th>3</th>\n",
              "      <th>4</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>nice hotel expensive parking got good deal sta...</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>ok nothing special charge diamond member hilto...</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>nice rooms not 4* experience hotel monaco seat...</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>unique, great stay, wonderful time hotel monac...</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>great stay great stay, went seahawk game aweso...</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                                              Review    0    1    2    3    4\n",
              "0  nice hotel expensive parking got good deal sta...  0.0  0.0  0.0  1.0  0.0\n",
              "1  ok nothing special charge diamond member hilto...  0.0  1.0  0.0  0.0  0.0\n",
              "2  nice rooms not 4* experience hotel monaco seat...  0.0  0.0  1.0  0.0  0.0\n",
              "3  unique, great stay, wonderful time hotel monac...  0.0  0.0  0.0  0.0  1.0\n",
              "4  great stay great stay, went seahawk game aweso...  0.0  0.0  0.0  0.0  1.0"
            ]
          },
          "metadata": {},
          "execution_count": 15
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Tyb-5kCSR-G1"
      },
      "source": [
        "dataset_new = dataset_new.rename(columns={0: \"Rating_1\", 1: \"Rating_2\", 2: \"Rating_3\", 3: \"Rating_4\", 4: \"Rating_5\"})"
      ],
      "execution_count": 16,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        },
        "id": "RAqEWtLCSiLi",
        "outputId": "08313a54-7a69-4a6c-b025-ae6eddffdca4"
      },
      "source": [
        "dataset_new.head()"
      ],
      "execution_count": 17,
      "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>Review</th>\n",
              "      <th>Rating_1</th>\n",
              "      <th>Rating_2</th>\n",
              "      <th>Rating_3</th>\n",
              "      <th>Rating_4</th>\n",
              "      <th>Rating_5</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>nice hotel expensive parking got good deal sta...</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>ok nothing special charge diamond member hilto...</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>nice rooms not 4* experience hotel monaco seat...</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>unique, great stay, wonderful time hotel monac...</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>great stay great stay, went seahawk game aweso...</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                                              Review  ...  Rating_5\n",
              "0  nice hotel expensive parking got good deal sta...  ...       0.0\n",
              "1  ok nothing special charge diamond member hilto...  ...       0.0\n",
              "2  nice rooms not 4* experience hotel monaco seat...  ...       0.0\n",
              "3  unique, great stay, wonderful time hotel monac...  ...       1.0\n",
              "4  great stay great stay, went seahawk game aweso...  ...       1.0\n",
              "\n",
              "[5 rows x 6 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 17
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_0JcYth2UG2y"
      },
      "source": [
        "Cleaning Text Data"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3SFyDtY1UDMg"
      },
      "source": [
        "# function for cleaning Review\n",
        "def standardize_text(df, field):\n",
        "    df[field] = df[field].str.replace(r\"http\\S+\", \"\")\n",
        "    df[field] = df[field].str.replace(r\"http\",\"\")\n",
        "    df[field] = df[field].str.replace(r\"@/S+\",\"\")\n",
        "    df[field] = df[field].str.replace(r\"[^A-Za-z0-9(),!?@\\'\\`\\\"\\_\\n]\", \" \")\n",
        "    df[field] = df[field].str.replace(r\"@\",\" at \")\n",
        "    df[field] = df[field].str.lower()\n",
        "    return df"
      ],
      "execution_count": 74,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        },
        "id": "ojpGo90vUJLp",
        "outputId": "acfef218-0635-47fd-a0bb-b022c8b5b66e"
      },
      "source": [
        "dataset_new = standardize_text(dataset_new, \"Review\")\n",
        "dataset_new.head()"
      ],
      "execution_count": 75,
      "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>Review</th>\n",
              "      <th>Rating_1</th>\n",
              "      <th>Rating_2</th>\n",
              "      <th>Rating_3</th>\n",
              "      <th>Rating_4</th>\n",
              "      <th>Rating_5</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>nice hotel expensive parking got good deal sta...</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>ok nothing special charge diamond member hilto...</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>nice rooms not 4  experience hotel monaco seat...</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>unique, great stay, wonderful time hotel monac...</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>great stay great stay, went seahawk game aweso...</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                                              Review  ...  Rating_5\n",
              "0  nice hotel expensive parking got good deal sta...  ...       0.0\n",
              "1  ok nothing special charge diamond member hilto...  ...       0.0\n",
              "2  nice rooms not 4  experience hotel monaco seat...  ...       0.0\n",
              "3  unique, great stay, wonderful time hotel monac...  ...       1.0\n",
              "4  great stay great stay, went seahawk game aweso...  ...       1.0\n",
              "\n",
              "[5 rows x 6 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 75
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6pzcbwDavde9"
      },
      "source": [
        "Get Training and Testing dataset with split data become 80 percent and 20 percent respectively"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yzQS-SlITX3G"
      },
      "source": [
        "reviews = dataset_new['Review']\n",
        "ratings = dataset_new[['Rating_1', 'Rating_2', 'Rating_3', 'Rating_4', 'Rating_5']]"
      ],
      "execution_count": 76,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XZiHdggxTD6K"
      },
      "source": [
        "from sklearn.model_selection import train_test_split\n",
        "\n",
        "training_reviews, testing_reviews, training_ratings, testing_ratings = train_test_split(reviews, ratings, test_size=0.2)\n"
      ],
      "execution_count": 77,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wdCtk0tVVUxa"
      },
      "source": [
        "# Tokenizing text and creating sequences for sentences"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PJOw5_AJuRuQ"
      },
      "source": [
        "Now you will tokenize text and create sequences for sentences as the first stage of preparing text for use with TensorFlow models."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FGa0dzs0nj6G"
      },
      "source": [
        "# Import Tokenizer and pad_sequences\n",
        "import tensorflow as tf\n",
        "from tensorflow.keras.preprocessing.text import Tokenizer\n",
        "from tensorflow.keras.preprocessing.sequence import pad_sequences"
      ],
      "execution_count": 85,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "plnVm__x0Hw8"
      },
      "source": [
        "# Define variables first to make easier tokenization\n",
        "vocab_size = 3000\n",
        "max_length = 250\n",
        "trunc_type = 'post'\n",
        "padding_type = 'post'\n",
        "oov_took = \"<OOV>\"\n",
        "\n",
        "tokenizer = Tokenizer(num_words = vocab_size, oov_token = oov_took, filters='!\"#$%&()*+,-./:;<=>?@[\\\\]^_`{|}~\\t\\n',\n",
        "                                                                   lower=True, split=' ', char_level=False,)\n",
        "\n",
        "\n",
        "tokenizer.fit_on_texts(training_reviews)\n",
        "#tokenizer.fit_on_texts(testing_reviews)\n",
        "\n",
        "# Sequences will be padded to the length of the longest individual sequence.\n",
        "training_sequences = tokenizer.texts_to_sequences(training_reviews)\n",
        "testing_sequences = tokenizer.texts_to_sequences(testing_reviews)\n",
        "\n",
        "\n",
        "training_padded = pad_sequences(training_sequences, maxlen=max_length, padding=padding_type,\n",
        "                       truncating=trunc_type)\n",
        "testing_padded = pad_sequences(testing_sequences, maxlen=max_length,\n",
        "                               padding=padding_type, truncating=trunc_type)"
      ],
      "execution_count": 89,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "h_hjfT7oXQcv"
      },
      "source": [
        "Head [here](https://www.tensorflow.org/api_docs/python/tf/keras/preprocessing/text/Tokenizer) if you’d like to check out the full TensorFlow documentation for *Tokenizer*.\n",
        "\n",
        "Head [here](https://www.tensorflow.org/api_docs/python/tf/keras/preprocessing/sequence/pad_sequences) if you’d like to check out the full TensorFlow documentation for *pad_sequences*.\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "c41wX6c1jktQ"
      },
      "source": [
        "## Train a Model with Embeddings and Bidirectional LSTM"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_g0V9913jx_S"
      },
      "source": [
        "# Build a basic reviews network\n",
        "# Note the embedding layer is first\n",
        "\n",
        "# Define embedding dimension as variable\n",
        "embedding_dim = 16 \n",
        "\n",
        "model = tf.keras.Sequential([\n",
        "                             tf.keras.layers.Embedding(vocab_size, embedding_dim, input_length=max_length),\n",
        "                             tf.keras.layers.Bidirectional(tf.keras.layers.LSTM(512, \n",
        "                                                       return_sequences=True)), \n",
        "                             tf.keras.layers.Bidirectional(tf.keras.layers.LSTM(256)), \n",
        "                             tf.keras.layers.Dense(512, kernel_regularizer=tf.keras.regularizers.l2(0.001),\n",
        "                                                   activation='relu'),\n",
        "                             tf.keras.layers.Dropout(0.5),\n",
        "                             tf.keras.layers.Dense(256, kernel_regularizer=tf.keras.regularizers.l2(0.001),\n",
        "                                                   activation='relu'),\n",
        "                             tf.keras.layers.Dropout(0.5),\n",
        "                             tf.keras.layers.Dense(5, activation='softmax')\n",
        "])"
      ],
      "execution_count": 90,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1uGWaHEmc4Ys"
      },
      "source": [
        "If you want to learn more on why word embeddings are used in NLP, check out this [useful post](https://towardsdatascience.com/why-do-we-use-embeddings-in-nlp-2f20e1b632d2).\n",
        "\n",
        "\n",
        "\n",
        "Tensorflow Documentation:\n",
        "\n",
        "\n",
        "[1] [Word Embedding](https://www.tensorflow.org/text/guide/word_embeddings)\n",
        "\n",
        "\n",
        "\n",
        "[2] [Embedding](https://www.tensorflow.org/api_docs/python/tf/keras/layers/Embedding)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "E8DoQlvBmc9w"
      },
      "source": [
        "# Compile function\n",
        "model.compile(loss='categorical_crossentropy',\n",
        "              optimizer='adam',\n",
        "              metrics=['accuracy'])"
      ],
      "execution_count": 91,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZEljOy48m44H",
        "outputId": "0d70055d-36d3-4208-dfbf-1022aaf68054"
      },
      "source": [
        "model.summary()"
      ],
      "execution_count": 92,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential_12\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "embedding_12 (Embedding)     (None, 250, 16)           48000     \n",
            "_________________________________________________________________\n",
            "bidirectional_24 (Bidirectio (None, 250, 1024)         2166784   \n",
            "_________________________________________________________________\n",
            "bidirectional_25 (Bidirectio (None, 512)               2623488   \n",
            "_________________________________________________________________\n",
            "dense_35 (Dense)             (None, 512)               262656    \n",
            "_________________________________________________________________\n",
            "dropout_21 (Dropout)         (None, 512)               0         \n",
            "_________________________________________________________________\n",
            "dense_36 (Dense)             (None, 256)               131328    \n",
            "_________________________________________________________________\n",
            "dropout_22 (Dropout)         (None, 256)               0         \n",
            "_________________________________________________________________\n",
            "dense_37 (Dense)             (None, 5)                 1285      \n",
            "=================================================================\n",
            "Total params: 5,233,541\n",
            "Trainable params: 5,233,541\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "AXqSfodBZhCi"
      },
      "source": [
        "# callback\n",
        "callback = tf.keras.callbacks.EarlyStopping(monitor='loss', patience=3)"
      ],
      "execution_count": 93,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "cUhNmN_SnPaA",
        "outputId": "c5d56433-281e-4587-d111-4faa022506aa"
      },
      "source": [
        "num_epochs = 30\n",
        "history = model.fit(training_padded, training_ratings, epochs=num_epochs, batch_size=128, validation_data=(testing_padded, testing_ratings), callbacks=[callback], verbose=2)"
      ],
      "execution_count": 94,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/30\n",
            "129/129 - 117s - loss: 1.7028 - accuracy: 0.4300 - val_loss: 1.4337 - val_accuracy: 0.4274\n",
            "Epoch 2/30\n",
            "129/129 - 108s - loss: 1.2969 - accuracy: 0.4632 - val_loss: 1.2117 - val_accuracy: 0.4704\n",
            "Epoch 3/30\n",
            "129/129 - 108s - loss: 1.1474 - accuracy: 0.5009 - val_loss: 1.1436 - val_accuracy: 0.5013\n",
            "Epoch 4/30\n",
            "129/129 - 108s - loss: 1.0889 - accuracy: 0.5176 - val_loss: 1.1574 - val_accuracy: 0.4904\n",
            "Epoch 5/30\n",
            "129/129 - 108s - loss: 1.0579 - accuracy: 0.5313 - val_loss: 1.1195 - val_accuracy: 0.5055\n",
            "Epoch 6/30\n",
            "129/129 - 108s - loss: 1.0299 - accuracy: 0.5454 - val_loss: 1.1244 - val_accuracy: 0.5040\n",
            "Epoch 7/30\n",
            "129/129 - 108s - loss: 1.0156 - accuracy: 0.5569 - val_loss: 1.1233 - val_accuracy: 0.5006\n",
            "Epoch 8/30\n",
            "129/129 - 108s - loss: 1.0123 - accuracy: 0.5589 - val_loss: 1.1208 - val_accuracy: 0.5265\n",
            "Epoch 9/30\n",
            "129/129 - 108s - loss: 0.9987 - accuracy: 0.5673 - val_loss: 1.1157 - val_accuracy: 0.5299\n",
            "Epoch 10/30\n",
            "129/129 - 108s - loss: 0.9839 - accuracy: 0.5806 - val_loss: 1.1228 - val_accuracy: 0.5394\n",
            "Epoch 11/30\n",
            "129/129 - 108s - loss: 0.9850 - accuracy: 0.5813 - val_loss: 1.2004 - val_accuracy: 0.5128\n",
            "Epoch 12/30\n",
            "129/129 - 109s - loss: 0.9622 - accuracy: 0.5959 - val_loss: 1.1264 - val_accuracy: 0.5418\n",
            "Epoch 13/30\n",
            "129/129 - 108s - loss: 0.9560 - accuracy: 0.5999 - val_loss: 1.1170 - val_accuracy: 0.5396\n",
            "Epoch 14/30\n",
            "129/129 - 108s - loss: 0.9489 - accuracy: 0.6038 - val_loss: 1.1196 - val_accuracy: 0.5440\n",
            "Epoch 15/30\n",
            "129/129 - 109s - loss: 0.9340 - accuracy: 0.6119 - val_loss: 1.1186 - val_accuracy: 0.5389\n",
            "Epoch 16/30\n",
            "129/129 - 109s - loss: 0.9291 - accuracy: 0.6119 - val_loss: 1.1114 - val_accuracy: 0.5399\n",
            "Epoch 17/30\n",
            "129/129 - 108s - loss: 0.9249 - accuracy: 0.6191 - val_loss: 1.1212 - val_accuracy: 0.5477\n",
            "Epoch 18/30\n",
            "129/129 - 108s - loss: 0.9122 - accuracy: 0.6229 - val_loss: 1.0825 - val_accuracy: 0.5533\n",
            "Epoch 19/30\n",
            "129/129 - 108s - loss: 0.9085 - accuracy: 0.6223 - val_loss: 1.1007 - val_accuracy: 0.5448\n",
            "Epoch 20/30\n",
            "129/129 - 108s - loss: 0.9109 - accuracy: 0.6292 - val_loss: 1.1066 - val_accuracy: 0.5416\n",
            "Epoch 21/30\n",
            "129/129 - 108s - loss: 0.9214 - accuracy: 0.6203 - val_loss: 1.1837 - val_accuracy: 0.5387\n",
            "Epoch 22/30\n",
            "129/129 - 108s - loss: 0.9018 - accuracy: 0.6348 - val_loss: 1.1082 - val_accuracy: 0.5438\n",
            "Epoch 23/30\n",
            "129/129 - 108s - loss: 0.8806 - accuracy: 0.6428 - val_loss: 1.1670 - val_accuracy: 0.5667\n",
            "Epoch 24/30\n",
            "129/129 - 108s - loss: 0.8766 - accuracy: 0.6445 - val_loss: 1.1111 - val_accuracy: 0.5592\n",
            "Epoch 25/30\n",
            "129/129 - 108s - loss: 0.8897 - accuracy: 0.6318 - val_loss: 1.0979 - val_accuracy: 0.5567\n",
            "Epoch 26/30\n",
            "129/129 - 108s - loss: 0.8694 - accuracy: 0.6495 - val_loss: 1.1306 - val_accuracy: 0.5631\n",
            "Epoch 27/30\n",
            "129/129 - 108s - loss: 0.8568 - accuracy: 0.6508 - val_loss: 1.1437 - val_accuracy: 0.5575\n",
            "Epoch 28/30\n",
            "129/129 - 108s - loss: 0.8946 - accuracy: 0.6296 - val_loss: 1.1179 - val_accuracy: 0.5570\n",
            "Epoch 29/30\n",
            "129/129 - 108s - loss: 0.8704 - accuracy: 0.6517 - val_loss: 1.1110 - val_accuracy: 0.5614\n",
            "Epoch 30/30\n",
            "129/129 - 108s - loss: 0.8566 - accuracy: 0.6528 - val_loss: 1.0585 - val_accuracy: 0.5748\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "brBcbxXcmfYO"
      },
      "source": [
        "# Plot the accuracy and loss"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "h9E7G_KGTvJI",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 541
        },
        "outputId": "99706dbe-6884-4ea0-83a9-05d8eafde94a"
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "\n",
        "\n",
        "def plot_graphs(history, string):\n",
        "  plt.plot(history.history[string])\n",
        "  plt.plot(history.history['val_'+string])\n",
        "  plt.xlabel(\"Epochs\")\n",
        "  plt.ylabel(string)\n",
        "  plt.legend([string, 'val_'+string])\n",
        "  plt.show()\n",
        "\n",
        "plot_graphs(history, \"accuracy\")\n",
        "plot_graphs(history, \"loss\")\n"
      ],
      "execution_count": 95,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fsj2SBdqU7f-"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}