Feedback_Model.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/mjain2/csci544-group32/blob/main/LSTM_Model.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Y0bjBmxduGt7"
      },
      "outputs": [],
      "source": [
        "import re \n",
        "import warnings\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import seaborn as sns\n",
        "import tensorflow as tf\n",
        "from wordcloud import WordCloud\n",
        "from matplotlib import pyplot as plt\n",
        "from sklearn.feature_extraction.text import CountVectorizer\n",
        "from tensorflow.keras.preprocessing.text import Tokenizer\n",
        "from tensorflow.keras.preprocessing.sequence import pad_sequences\n",
        "from sklearn.metrics import mean_squared_error\n",
        "warnings.filterwarnings(\"ignore\", category=FutureWarning)\n",
        "sns.set_theme()\n",
        "sns.set(rc = {'figure.figsize':(15,8)})\n",
        "%matplotlib inline "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "XDI5Tz9JRcbU",
        "outputId": "39778ab1-db2f-4f9d-badd-57a5d0ff3166"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/\n",
            "Collecting language_tool_python\n",
            "  Downloading language_tool_python-2.7.1-py3-none-any.whl (34 kB)\n",
            "Requirement already satisfied: tqdm in /usr/local/lib/python3.7/dist-packages (from language_tool_python) (4.64.1)\n",
            "Requirement already satisfied: requests in /usr/local/lib/python3.7/dist-packages (from language_tool_python) (2.23.0)\n",
            "Requirement already satisfied: urllib3!=1.25.0,!=1.25.1,<1.26,>=1.21.1 in /usr/local/lib/python3.7/dist-packages (from requests->language_tool_python) (1.24.3)\n",
            "Requirement already satisfied: chardet<4,>=3.0.2 in /usr/local/lib/python3.7/dist-packages (from requests->language_tool_python) (3.0.4)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.7/dist-packages (from requests->language_tool_python) (2022.9.24)\n",
            "Requirement already satisfied: idna<3,>=2.5 in /usr/local/lib/python3.7/dist-packages (from requests->language_tool_python) (2.10)\n",
            "Installing collected packages: language-tool-python\n",
            "Successfully installed language-tool-python-2.7.1\n"
          ]
        }
      ],
      "source": [
        "!pip install --upgrade language_tool_python\n",
        "import language_tool_python"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "7ijtOKsSuZDF",
        "outputId": "6d9ee517-e69e-47d6-b846-1df1258e6442"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mounted at /content/gdrive\n"
          ]
        }
      ],
      "source": [
        "from google.colab import drive\n",
        "drive.mount(\"/content/gdrive\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "u-CsgYn7uLWW"
      },
      "outputs": [],
      "source": [
        "train_df = pd.read_csv('/content/gdrive/MyDrive/Colab Notebooks/NLP/Project/feedback-prize-english-language-learning/train.csv')\n",
        "test_df = pd.read_csv('/content/gdrive/MyDrive/Colab Notebooks/NLP/Project/feedback-prize-english-language-learning/test.csv')"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ObBxhq2puhi1"
      },
      "source": [
        "### Words Count Analysis"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "fQvv7izHuVO8",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "4f25fb8b-24ce-4f31-de80-9cb3b9fddf0d"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Average length: 430\n",
            "Max length: 1260\n"
          ]
        }
      ],
      "source": [
        "train_df['num_words'] = train_df['full_text'].apply(lambda x: len(x.split()))\n",
        "avg_words = round(train_df['num_words'].mean())\n",
        "max_words = round(train_df['num_words'].max())\n",
        "print('Average length: {}'.format(avg_words))\n",
        "print('Max length: {}'.format(max_words))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 520
        },
        "id": "SRqdiTO_ulCo",
        "outputId": "f5057e84-5f29-4da4-af06-761f28be5af5"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7f1f04bcfdd0>"
            ]
          },
          "metadata": {},
          "execution_count": 6
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1080x576 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "sns.set(rc = {'figure.figsize':(15,8)})\n",
        "sns.distplot(train_df['num_words'], kde=True)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "C_eN6k0-u1pA"
      },
      "source": [
        "### Dataset Preparation"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "pY6VJ7l7uop3"
      },
      "outputs": [],
      "source": [
        "tokenizer = Tokenizer(oov_token=\"<OOV>\")\n",
        "tokenizer.fit_on_texts(train_df['full_text'])\n",
        "word_index = tokenizer.word_index\n",
        "\n",
        "train_seq = tokenizer.texts_to_sequences(train_df['full_text'])\n",
        "pad_train = pad_sequences(train_seq, maxlen=max_words, truncating='post')\n",
        "\n",
        "test_seq = tokenizer.texts_to_sequences(test_df['full_text'])\n",
        "pad_test = pad_sequences(test_seq, maxlen=max_words, truncating='post')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "SB1GsJiJuyC-",
        "outputId": "6de8337b-dd7b-41d0-fe5e-423fe4daea22"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "22170\n"
          ]
        }
      ],
      "source": [
        "word_idx_count = len(word_index)\n",
        "print(word_idx_count)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "SjHn-B7LvBeB"
      },
      "outputs": [],
      "source": [
        "X = pad_train\n",
        "y = [train_df['cohesion'].values, train_df['syntax'].values, train_df['vocabulary'].values,\n",
        "      train_df['phraseology'].values, train_df['grammar'].values, train_df['conventions'].values]"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "AY9cPJrlvJAo"
      },
      "source": [
        "### LSTM"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Ifq-C-r_vDku"
      },
      "outputs": [],
      "source": [
        "def get_lstm_model():\n",
        "\n",
        "    inputs = tf.keras.layers.Input(shape=(max_words))\n",
        "    embeddings = tf.keras.layers.Embedding(word_idx_count + 1, 512, input_length=max_words)(inputs)\n",
        "    x = tf.keras.layers.Bidirectional(tf.keras.layers.LSTM(512, return_sequences=True))(embeddings)\n",
        "    x = tf.keras.layers.Bidirectional(tf.keras.layers.LSTM(128, return_sequences=True))(x)\n",
        "    x = tf.keras.layers.Bidirectional(tf.keras.layers.LSTM(128, return_sequences=True))(x)\n",
        "    x = tf.keras.layers.Bidirectional(tf.keras.layers.LSTM(64))(x)\n",
        "    \n",
        "    output1 = tf.keras.layers.Dense(128, activation='relu', kernel_initializer=tf.keras.initializers.RandomNormal(mean=0.0, stddev=0.05), \n",
        "                                    bias_initializer='zeros',\n",
        "                                    kernel_regularizer=tf.keras.regularizers.L1L2(l1=1e-5, l2=1e-4), \n",
        "                                    bias_regularizer=tf.keras.regularizers.L2(1e-4), \n",
        "                                    activity_regularizer=tf.keras.regularizers.L2(1e-5))(x)\n",
        "    output1 = tf.keras.layers.Dropout(0.4)(output1)\n",
        "    output1 = tf.keras.layers.Dense(64, activation='relu', kernel_initializer=tf.keras.initializers.RandomNormal(mean=0.0, stddev=0.05), \n",
        "                                    bias_initializer='zeros',\n",
        "                                    kernel_regularizer=tf.keras.regularizers.L1L2(l1=1e-5, l2=1e-4), \n",
        "                                    bias_regularizer=tf.keras.regularizers.L2(1e-4), \n",
        "                                    activity_regularizer=tf.keras.regularizers.L2(1e-5))(output1)\n",
        "    output1 = tf.keras.layers.Dropout(0.4)(output1)\n",
        "    output1 = tf.keras.layers.Dense(1, activation='relu', name='cohesion')(output1)\n",
        "    \n",
        "    output2 = tf.keras.layers.Dense(128, activation='relu', kernel_initializer=tf.keras.initializers.RandomNormal(mean=0.0, stddev=0.05), \n",
        "                                    bias_initializer='zeros',\n",
        "                                    kernel_regularizer=tf.keras.regularizers.L1L2(l1=1e-5, l2=1e-4), \n",
        "                                    bias_regularizer=tf.keras.regularizers.L2(1e-4), \n",
        "                                    activity_regularizer=tf.keras.regularizers.L2(1e-5))(x)\n",
        "    output2 = tf.keras.layers.Dropout(0.4)(output2)\n",
        "    output2 = tf.keras.layers.Dense(64, activation='relu', kernel_initializer=tf.keras.initializers.RandomNormal(mean=0.0, stddev=0.05), \n",
        "                                    bias_initializer='zeros',\n",
        "                                    kernel_regularizer=tf.keras.regularizers.L1L2(l1=1e-5, l2=1e-4), \n",
        "                                    bias_regularizer=tf.keras.regularizers.L2(1e-4), \n",
        "                                    activity_regularizer=tf.keras.regularizers.L2(1e-5))(output2)\n",
        "    output2 = tf.keras.layers.Dropout(0.4)(output2)\n",
        "    output2 = tf.keras.layers.Dense(1, activation='relu', name='syntax')(output2)\n",
        "    \n",
        "    output3 = tf.keras.layers.Dense(128, activation='relu', kernel_initializer=tf.keras.initializers.RandomNormal(mean=0.0, stddev=0.05), \n",
        "                                    bias_initializer='zeros',\n",
        "                                    kernel_regularizer=tf.keras.regularizers.L1L2(l1=1e-5, l2=1e-4), \n",
        "                                    bias_regularizer=tf.keras.regularizers.L2(1e-4), \n",
        "                                    activity_regularizer=tf.keras.regularizers.L2(1e-5))(x)\n",
        "    output3 = tf.keras.layers.Dropout(0.4)(output3)\n",
        "    output3 = tf.keras.layers.Dense(64, activation='relu', kernel_initializer=tf.keras.initializers.RandomNormal(mean=0.0, stddev=0.05), \n",
        "                                    bias_initializer='zeros',\n",
        "                                    kernel_regularizer=tf.keras.regularizers.L1L2(l1=1e-5, l2=1e-4), \n",
        "                                    bias_regularizer=tf.keras.regularizers.L2(1e-4), \n",
        "                                    activity_regularizer=tf.keras.regularizers.L2(1e-5))(output3)\n",
        "    output3 = tf.keras.layers.Dropout(0.4)(output3)\n",
        "    output3 = tf.keras.layers.Dense(1, activation='relu', name='vocabulary')(output3)\n",
        "    \n",
        "    output4 = tf.keras.layers.Dense(128, activation='relu', kernel_initializer=tf.keras.initializers.RandomNormal(mean=0.0, stddev=0.05), \n",
        "                                    bias_initializer='zeros',\n",
        "                                    kernel_regularizer=tf.keras.regularizers.L1L2(l1=1e-5, l2=1e-4), \n",
        "                                    bias_regularizer=tf.keras.regularizers.L2(1e-4), \n",
        "                                    activity_regularizer=tf.keras.regularizers.L2(1e-5))(x)\n",
        "    output4 = tf.keras.layers.Dropout(0.4)(output4)\n",
        "    output4 = tf.keras.layers.Dense(64, activation='relu', kernel_initializer=tf.keras.initializers.RandomNormal(mean=0.0, stddev=0.05), \n",
        "                                    bias_initializer='zeros',\n",
        "                                    kernel_regularizer=tf.keras.regularizers.L1L2(l1=1e-5, l2=1e-4), \n",
        "                                    bias_regularizer=tf.keras.regularizers.L2(1e-4), \n",
        "                                    activity_regularizer=tf.keras.regularizers.L2(1e-5))(output4)\n",
        "    output4 = tf.keras.layers.Dropout(0.4)(output4)\n",
        "    output4 = tf.keras.layers.Dense(1, activation='relu', name='phraseology')(output4)\n",
        "    \n",
        "    output5 = tf.keras.layers.Dense(128, activation='relu', kernel_initializer=tf.keras.initializers.RandomNormal(mean=0.0, stddev=0.05), \n",
        "                                    bias_initializer='zeros',\n",
        "                                    kernel_regularizer=tf.keras.regularizers.L1L2(l1=1e-5, l2=1e-4), \n",
        "                                    bias_regularizer=tf.keras.regularizers.L2(1e-4), \n",
        "                                    activity_regularizer=tf.keras.regularizers.L2(1e-5))(x)\n",
        "    output5 = tf.keras.layers.Dropout(0.4)(output5)\n",
        "    output5 = tf.keras.layers.Dense(64, activation='relu', kernel_initializer=tf.keras.initializers.RandomNormal(mean=0.0, stddev=0.05), \n",
        "                                    bias_initializer='zeros',\n",
        "                                    kernel_regularizer=tf.keras.regularizers.L1L2(l1=1e-5, l2=1e-4), \n",
        "                                    bias_regularizer=tf.keras.regularizers.L2(1e-4), \n",
        "                                    activity_regularizer=tf.keras.regularizers.L2(1e-5))(output5)\n",
        "    output5 = tf.keras.layers.Dropout(0.4)(output5)\n",
        "    output5 = tf.keras.layers.Dense(1, activation='relu', name='grammar')(output5)  \n",
        "    \n",
        "    output6 = tf.keras.layers.Dense(128, activation='relu', kernel_initializer=tf.keras.initializers.RandomNormal(mean=0.0, stddev=0.05), \n",
        "                                    bias_initializer='zeros',\n",
        "                                    kernel_regularizer=tf.keras.regularizers.L1L2(l1=1e-5, l2=1e-4), \n",
        "                                    bias_regularizer=tf.keras.regularizers.L2(1e-4), \n",
        "                                    activity_regularizer=tf.keras.regularizers.L2(1e-5))(x)\n",
        "    output6 = tf.keras.layers.Dropout(0.4)(output6)\n",
        "    output6 = tf.keras.layers.Dense(64, activation='relu', kernel_initializer=tf.keras.initializers.RandomNormal(mean=0.0, stddev=0.05), \n",
        "                                    bias_initializer='zeros',\n",
        "                                    kernel_regularizer=tf.keras.regularizers.L1L2(l1=1e-5, l2=1e-4), \n",
        "                                    bias_regularizer=tf.keras.regularizers.L2(1e-4), \n",
        "                                    activity_regularizer=tf.keras.regularizers.L2(1e-5))(output6)\n",
        "    output6 = tf.keras.layers.Dropout(0.4)(output6)\n",
        "    output6 = tf.keras.layers.Dense(1, activation='relu', name='conventions')(output6)\n",
        "    \n",
        "    model = tf.keras.models.Model(inputs=inputs, outputs=[output1, output2, output3, output4, output5, output6])\n",
        "    \n",
        "    model.compile(\n",
        "        optimizer=tf.keras.optimizers.Adam(),\n",
        "        loss = tf.keras.losses.mean_squared_error\n",
        "    )\n",
        "    \n",
        "    return model"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Nnr6zHKGvLl3"
      },
      "outputs": [],
      "source": [
        "model = get_lstm_model()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-X9CzTNKvN_f",
        "outputId": "89775574-4cd6-4aa4-b553-b1a497908ee8"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            " Layer (type)                   Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            " input_1 (InputLayer)           [(None, 1260)]       0           []                               \n",
            "                                                                                                  \n",
            " embedding (Embedding)          (None, 1260, 512)    11351552    ['input_1[0][0]']                \n",
            "                                                                                                  \n",
            " bidirectional (Bidirectional)  (None, 1260, 1024)   4198400     ['embedding[0][0]']              \n",
            "                                                                                                  \n",
            " bidirectional_1 (Bidirectional  (None, 1260, 256)   1180672     ['bidirectional[0][0]']          \n",
            " )                                                                                                \n",
            "                                                                                                  \n",
            " bidirectional_2 (Bidirectional  (None, 1260, 256)   394240      ['bidirectional_1[0][0]']        \n",
            " )                                                                                                \n",
            "                                                                                                  \n",
            " bidirectional_3 (Bidirectional  (None, 128)         164352      ['bidirectional_2[0][0]']        \n",
            " )                                                                                                \n",
            "                                                                                                  \n",
            " dense (Dense)                  (None, 128)          16512       ['bidirectional_3[0][0]']        \n",
            "                                                                                                  \n",
            " dense_2 (Dense)                (None, 128)          16512       ['bidirectional_3[0][0]']        \n",
            "                                                                                                  \n",
            " dense_4 (Dense)                (None, 128)          16512       ['bidirectional_3[0][0]']        \n",
            "                                                                                                  \n",
            " dense_6 (Dense)                (None, 128)          16512       ['bidirectional_3[0][0]']        \n",
            "                                                                                                  \n",
            " dense_8 (Dense)                (None, 128)          16512       ['bidirectional_3[0][0]']        \n",
            "                                                                                                  \n",
            " dense_10 (Dense)               (None, 128)          16512       ['bidirectional_3[0][0]']        \n",
            "                                                                                                  \n",
            " dropout (Dropout)              (None, 128)          0           ['dense[0][0]']                  \n",
            "                                                                                                  \n",
            " dropout_2 (Dropout)            (None, 128)          0           ['dense_2[0][0]']                \n",
            "                                                                                                  \n",
            " dropout_4 (Dropout)            (None, 128)          0           ['dense_4[0][0]']                \n",
            "                                                                                                  \n",
            " dropout_6 (Dropout)            (None, 128)          0           ['dense_6[0][0]']                \n",
            "                                                                                                  \n",
            " dropout_8 (Dropout)            (None, 128)          0           ['dense_8[0][0]']                \n",
            "                                                                                                  \n",
            " dropout_10 (Dropout)           (None, 128)          0           ['dense_10[0][0]']               \n",
            "                                                                                                  \n",
            " dense_1 (Dense)                (None, 64)           8256        ['dropout[0][0]']                \n",
            "                                                                                                  \n",
            " dense_3 (Dense)                (None, 64)           8256        ['dropout_2[0][0]']              \n",
            "                                                                                                  \n",
            " dense_5 (Dense)                (None, 64)           8256        ['dropout_4[0][0]']              \n",
            "                                                                                                  \n",
            " dense_7 (Dense)                (None, 64)           8256        ['dropout_6[0][0]']              \n",
            "                                                                                                  \n",
            " dense_9 (Dense)                (None, 64)           8256        ['dropout_8[0][0]']              \n",
            "                                                                                                  \n",
            " dense_11 (Dense)               (None, 64)           8256        ['dropout_10[0][0]']             \n",
            "                                                                                                  \n",
            " dropout_1 (Dropout)            (None, 64)           0           ['dense_1[0][0]']                \n",
            "                                                                                                  \n",
            " dropout_3 (Dropout)            (None, 64)           0           ['dense_3[0][0]']                \n",
            "                                                                                                  \n",
            " dropout_5 (Dropout)            (None, 64)           0           ['dense_5[0][0]']                \n",
            "                                                                                                  \n",
            " dropout_7 (Dropout)            (None, 64)           0           ['dense_7[0][0]']                \n",
            "                                                                                                  \n",
            " dropout_9 (Dropout)            (None, 64)           0           ['dense_9[0][0]']                \n",
            "                                                                                                  \n",
            " dropout_11 (Dropout)           (None, 64)           0           ['dense_11[0][0]']               \n",
            "                                                                                                  \n",
            " cohesion (Dense)               (None, 1)            65          ['dropout_1[0][0]']              \n",
            "                                                                                                  \n",
            " syntax (Dense)                 (None, 1)            65          ['dropout_3[0][0]']              \n",
            "                                                                                                  \n",
            " vocabulary (Dense)             (None, 1)            65          ['dropout_5[0][0]']              \n",
            "                                                                                                  \n",
            " phraseology (Dense)            (None, 1)            65          ['dropout_7[0][0]']              \n",
            "                                                                                                  \n",
            " grammar (Dense)                (None, 1)            65          ['dropout_9[0][0]']              \n",
            "                                                                                                  \n",
            " conventions (Dense)            (None, 1)            65          ['dropout_11[0][0]']             \n",
            "                                                                                                  \n",
            "==================================================================================================\n",
            "Total params: 17,438,214\n",
            "Trainable params: 17,438,214\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "None\n"
          ]
        }
      ],
      "source": [
        "print(model.summary())"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZtNCmqupvQsw",
        "outputId": "53decb67-6f8f-4818-a4aa-ad8659de16e9"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch 1/10\n",
            "98/98 [==============================] - 52s 531ms/step - loss: 5.1831 - cohesion_loss: 0.8340 - syntax_loss: 0.8529 - vocabulary_loss: 0.8009 - phraseology_loss: 0.8518 - grammar_loss: 0.8985 - conventions_loss: 0.8707 - val_loss: 2.8820 - val_cohesion_loss: 0.4582 - val_syntax_loss: 0.4594 - val_vocabulary_loss: 0.3907 - val_phraseology_loss: 0.5041 - val_grammar_loss: 0.5098 - val_conventions_loss: 0.4884\n",
            "Epoch 2/10\n",
            "98/98 [==============================] - 49s 496ms/step - loss: 4.8118 - cohesion_loss: 0.7716 - syntax_loss: 0.7940 - vocabulary_loss: 0.7489 - phraseology_loss: 0.7941 - grammar_loss: 0.8381 - conventions_loss: 0.7945 - val_loss: 2.6835 - val_cohesion_loss: 0.4572 - val_syntax_loss: 0.4221 - val_vocabulary_loss: 0.3308 - val_phraseology_loss: 0.4301 - val_grammar_loss: 0.4963 - val_conventions_loss: 0.4775\n",
            "Epoch 3/10\n",
            "98/98 [==============================] - 48s 494ms/step - loss: 4.6656 - cohesion_loss: 0.7366 - syntax_loss: 0.7758 - vocabulary_loss: 0.7334 - phraseology_loss: 0.7729 - grammar_loss: 0.7926 - conventions_loss: 0.7864 - val_loss: 2.8733 - val_cohesion_loss: 0.5352 - val_syntax_loss: 0.4709 - val_vocabulary_loss: 0.3830 - val_phraseology_loss: 0.4153 - val_grammar_loss: 0.5064 - val_conventions_loss: 0.4960\n",
            "Epoch 4/10\n",
            "98/98 [==============================] - 48s 494ms/step - loss: 4.4603 - cohesion_loss: 0.6967 - syntax_loss: 0.7231 - vocabulary_loss: 0.6976 - phraseology_loss: 0.7538 - grammar_loss: 0.7798 - conventions_loss: 0.7436 - val_loss: 2.6129 - val_cohesion_loss: 0.4407 - val_syntax_loss: 0.4153 - val_vocabulary_loss: 0.3324 - val_phraseology_loss: 0.4254 - val_grammar_loss: 0.4943 - val_conventions_loss: 0.4399\n",
            "Epoch 5/10\n",
            "98/98 [==============================] - 49s 496ms/step - loss: 4.4600 - cohesion_loss: 0.7232 - syntax_loss: 0.7529 - vocabulary_loss: 0.7018 - phraseology_loss: 0.7219 - grammar_loss: 0.7673 - conventions_loss: 0.7289 - val_loss: 2.6372 - val_cohesion_loss: 0.4455 - val_syntax_loss: 0.4093 - val_vocabulary_loss: 0.3543 - val_phraseology_loss: 0.4179 - val_grammar_loss: 0.5037 - val_conventions_loss: 0.4436\n",
            "Epoch 6/10\n",
            "98/98 [==============================] - 48s 495ms/step - loss: 4.3663 - cohesion_loss: 0.6824 - syntax_loss: 0.7396 - vocabulary_loss: 0.6663 - phraseology_loss: 0.7225 - grammar_loss: 0.7761 - conventions_loss: 0.7170 - val_loss: 2.6172 - val_cohesion_loss: 0.4419 - val_syntax_loss: 0.4523 - val_vocabulary_loss: 0.3311 - val_phraseology_loss: 0.4070 - val_grammar_loss: 0.4970 - val_conventions_loss: 0.4262\n",
            "Epoch 7/10\n",
            "98/98 [==============================] - 49s 496ms/step - loss: 4.1718 - cohesion_loss: 0.6747 - syntax_loss: 0.7292 - vocabulary_loss: 0.6380 - phraseology_loss: 0.6639 - grammar_loss: 0.7067 - conventions_loss: 0.6984 - val_loss: 2.7070 - val_cohesion_loss: 0.4826 - val_syntax_loss: 0.4479 - val_vocabulary_loss: 0.3351 - val_phraseology_loss: 0.4248 - val_grammar_loss: 0.5203 - val_conventions_loss: 0.4358\n",
            "Epoch 8/10\n",
            "98/98 [==============================] - 49s 496ms/step - loss: 4.1143 - cohesion_loss: 0.6755 - syntax_loss: 0.6951 - vocabulary_loss: 0.6273 - phraseology_loss: 0.6663 - grammar_loss: 0.7117 - conventions_loss: 0.6787 - val_loss: 2.8899 - val_cohesion_loss: 0.4789 - val_syntax_loss: 0.4626 - val_vocabulary_loss: 0.3773 - val_phraseology_loss: 0.4454 - val_grammar_loss: 0.5542 - val_conventions_loss: 0.5127\n",
            "Epoch 9/10\n",
            "98/98 [==============================] - 49s 497ms/step - loss: 3.9019 - cohesion_loss: 0.6395 - syntax_loss: 0.6121 - vocabulary_loss: 0.5984 - phraseology_loss: 0.6449 - grammar_loss: 0.7085 - conventions_loss: 0.6403 - val_loss: 2.8616 - val_cohesion_loss: 0.5053 - val_syntax_loss: 0.4513 - val_vocabulary_loss: 0.3653 - val_phraseology_loss: 0.5171 - val_grammar_loss: 0.5039 - val_conventions_loss: 0.4608\n",
            "Epoch 10/10\n",
            "98/98 [==============================] - 49s 496ms/step - loss: 3.7413 - cohesion_loss: 0.5961 - syntax_loss: 0.6167 - vocabulary_loss: 0.5697 - phraseology_loss: 0.6282 - grammar_loss: 0.6697 - conventions_loss: 0.6036 - val_loss: 2.7670 - val_cohesion_loss: 0.4553 - val_syntax_loss: 0.4620 - val_vocabulary_loss: 0.3534 - val_phraseology_loss: 0.4425 - val_grammar_loss: 0.5324 - val_conventions_loss: 0.4647\n"
          ]
        }
      ],
      "source": [
        "history = model.fit(X, y, epochs=10, validation_split=0.2, shuffle=True, verbose=1)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "jv3q7nOgvVNz"
      },
      "outputs": [],
      "source": [
        "def plot_loss(history):\n",
        "    plt.plot(history.history['loss'], label='loss')\n",
        "    plt.plot(history.history['val_loss'], label='val_loss')\n",
        "    plt.xlabel('Epoch')\n",
        "    plt.ylabel('Error [MSE]')\n",
        "    plt.legend()\n",
        "    plt.grid(True)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 502
        },
        "id": "HLujd8wtvbk1",
        "outputId": "ee68f996-8ef7-4e8e-f47a-6b1393f2bed0"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1080x576 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "plot_loss(history)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QrVqDDtJvv_Y"
      },
      "source": [
        "### Saving LSTM"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "yCu66XlNvuIC"
      },
      "outputs": [],
      "source": [
        "model.save('/content/gdrive/MyDrive/Colab Notebooks/NLP/Project/feedback-prize-english-language-learning/lstm_model.h5')"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "95LfCj-nF-5-"
      },
      "source": [
        "### Loading the LSTM model "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "VPTEBgwHd8Lg"
      },
      "outputs": [],
      "source": [
        "from tensorflow import keras"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "QKfEXVwGR0rx"
      },
      "outputs": [],
      "source": [
        "model = keras.models.load_model('/content/gdrive/MyDrive/Colab Notebooks/NLP/Project/feedback-prize-english-language-learning/lstm_model.h5')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 224
        },
        "id": "FZqQ9Dmlvykl",
        "outputId": "4bada3ec-cb8f-4cd0-8d6a-b1b5264f6465"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "123/123 [==============================] - 23s 190ms/step\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "        text_id  cohesion    syntax  vocabulary  phraseology   grammar  \\\n",
              "0  0016926B079C  3.282510  3.273651    3.369876     3.128019  3.306530   \n",
              "1  0022683E9EA5  2.927514  2.847336    2.993550     2.763618  2.923829   \n",
              "2  00299B378633  3.197886  3.168996    3.275342     3.037946  3.211999   \n",
              "3  003885A45F42  3.421712  3.453564    3.535726     3.284441  3.472092   \n",
              "4  0049B1DF5CCC  3.068267  3.010552    3.134091     2.900490  3.070340   \n",
              "\n",
              "   conventions  \n",
              "0     3.335952  \n",
              "1     2.926355  \n",
              "2     3.232405  \n",
              "3     3.515730  \n",
              "4     3.079011  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-9c25e051-fafb-45ec-83aa-c466400d58a0\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <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>text_id</th>\n",
              "      <th>cohesion</th>\n",
              "      <th>syntax</th>\n",
              "      <th>vocabulary</th>\n",
              "      <th>phraseology</th>\n",
              "      <th>grammar</th>\n",
              "      <th>conventions</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0016926B079C</td>\n",
              "      <td>3.282510</td>\n",
              "      <td>3.273651</td>\n",
              "      <td>3.369876</td>\n",
              "      <td>3.128019</td>\n",
              "      <td>3.306530</td>\n",
              "      <td>3.335952</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0022683E9EA5</td>\n",
              "      <td>2.927514</td>\n",
              "      <td>2.847336</td>\n",
              "      <td>2.993550</td>\n",
              "      <td>2.763618</td>\n",
              "      <td>2.923829</td>\n",
              "      <td>2.926355</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>00299B378633</td>\n",
              "      <td>3.197886</td>\n",
              "      <td>3.168996</td>\n",
              "      <td>3.275342</td>\n",
              "      <td>3.037946</td>\n",
              "      <td>3.211999</td>\n",
              "      <td>3.232405</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>003885A45F42</td>\n",
              "      <td>3.421712</td>\n",
              "      <td>3.453564</td>\n",
              "      <td>3.535726</td>\n",
              "      <td>3.284441</td>\n",
              "      <td>3.472092</td>\n",
              "      <td>3.515730</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>0049B1DF5CCC</td>\n",
              "      <td>3.068267</td>\n",
              "      <td>3.010552</td>\n",
              "      <td>3.134091</td>\n",
              "      <td>2.900490</td>\n",
              "      <td>3.070340</td>\n",
              "      <td>3.079011</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-9c25e051-fafb-45ec-83aa-c466400d58a0')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-9c25e051-fafb-45ec-83aa-c466400d58a0 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-9c25e051-fafb-45ec-83aa-c466400d58a0');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 14
        }
      ],
      "source": [
        "X_test = pad_train\n",
        "results = model.predict(X_test)\n",
        "lstm_submission_train = pd.DataFrame()\n",
        "lstm_submission_train['text_id'] = train_df['text_id']\n",
        "for i in range(6):\n",
        "    lstm_submission_train[i] = pd.Series([float(x) for x in results[i]])\n",
        "lstm_submission_train.columns = train_df.drop(columns=['full_text', 'num_words']).columns\n",
        "lstm_submission_train['text_id'] = train_df['text_id']\n",
        "\n",
        "lstm_submission_train.head()"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "target_cols = ['cohesion', 'syntax', 'vocabulary', 'phraseology', 'grammar', 'conventions']"
      ],
      "metadata": {
        "id": "OAD0ufPeWn20"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "X_test = pad_test\n",
        "results = model.predict(X_test)\n",
        "lstm_submission = pd.DataFrame()\n",
        "lstm_submission['text_id'] = test_df['text_id']\n",
        "for i in range(6):\n",
        "    lstm_submission[i] = pd.Series([float(x) for x in results[i]])\n",
        "lstm_submission.columns = train_df.drop(columns=['full_text', 'num_words']).columns\n",
        "lstm_submission['text_id'] = test_df['text_id']\n",
        "\n",
        "lstm_submission.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 161
        },
        "id": "LAxwBBJw2GJn",
        "outputId": "a8989a42-5b31-4f92-8117-24bba201adbe"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1/1 [==============================] - 0s 186ms/step\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "        text_id  cohesion    syntax  vocabulary  phraseology   grammar  \\\n",
              "0  0000C359D63E  2.848783  2.757720    2.916774     2.689219  2.842958   \n",
              "1  000BAD50D026  3.459145  3.497568    3.576564     3.321732  3.509279   \n",
              "2  00367BB2546B  3.021687  2.954948    3.085000     2.853040  3.020295   \n",
              "\n",
              "   conventions  \n",
              "0     2.843955  \n",
              "1     3.565041  \n",
              "2     3.025939  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-9db39234-dce1-448f-831a-9b5ff588b1cb\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <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>text_id</th>\n",
              "      <th>cohesion</th>\n",
              "      <th>syntax</th>\n",
              "      <th>vocabulary</th>\n",
              "      <th>phraseology</th>\n",
              "      <th>grammar</th>\n",
              "      <th>conventions</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0000C359D63E</td>\n",
              "      <td>2.848783</td>\n",
              "      <td>2.757720</td>\n",
              "      <td>2.916774</td>\n",
              "      <td>2.689219</td>\n",
              "      <td>2.842958</td>\n",
              "      <td>2.843955</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>000BAD50D026</td>\n",
              "      <td>3.459145</td>\n",
              "      <td>3.497568</td>\n",
              "      <td>3.576564</td>\n",
              "      <td>3.321732</td>\n",
              "      <td>3.509279</td>\n",
              "      <td>3.565041</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>00367BB2546B</td>\n",
              "      <td>3.021687</td>\n",
              "      <td>2.954948</td>\n",
              "      <td>3.085000</td>\n",
              "      <td>2.853040</td>\n",
              "      <td>3.020295</td>\n",
              "      <td>3.025939</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-9db39234-dce1-448f-831a-9b5ff588b1cb')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-9db39234-dce1-448f-831a-9b5ff588b1cb button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-9db39234-dce1-448f-831a-9b5ff588b1cb');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 24
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def comp_score(y_true,y_pred):\n",
        "    rmse_scores = []\n",
        "    for i in range(len(target_cols)):\n",
        "        rmse_scores.append(np.sqrt(mean_squared_error(y_true[:,i],y_pred[:,i])))\n",
        "    return np.mean(rmse_scores)"
      ],
      "metadata": {
        "id": "gjBpxWoqWZVp"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(comp_score(np.array(train_df[target_cols]), np.array(lstm_submission_train[target_cols])))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "0Tcy0pYMWJ41",
        "outputId": "d59f5482-c3e2-4d4f-a336-38a7eb43fce5"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "0.5985894818390786\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "uDA4jNWK3LHB"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gKCcQa-WRToQ"
      },
      "source": [
        "### Finding the stronger and weaker part of Essay "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "vA108mfdSJI0"
      },
      "outputs": [],
      "source": [
        "para_list = []\n",
        "key_list = []\n",
        "\n",
        "for index in test_df.index:\n",
        "  key = test_df['text_id'].iloc[index]\n",
        "  text_list = test_df['full_text'].iloc[index].split(\"\\n\")\n",
        "  final_text = []\n",
        "  for text in text_list:\n",
        "    if len(text) > 10:\n",
        "      final_text.append(text)\n",
        "\n",
        "  para_list.extend(final_text)\n",
        "  key_list.extend([key for i in range(len(final_text))])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "XfBhkIAiTJKp"
      },
      "outputs": [],
      "source": [
        "para_df = pd.DataFrame()\n",
        "para_df['text_id'] = key_list\n",
        "para_df['full_text'] = para_list"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 224
        },
        "id": "KSxmvwrOofak",
        "outputId": "fe35723c-ebf6-43b2-f98d-66c4c3fdbfbc"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1/1 [==============================] - 0s 193ms/step\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "        text_id  cohesion    syntax  vocabulary  phraseology   grammar  \\\n",
              "0  0000C359D63E  3.309421  3.308018    3.401232     3.158089  3.338328   \n",
              "1  0000C359D63E  3.158731  3.121032    3.231731     2.996434  3.169460   \n",
              "2  0000C359D63E  2.723413  2.608602    2.782124     2.558853  2.707006   \n",
              "3  0000C359D63E  3.456025  3.493727    3.573276     3.318552  3.506110   \n",
              "4  0000C359D63E  2.788485  2.687580    2.856175     2.627550  2.779032   \n",
              "\n",
              "   conventions                                          full_text  \n",
              "0     3.369349  when a person has no experience on a job their...  \n",
              "1     3.185546  yes, even thought you may not have experience ...  \n",
              "2     2.702266  If you dont have experence in a restaurant for...  \n",
              "3     3.560393  My dad has always talk with my cousins that wh...  \n",
              "4     2.778413  In every job their is always going to be a per...  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-bd6f3d7e-1caf-4448-a472-cda44119c1ae\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <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>text_id</th>\n",
              "      <th>cohesion</th>\n",
              "      <th>syntax</th>\n",
              "      <th>vocabulary</th>\n",
              "      <th>phraseology</th>\n",
              "      <th>grammar</th>\n",
              "      <th>conventions</th>\n",
              "      <th>full_text</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0000C359D63E</td>\n",
              "      <td>3.309421</td>\n",
              "      <td>3.308018</td>\n",
              "      <td>3.401232</td>\n",
              "      <td>3.158089</td>\n",
              "      <td>3.338328</td>\n",
              "      <td>3.369349</td>\n",
              "      <td>when a person has no experience on a job their...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0000C359D63E</td>\n",
              "      <td>3.158731</td>\n",
              "      <td>3.121032</td>\n",
              "      <td>3.231731</td>\n",
              "      <td>2.996434</td>\n",
              "      <td>3.169460</td>\n",
              "      <td>3.185546</td>\n",
              "      <td>yes, even thought you may not have experience ...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>0000C359D63E</td>\n",
              "      <td>2.723413</td>\n",
              "      <td>2.608602</td>\n",
              "      <td>2.782124</td>\n",
              "      <td>2.558853</td>\n",
              "      <td>2.707006</td>\n",
              "      <td>2.702266</td>\n",
              "      <td>If you dont have experence in a restaurant for...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>0000C359D63E</td>\n",
              "      <td>3.456025</td>\n",
              "      <td>3.493727</td>\n",
              "      <td>3.573276</td>\n",
              "      <td>3.318552</td>\n",
              "      <td>3.506110</td>\n",
              "      <td>3.560393</td>\n",
              "      <td>My dad has always talk with my cousins that wh...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>0000C359D63E</td>\n",
              "      <td>2.788485</td>\n",
              "      <td>2.687580</td>\n",
              "      <td>2.856175</td>\n",
              "      <td>2.627550</td>\n",
              "      <td>2.779032</td>\n",
              "      <td>2.778413</td>\n",
              "      <td>In every job their is always going to be a per...</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-bd6f3d7e-1caf-4448-a472-cda44119c1ae')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-bd6f3d7e-1caf-4448-a472-cda44119c1ae button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-bd6f3d7e-1caf-4448-a472-cda44119c1ae');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 17
        }
      ],
      "source": [
        "para_seq = tokenizer.texts_to_sequences(para_df['full_text'])\n",
        "pad_para = pad_sequences(para_seq, maxlen=max_words, truncating='post')\n",
        "\n",
        "X_test_para = pad_para\n",
        "results = model.predict(X_test_para)\n",
        "\n",
        "lstm_submission_para = pd.DataFrame()\n",
        "lstm_submission_para['text_id'] = para_df['text_id']\n",
        "\n",
        "for i in range(6):\n",
        "    lstm_submission_para[i] = pd.Series([float(x) for x in results[i]])\n",
        "lstm_submission_para.columns = train_df.drop(columns=['full_text', 'num_words']).columns\n",
        "lstm_submission_para['text_id'] = para_df['text_id']\n",
        "lstm_submission_para['full_text'] = para_df['full_text']\n",
        "\n",
        "lstm_submission_para.head()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 488
        },
        "id": "SbXj6284URVi",
        "outputId": "ef41dc0b-eaaa-4c49-bc3c-803044e60167"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "         text_id  cohesion    syntax  vocabulary  phraseology   grammar  \\\n",
              "0   0000C359D63E  3.309421  3.308018    3.401232     3.158089  3.338328   \n",
              "1   0000C359D63E  3.158731  3.121032    3.231731     2.996434  3.169460   \n",
              "2   0000C359D63E  2.723413  2.608602    2.782124     2.558853  2.707006   \n",
              "3   0000C359D63E  3.456025  3.493727    3.573276     3.318552  3.506110   \n",
              "4   0000C359D63E  2.788485  2.687580    2.856175     2.627550  2.779032   \n",
              "5   0000C359D63E  2.848713  2.757782    2.916678     2.689212  2.842931   \n",
              "6   000BAD50D026  3.307261  3.302990    3.396756     3.153998  3.332157   \n",
              "7   000BAD50D026  3.301543  3.296459    3.390664     3.148166  3.326500   \n",
              "8   000BAD50D026  3.459129  3.497639    3.576541     3.321793  3.509279   \n",
              "9   00367BB2546B  3.248961  3.232156    3.332257     3.093129  3.269393   \n",
              "10  00367BB2546B  3.036406  2.973136    3.101053     2.868746  3.037072   \n",
              "11  00367BB2546B  3.322196  3.324109    3.416587     3.172517  3.353690   \n",
              "12  00367BB2546B  3.177976  3.143913    3.252408     3.015637  3.186189   \n",
              "13  00367BB2546B  3.021718  2.954892    3.085028     2.853001  3.020224   \n",
              "\n",
              "    conventions                                          full_text  \n",
              "0      3.369349  when a person has no experience on a job their...  \n",
              "1      3.185546  yes, even thought you may not have experience ...  \n",
              "2      2.702266  If you dont have experence in a restaurant for...  \n",
              "3      3.560393  My dad has always talk with my cousins that wh...  \n",
              "4      2.778413  In every job their is always going to be a per...  \n",
              "5      2.843895  I think yes, you can be a good candidate to be...  \n",
              "6      3.368976  Do you think students would benefit from being...  \n",
              "7      3.361804  Yes! its benefit for student who attend classe...  \n",
              "8      3.565007  some student attend classes from home because ...  \n",
              "9      3.294416  Thomas Jefferson once states that \"it is wonde...  \n",
              "10     3.043085  The first reason that always doing is better t...  \n",
              "11     3.386588  Secondly,accomplishing your goals makes you wa...  \n",
              "12     3.209408  Finally it give you the opportunity to learn n...  \n",
              "13     3.025904  Although some may disagree with me and say tha...  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-8fdffc3f-11a4-4184-8980-c0229e11fa44\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <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>text_id</th>\n",
              "      <th>cohesion</th>\n",
              "      <th>syntax</th>\n",
              "      <th>vocabulary</th>\n",
              "      <th>phraseology</th>\n",
              "      <th>grammar</th>\n",
              "      <th>conventions</th>\n",
              "      <th>full_text</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0000C359D63E</td>\n",
              "      <td>3.309421</td>\n",
              "      <td>3.308018</td>\n",
              "      <td>3.401232</td>\n",
              "      <td>3.158089</td>\n",
              "      <td>3.338328</td>\n",
              "      <td>3.369349</td>\n",
              "      <td>when a person has no experience on a job their...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0000C359D63E</td>\n",
              "      <td>3.158731</td>\n",
              "      <td>3.121032</td>\n",
              "      <td>3.231731</td>\n",
              "      <td>2.996434</td>\n",
              "      <td>3.169460</td>\n",
              "      <td>3.185546</td>\n",
              "      <td>yes, even thought you may not have experience ...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>0000C359D63E</td>\n",
              "      <td>2.723413</td>\n",
              "      <td>2.608602</td>\n",
              "      <td>2.782124</td>\n",
              "      <td>2.558853</td>\n",
              "      <td>2.707006</td>\n",
              "      <td>2.702266</td>\n",
              "      <td>If you dont have experence in a restaurant for...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>0000C359D63E</td>\n",
              "      <td>3.456025</td>\n",
              "      <td>3.493727</td>\n",
              "      <td>3.573276</td>\n",
              "      <td>3.318552</td>\n",
              "      <td>3.506110</td>\n",
              "      <td>3.560393</td>\n",
              "      <td>My dad has always talk with my cousins that wh...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>0000C359D63E</td>\n",
              "      <td>2.788485</td>\n",
              "      <td>2.687580</td>\n",
              "      <td>2.856175</td>\n",
              "      <td>2.627550</td>\n",
              "      <td>2.779032</td>\n",
              "      <td>2.778413</td>\n",
              "      <td>In every job their is always going to be a per...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>0000C359D63E</td>\n",
              "      <td>2.848713</td>\n",
              "      <td>2.757782</td>\n",
              "      <td>2.916678</td>\n",
              "      <td>2.689212</td>\n",
              "      <td>2.842931</td>\n",
              "      <td>2.843895</td>\n",
              "      <td>I think yes, you can be a good candidate to be...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>000BAD50D026</td>\n",
              "      <td>3.307261</td>\n",
              "      <td>3.302990</td>\n",
              "      <td>3.396756</td>\n",
              "      <td>3.153998</td>\n",
              "      <td>3.332157</td>\n",
              "      <td>3.368976</td>\n",
              "      <td>Do you think students would benefit from being...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>000BAD50D026</td>\n",
              "      <td>3.301543</td>\n",
              "      <td>3.296459</td>\n",
              "      <td>3.390664</td>\n",
              "      <td>3.148166</td>\n",
              "      <td>3.326500</td>\n",
              "      <td>3.361804</td>\n",
              "      <td>Yes! its benefit for student who attend classe...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>000BAD50D026</td>\n",
              "      <td>3.459129</td>\n",
              "      <td>3.497639</td>\n",
              "      <td>3.576541</td>\n",
              "      <td>3.321793</td>\n",
              "      <td>3.509279</td>\n",
              "      <td>3.565007</td>\n",
              "      <td>some student attend classes from home because ...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9</th>\n",
              "      <td>00367BB2546B</td>\n",
              "      <td>3.248961</td>\n",
              "      <td>3.232156</td>\n",
              "      <td>3.332257</td>\n",
              "      <td>3.093129</td>\n",
              "      <td>3.269393</td>\n",
              "      <td>3.294416</td>\n",
              "      <td>Thomas Jefferson once states that \"it is wonde...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10</th>\n",
              "      <td>00367BB2546B</td>\n",
              "      <td>3.036406</td>\n",
              "      <td>2.973136</td>\n",
              "      <td>3.101053</td>\n",
              "      <td>2.868746</td>\n",
              "      <td>3.037072</td>\n",
              "      <td>3.043085</td>\n",
              "      <td>The first reason that always doing is better t...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>11</th>\n",
              "      <td>00367BB2546B</td>\n",
              "      <td>3.322196</td>\n",
              "      <td>3.324109</td>\n",
              "      <td>3.416587</td>\n",
              "      <td>3.172517</td>\n",
              "      <td>3.353690</td>\n",
              "      <td>3.386588</td>\n",
              "      <td>Secondly,accomplishing your goals makes you wa...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>12</th>\n",
              "      <td>00367BB2546B</td>\n",
              "      <td>3.177976</td>\n",
              "      <td>3.143913</td>\n",
              "      <td>3.252408</td>\n",
              "      <td>3.015637</td>\n",
              "      <td>3.186189</td>\n",
              "      <td>3.209408</td>\n",
              "      <td>Finally it give you the opportunity to learn n...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>13</th>\n",
              "      <td>00367BB2546B</td>\n",
              "      <td>3.021718</td>\n",
              "      <td>2.954892</td>\n",
              "      <td>3.085028</td>\n",
              "      <td>2.853001</td>\n",
              "      <td>3.020224</td>\n",
              "      <td>3.025904</td>\n",
              "      <td>Although some may disagree with me and say tha...</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-8fdffc3f-11a4-4184-8980-c0229e11fa44')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-8fdffc3f-11a4-4184-8980-c0229e11fa44 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-8fdffc3f-11a4-4184-8980-c0229e11fa44');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 18
        }
      ],
      "source": [
        "lstm_submission_para"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "s1fZZ2HEDAy2"
      },
      "source": [
        "### Grammar"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "9lb-uXJwEMRu",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "9934dbac-a828-49b8-ccbd-71eec0193723"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "Downloading LanguageTool 5.7: 100%|██████████| 225M/225M [00:03<00:00, 66.7MB/s]\n",
            "INFO:language_tool_python.download_lt:Unzipping /tmp/tmpfv_miv46.zip to /root/.cache/language_tool_python.\n",
            "INFO:language_tool_python.download_lt:Downloaded https://www.languagetool.org/download/LanguageTool-5.7.zip to /root/.cache/language_tool_python.\n"
          ]
        }
      ],
      "source": [
        "tool = language_tool_python.LanguageTool('en-US')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "kENLYsJ0X-Pi"
      },
      "outputs": [],
      "source": [
        "def get_feedback(df):\n",
        "\n",
        "  issues_list = []\n",
        "\n",
        "  for ind in df.index:\n",
        "    text = df[\"full_text\"].iloc[ind]\n",
        "    \n",
        "    # get the matches\n",
        "    matches = tool.check(text)\n",
        "    \n",
        "    my_mistakes = []\n",
        "    my_corrections = []\n",
        "    start_positions = []\n",
        "    end_positions = []\n",
        "    \n",
        "    for rules in matches:\n",
        "        if len(rules.replacements)>0:\n",
        "            start_positions.append(rules.offset)\n",
        "            end_positions.append(rules.errorLength+rules.offset)\n",
        "            my_mistakes.append(text[rules.offset:rules.errorLength+rules.offset])\n",
        "            my_corrections.append(rules.replacements[0])\n",
        "        \n",
        "    \n",
        "        \n",
        "    my_new_text = list(text)\n",
        "    \n",
        "    \n",
        "    for m in range(len(start_positions)):\n",
        "        for i in range(len(text)):\n",
        "            my_new_text[start_positions[m]] = my_corrections[m]\n",
        "            if (i>start_positions[m] and i<end_positions[m]):\n",
        "                my_new_text[i]=\"\"\n",
        "        \n",
        "    my_new_text = \"\".join(my_new_text)\n",
        "\n",
        "    issues_list.append(list(zip(my_mistakes,my_corrections)))\n",
        "  \n",
        "  return issues_list"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "6waD8vzZEDb-"
      },
      "outputs": [],
      "source": [
        "issues_list = get_feedback(lstm_submission_para)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "kZ1ss_3yFIs1"
      },
      "outputs": [],
      "source": [
        "lstm_submission_para['issues_list'] = issues_list"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 250
        },
        "id": "_EcisgBuFivJ",
        "outputId": "c91655fb-5524-41a3-cafc-b36269f7154e"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "        text_id  cohesion    syntax  vocabulary  phraseology   grammar  \\\n",
              "0  0000C359D63E  3.309421  3.308018    3.401232     3.158089  3.338328   \n",
              "1  0000C359D63E  3.158731  3.121032    3.231731     2.996434  3.169460   \n",
              "2  0000C359D63E  2.723413  2.608602    2.782124     2.558853  2.707006   \n",
              "3  0000C359D63E  3.456025  3.493727    3.573276     3.318552  3.506110   \n",
              "4  0000C359D63E  2.788485  2.687580    2.856175     2.627550  2.779032   \n",
              "\n",
              "   conventions                                          full_text  \\\n",
              "0     3.369349  when a person has no experience on a job their...   \n",
              "1     3.185546  yes, even thought you may not have experience ...   \n",
              "2     2.702266  If you dont have experence in a restaurant for...   \n",
              "3     3.560393  My dad has always talk with my cousins that wh...   \n",
              "4     2.778413  In every job their is always going to be a per...   \n",
              "\n",
              "                                         issues_list  \n",
              "0  [(when, When), (their, there), (explane, enpla...  \n",
              "1                        [(yes, Yes), (,you, , you)]  \n",
              "2  [(dont, don't), (experence, experience), (expe...  \n",
              "3  [(talk, talked), (their, there), (cosuin, cous...  \n",
              "4  [(their, there), (the learn, to learn), (thay,...  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-151cd12a-106c-4642-b0b5-4f76ad964845\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <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>text_id</th>\n",
              "      <th>cohesion</th>\n",
              "      <th>syntax</th>\n",
              "      <th>vocabulary</th>\n",
              "      <th>phraseology</th>\n",
              "      <th>grammar</th>\n",
              "      <th>conventions</th>\n",
              "      <th>full_text</th>\n",
              "      <th>issues_list</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0000C359D63E</td>\n",
              "      <td>3.309421</td>\n",
              "      <td>3.308018</td>\n",
              "      <td>3.401232</td>\n",
              "      <td>3.158089</td>\n",
              "      <td>3.338328</td>\n",
              "      <td>3.369349</td>\n",
              "      <td>when a person has no experience on a job their...</td>\n",
              "      <td>[(when, When), (their, there), (explane, enpla...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0000C359D63E</td>\n",
              "      <td>3.158731</td>\n",
              "      <td>3.121032</td>\n",
              "      <td>3.231731</td>\n",
              "      <td>2.996434</td>\n",
              "      <td>3.169460</td>\n",
              "      <td>3.185546</td>\n",
              "      <td>yes, even thought you may not have experience ...</td>\n",
              "      <td>[(yes, Yes), (,you, , you)]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>0000C359D63E</td>\n",
              "      <td>2.723413</td>\n",
              "      <td>2.608602</td>\n",
              "      <td>2.782124</td>\n",
              "      <td>2.558853</td>\n",
              "      <td>2.707006</td>\n",
              "      <td>2.702266</td>\n",
              "      <td>If you dont have experence in a restaurant for...</td>\n",
              "      <td>[(dont, don't), (experence, experience), (expe...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>0000C359D63E</td>\n",
              "      <td>3.456025</td>\n",
              "      <td>3.493727</td>\n",
              "      <td>3.573276</td>\n",
              "      <td>3.318552</td>\n",
              "      <td>3.506110</td>\n",
              "      <td>3.560393</td>\n",
              "      <td>My dad has always talk with my cousins that wh...</td>\n",
              "      <td>[(talk, talked), (their, there), (cosuin, cous...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>0000C359D63E</td>\n",
              "      <td>2.788485</td>\n",
              "      <td>2.687580</td>\n",
              "      <td>2.856175</td>\n",
              "      <td>2.627550</td>\n",
              "      <td>2.779032</td>\n",
              "      <td>2.778413</td>\n",
              "      <td>In every job their is always going to be a per...</td>\n",
              "      <td>[(their, there), (the learn, to learn), (thay,...</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-151cd12a-106c-4642-b0b5-4f76ad964845')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-151cd12a-106c-4642-b0b5-4f76ad964845 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-151cd12a-106c-4642-b0b5-4f76ad964845');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 23
        }
      ],
      "source": [
        "lstm_submission_para.head()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0JrB-6-bcMo4"
      },
      "source": [
        "### Weak paragraph"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "ebhob4vcb_QT"
      },
      "outputs": [],
      "source": [
        "lstm_submission[\"mean\"] = lstm_submission[['cohesion', 'syntax', 'vocabulary', 'phraseology', 'grammar','conventions']].mean(axis=1)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Rs7OifDpfIr5"
      },
      "outputs": [],
      "source": [
        "id_mean_dict = {}\n",
        "\n",
        "for index in lstm_submission.index:\n",
        "  id = lstm_submission[\"text_id\"].iloc[index]\n",
        "  id_mean_dict[id] = lstm_submission[\"mean\"].iloc[index]\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "rLZIK4bXgQk0"
      },
      "outputs": [],
      "source": [
        "lstm_submission_para[\"strength\"] = \"Weak\""
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "h4NnO3rEgIqh",
        "outputId": "c1c97f9c-acad-4c46-f83a-bdfbe89906a1"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/pandas/core/indexing.py:1732: SettingWithCopyWarning: \n",
            "A value is trying to be set on a copy of a slice from a DataFrame\n",
            "\n",
            "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
            "  self._setitem_single_block(indexer, value, name)\n"
          ]
        }
      ],
      "source": [
        "for key,avg in id_mean_dict.items():\n",
        "  \n",
        "  temp  = lstm_submission_para[lstm_submission_para[\"text_id\"] == key ]\n",
        "  s_index = temp.loc[(temp['cohesion'] > avg) & (temp['syntax'] > avg) & (temp['vocabulary'] > avg) & (temp['phraseology'] > avg) & (temp['grammar'] > avg) & (temp['conventions'] > avg)].index\n",
        "  lstm_submission_para[\"strength\"].loc[s_index] = \"Strong\""
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 775
        },
        "id": "RLoViWRWiOkl",
        "outputId": "a740d8a9-7d09-48f9-8f79-d858780137ae"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "         text_id  cohesion    syntax  vocabulary  phraseology   grammar  \\\n",
              "0   0000C359D63E  3.309421  3.308018    3.401232     3.158089  3.338328   \n",
              "1   0000C359D63E  3.158731  3.121032    3.231731     2.996434  3.169460   \n",
              "2   0000C359D63E  2.723413  2.608602    2.782124     2.558853  2.707006   \n",
              "3   0000C359D63E  3.456025  3.493727    3.573276     3.318552  3.506110   \n",
              "4   0000C359D63E  2.788485  2.687580    2.856175     2.627550  2.779032   \n",
              "5   0000C359D63E  2.848713  2.757782    2.916678     2.689212  2.842931   \n",
              "6   000BAD50D026  3.307261  3.302990    3.396756     3.153998  3.332157   \n",
              "7   000BAD50D026  3.301543  3.296459    3.390664     3.148166  3.326500   \n",
              "8   000BAD50D026  3.459129  3.497639    3.576541     3.321793  3.509279   \n",
              "9   00367BB2546B  3.248961  3.232156    3.332257     3.093129  3.269393   \n",
              "10  00367BB2546B  3.036406  2.973136    3.101053     2.868746  3.037072   \n",
              "11  00367BB2546B  3.322196  3.324109    3.416587     3.172517  3.353690   \n",
              "12  00367BB2546B  3.177976  3.143913    3.252408     3.015637  3.186189   \n",
              "13  00367BB2546B  3.021718  2.954892    3.085028     2.853001  3.020224   \n",
              "\n",
              "    conventions                                          full_text  \\\n",
              "0      3.369349  when a person has no experience on a job their...   \n",
              "1      3.185546  yes, even thought you may not have experience ...   \n",
              "2      2.702266  If you dont have experence in a restaurant for...   \n",
              "3      3.560393  My dad has always talk with my cousins that wh...   \n",
              "4      2.778413  In every job their is always going to be a per...   \n",
              "5      2.843895  I think yes, you can be a good candidate to be...   \n",
              "6      3.368976  Do you think students would benefit from being...   \n",
              "7      3.361804  Yes! its benefit for student who attend classe...   \n",
              "8      3.565007  some student attend classes from home because ...   \n",
              "9      3.294416  Thomas Jefferson once states that \"it is wonde...   \n",
              "10     3.043085  The first reason that always doing is better t...   \n",
              "11     3.386588  Secondly,accomplishing your goals makes you wa...   \n",
              "12     3.209408  Finally it give you the opportunity to learn n...   \n",
              "13     3.025904  Although some may disagree with me and say tha...   \n",
              "\n",
              "                                          issues_list strength  \n",
              "0   [(when, When), (their, there), (explane, enpla...   Strong  \n",
              "1                         [(yes, Yes), (,you, , you)]   Strong  \n",
              "2   [(dont, don't), (experence, experience), (expe...     Weak  \n",
              "3   [(talk, talked), (their, there), (cosuin, cous...   Strong  \n",
              "4   [(their, there), (the learn, to learn), (thay,...     Weak  \n",
              "5   [(a, an), (dont, don't), (going, go), (Practic...     Weak  \n",
              "6                     [(able attend, able to attend)]     Weak  \n",
              "7   [(its, Its), (student want, students want), (t...     Weak  \n",
              "8   [(some, Some), (student attend, students atten...     Weak  \n",
              "9   [(,I, , I), (always, Always), (,it, , it), (,y...   Strong  \n",
              "10  [(when, When), (for, For), (,when, , when), (,...     Weak  \n",
              "11  [(,accomplishing, , accomplishing), (when, Whe...   Strong  \n",
              "12  [(Finally, Finally,), (give, gives), (by, By),...   Strong  \n",
              "13  [( ,, , ), (day to day, day-to-day), (in, In),...     Weak  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-fba15617-f5c9-4a37-820b-f22de30a7a9e\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <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>text_id</th>\n",
              "      <th>cohesion</th>\n",
              "      <th>syntax</th>\n",
              "      <th>vocabulary</th>\n",
              "      <th>phraseology</th>\n",
              "      <th>grammar</th>\n",
              "      <th>conventions</th>\n",
              "      <th>full_text</th>\n",
              "      <th>issues_list</th>\n",
              "      <th>strength</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0000C359D63E</td>\n",
              "      <td>3.309421</td>\n",
              "      <td>3.308018</td>\n",
              "      <td>3.401232</td>\n",
              "      <td>3.158089</td>\n",
              "      <td>3.338328</td>\n",
              "      <td>3.369349</td>\n",
              "      <td>when a person has no experience on a job their...</td>\n",
              "      <td>[(when, When), (their, there), (explane, enpla...</td>\n",
              "      <td>Strong</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0000C359D63E</td>\n",
              "      <td>3.158731</td>\n",
              "      <td>3.121032</td>\n",
              "      <td>3.231731</td>\n",
              "      <td>2.996434</td>\n",
              "      <td>3.169460</td>\n",
              "      <td>3.185546</td>\n",
              "      <td>yes, even thought you may not have experience ...</td>\n",
              "      <td>[(yes, Yes), (,you, , you)]</td>\n",
              "      <td>Strong</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>0000C359D63E</td>\n",
              "      <td>2.723413</td>\n",
              "      <td>2.608602</td>\n",
              "      <td>2.782124</td>\n",
              "      <td>2.558853</td>\n",
              "      <td>2.707006</td>\n",
              "      <td>2.702266</td>\n",
              "      <td>If you dont have experence in a restaurant for...</td>\n",
              "      <td>[(dont, don't), (experence, experience), (expe...</td>\n",
              "      <td>Weak</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>0000C359D63E</td>\n",
              "      <td>3.456025</td>\n",
              "      <td>3.493727</td>\n",
              "      <td>3.573276</td>\n",
              "      <td>3.318552</td>\n",
              "      <td>3.506110</td>\n",
              "      <td>3.560393</td>\n",
              "      <td>My dad has always talk with my cousins that wh...</td>\n",
              "      <td>[(talk, talked), (their, there), (cosuin, cous...</td>\n",
              "      <td>Strong</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>0000C359D63E</td>\n",
              "      <td>2.788485</td>\n",
              "      <td>2.687580</td>\n",
              "      <td>2.856175</td>\n",
              "      <td>2.627550</td>\n",
              "      <td>2.779032</td>\n",
              "      <td>2.778413</td>\n",
              "      <td>In every job their is always going to be a per...</td>\n",
              "      <td>[(their, there), (the learn, to learn), (thay,...</td>\n",
              "      <td>Weak</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>0000C359D63E</td>\n",
              "      <td>2.848713</td>\n",
              "      <td>2.757782</td>\n",
              "      <td>2.916678</td>\n",
              "      <td>2.689212</td>\n",
              "      <td>2.842931</td>\n",
              "      <td>2.843895</td>\n",
              "      <td>I think yes, you can be a good candidate to be...</td>\n",
              "      <td>[(a, an), (dont, don't), (going, go), (Practic...</td>\n",
              "      <td>Weak</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>000BAD50D026</td>\n",
              "      <td>3.307261</td>\n",
              "      <td>3.302990</td>\n",
              "      <td>3.396756</td>\n",
              "      <td>3.153998</td>\n",
              "      <td>3.332157</td>\n",
              "      <td>3.368976</td>\n",
              "      <td>Do you think students would benefit from being...</td>\n",
              "      <td>[(able attend, able to attend)]</td>\n",
              "      <td>Weak</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>000BAD50D026</td>\n",
              "      <td>3.301543</td>\n",
              "      <td>3.296459</td>\n",
              "      <td>3.390664</td>\n",
              "      <td>3.148166</td>\n",
              "      <td>3.326500</td>\n",
              "      <td>3.361804</td>\n",
              "      <td>Yes! its benefit for student who attend classe...</td>\n",
              "      <td>[(its, Its), (student want, students want), (t...</td>\n",
              "      <td>Weak</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>000BAD50D026</td>\n",
              "      <td>3.459129</td>\n",
              "      <td>3.497639</td>\n",
              "      <td>3.576541</td>\n",
              "      <td>3.321793</td>\n",
              "      <td>3.509279</td>\n",
              "      <td>3.565007</td>\n",
              "      <td>some student attend classes from home because ...</td>\n",
              "      <td>[(some, Some), (student attend, students atten...</td>\n",
              "      <td>Weak</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9</th>\n",
              "      <td>00367BB2546B</td>\n",
              "      <td>3.248961</td>\n",
              "      <td>3.232156</td>\n",
              "      <td>3.332257</td>\n",
              "      <td>3.093129</td>\n",
              "      <td>3.269393</td>\n",
              "      <td>3.294416</td>\n",
              "      <td>Thomas Jefferson once states that \"it is wonde...</td>\n",
              "      <td>[(,I, , I), (always, Always), (,it, , it), (,y...</td>\n",
              "      <td>Strong</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10</th>\n",
              "      <td>00367BB2546B</td>\n",
              "      <td>3.036406</td>\n",
              "      <td>2.973136</td>\n",
              "      <td>3.101053</td>\n",
              "      <td>2.868746</td>\n",
              "      <td>3.037072</td>\n",
              "      <td>3.043085</td>\n",
              "      <td>The first reason that always doing is better t...</td>\n",
              "      <td>[(when, When), (for, For), (,when, , when), (,...</td>\n",
              "      <td>Weak</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>11</th>\n",
              "      <td>00367BB2546B</td>\n",
              "      <td>3.322196</td>\n",
              "      <td>3.324109</td>\n",
              "      <td>3.416587</td>\n",
              "      <td>3.172517</td>\n",
              "      <td>3.353690</td>\n",
              "      <td>3.386588</td>\n",
              "      <td>Secondly,accomplishing your goals makes you wa...</td>\n",
              "      <td>[(,accomplishing, , accomplishing), (when, Whe...</td>\n",
              "      <td>Strong</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>12</th>\n",
              "      <td>00367BB2546B</td>\n",
              "      <td>3.177976</td>\n",
              "      <td>3.143913</td>\n",
              "      <td>3.252408</td>\n",
              "      <td>3.015637</td>\n",
              "      <td>3.186189</td>\n",
              "      <td>3.209408</td>\n",
              "      <td>Finally it give you the opportunity to learn n...</td>\n",
              "      <td>[(Finally, Finally,), (give, gives), (by, By),...</td>\n",
              "      <td>Strong</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>13</th>\n",
              "      <td>00367BB2546B</td>\n",
              "      <td>3.021718</td>\n",
              "      <td>2.954892</td>\n",
              "      <td>3.085028</td>\n",
              "      <td>2.853001</td>\n",
              "      <td>3.020224</td>\n",
              "      <td>3.025904</td>\n",
              "      <td>Although some may disagree with me and say tha...</td>\n",
              "      <td>[( ,, , ), (day to day, day-to-day), (in, In),...</td>\n",
              "      <td>Weak</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-fba15617-f5c9-4a37-820b-f22de30a7a9e')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-fba15617-f5c9-4a37-820b-f22de30a7a9e button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-fba15617-f5c9-4a37-820b-f22de30a7a9e');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 28
        }
      ],
      "source": [
        "lstm_submission_para"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Making correction in text "
      ],
      "metadata": {
        "id": "5wz2O0qTZ4PO"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "updated_text = []\n",
        "for ind in lstm_submission_para.index:\n",
        "  full_text = lstm_submission_para.iloc[ind][\"full_text\"]\n",
        "  issues_list = lstm_submission_para.iloc[ind][\"issues_list\"]\n",
        "\n",
        "  for issue in issues_list:\n",
        "    full_text = full_text.replace(issue[0], issue[1])\n",
        "  \n",
        "  updated_text.append(full_text)"
      ],
      "metadata": {
        "id": "RImEAyCSYdn6"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "lstm_submission_para['updated_text'] = updated_text"
      ],
      "metadata": {
        "id": "NhBZ9BgzYdly"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Creating new dataframe with combining paragraph to essay "
      ],
      "metadata": {
        "id": "2z8mAb32Z9qA"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "lstm_submission_para['new_essay'] =lstm_submission_para.groupby(['text_id'])['updated_text'].transform(lambda x : '\\n'.join(x))"
      ],
      "metadata": {
        "id": "f-3kpy1-Yde0"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "lstm_submission_para_new =  lstm_submission_para[['text_id', 'new_essay']].drop_duplicates()"
      ],
      "metadata": {
        "id": "6Ufb9PHAYdaD"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "lstm_submission_para_new"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 143
        },
        "id": "VddrkSQabBou",
        "outputId": "00c37a58-538b-4687-8364-61e7e6c2cbfc"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "        text_id                                          new_essay\n",
              "0  0000C359D63E  When a person have no experience on a job ther...\n",
              "6  000BAD50D026  Do you think students would benefit from being...\n",
              "9  00367BB2546B  Thomas Jefferson once states that \"it is wonde..."
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-6f3aebc8-6ef5-46d5-ae3f-6a60dfd66b7f\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <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>text_id</th>\n",
              "      <th>new_essay</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0000C359D63E</td>\n",
              "      <td>When a person have no experience on a job ther...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>000BAD50D026</td>\n",
              "      <td>Do you think students would benefit from being...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9</th>\n",
              "      <td>00367BB2546B</td>\n",
              "      <td>Thomas Jefferson once states that \"it is wonde...</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-6f3aebc8-6ef5-46d5-ae3f-6a60dfd66b7f')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-6f3aebc8-6ef5-46d5-ae3f-6a60dfd66b7f button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-6f3aebc8-6ef5-46d5-ae3f-6a60dfd66b7f');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 50
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Evaluating essay on updated text"
      ],
      "metadata": {
        "id": "_hNqgXFcbzy-"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "test_seq = tokenizer.texts_to_sequences(lstm_submission_para_new['new_essay'])\n",
        "pad_test = pad_sequences(test_seq, maxlen=max_words, truncating='post')\n",
        "\n",
        "X_test = pad_test\n",
        "results = model.predict(X_test)\n",
        "lstm_submission_new = pd.DataFrame()\n",
        "lstm_submission_new['text_id'] = test_df['text_id']\n",
        "for i in range(6):\n",
        "    lstm_submission_new[i] = pd.Series([float(x) for x in results[i]])\n",
        "lstm_submission_new.columns = train_df.drop(columns=['full_text', 'num_words']).columns\n",
        "lstm_submission_new['text_id'] = test_df['text_id']\n",
        "\n",
        "lstm_submission_new.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 161
        },
        "id": "_uxLvaYRbziF",
        "outputId": "2747e57c-c57d-4b5b-ba05-591ed246c7aa"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1/1 [==============================] - 0s 178ms/step\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "        text_id  cohesion    syntax  vocabulary  phraseology   grammar  \\\n",
              "0  0000C359D63E  3.052899  2.994851    3.121371     2.888386  3.057064   \n",
              "1  000BAD50D026  3.460105  3.498620    3.577665     3.322809  3.510040   \n",
              "2  00367BB2546B  3.021687  2.954948    3.085000     2.853040  3.020295   \n",
              "\n",
              "   conventions  \n",
              "0     3.065004  \n",
              "1     3.566659  \n",
              "2     3.025939  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-1abc62c6-ce30-4225-8ea6-d71524b47856\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <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>text_id</th>\n",
              "      <th>cohesion</th>\n",
              "      <th>syntax</th>\n",
              "      <th>vocabulary</th>\n",
              "      <th>phraseology</th>\n",
              "      <th>grammar</th>\n",
              "      <th>conventions</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0000C359D63E</td>\n",
              "      <td>3.052899</td>\n",
              "      <td>2.994851</td>\n",
              "      <td>3.121371</td>\n",
              "      <td>2.888386</td>\n",
              "      <td>3.057064</td>\n",
              "      <td>3.065004</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>000BAD50D026</td>\n",
              "      <td>3.460105</td>\n",
              "      <td>3.498620</td>\n",
              "      <td>3.577665</td>\n",
              "      <td>3.322809</td>\n",
              "      <td>3.510040</td>\n",
              "      <td>3.566659</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>00367BB2546B</td>\n",
              "      <td>3.021687</td>\n",
              "      <td>2.954948</td>\n",
              "      <td>3.085000</td>\n",
              "      <td>2.853040</td>\n",
              "      <td>3.020295</td>\n",
              "      <td>3.025939</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-1abc62c6-ce30-4225-8ea6-d71524b47856')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-1abc62c6-ce30-4225-8ea6-d71524b47856 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-1abc62c6-ce30-4225-8ea6-d71524b47856');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 52
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "lstm_submission"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 143
        },
        "id": "apCG3ZRlfUkn",
        "outputId": "8ad06954-1e35-4030-cb31-cfc691db5280"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "        text_id  old_cohesion  old_syntax  old_vocabulary  old_phraseology  \\\n",
              "0  0000C359D63E      2.848783    2.757720        2.916774         2.689219   \n",
              "1  000BAD50D026      3.459145    3.497568        3.576564         3.321732   \n",
              "2  00367BB2546B      3.021687    2.954948        3.085000         2.853040   \n",
              "\n",
              "   old_grammar  old_conventions  \n",
              "0     2.842958         2.843955  \n",
              "1     3.509279         3.565041  \n",
              "2     3.020295         3.025939  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-491e539a-4bee-4fa3-be3f-725b54a376a8\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <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>text_id</th>\n",
              "      <th>old_cohesion</th>\n",
              "      <th>old_syntax</th>\n",
              "      <th>old_vocabulary</th>\n",
              "      <th>old_phraseology</th>\n",
              "      <th>old_grammar</th>\n",
              "      <th>old_conventions</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0000C359D63E</td>\n",
              "      <td>2.848783</td>\n",
              "      <td>2.757720</td>\n",
              "      <td>2.916774</td>\n",
              "      <td>2.689219</td>\n",
              "      <td>2.842958</td>\n",
              "      <td>2.843955</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>000BAD50D026</td>\n",
              "      <td>3.459145</td>\n",
              "      <td>3.497568</td>\n",
              "      <td>3.576564</td>\n",
              "      <td>3.321732</td>\n",
              "      <td>3.509279</td>\n",
              "      <td>3.565041</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>00367BB2546B</td>\n",
              "      <td>3.021687</td>\n",
              "      <td>2.954948</td>\n",
              "      <td>3.085000</td>\n",
              "      <td>2.853040</td>\n",
              "      <td>3.020295</td>\n",
              "      <td>3.025939</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-491e539a-4bee-4fa3-be3f-725b54a376a8')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-491e539a-4bee-4fa3-be3f-725b54a376a8 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-491e539a-4bee-4fa3-be3f-725b54a376a8');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 65
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Compare results"
      ],
      "metadata": {
        "id": "NEEdQXoec35w"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "lstm_submission.columns = ['text_id',\t'old_cohesion',\t'old_syntax',\t'old_vocabulary',\t'old_phraseology',\t'old_grammar',\t'old_conventions']"
      ],
      "metadata": {
        "id": "M8GsJC3tc7m_"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "merge_df = pd.merge(lstm_submission, lstm_submission_new, on=[\"text_id\"])"
      ],
      "metadata": {
        "id": "CaF7UomMfKBH"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "merge_df"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 207
        },
        "id": "M-bshWCIfJ-j",
        "outputId": "acaa4e69-520b-4ece-b51d-7635dccdddfa"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "        text_id  old_cohesion  old_syntax  old_vocabulary  old_phraseology  \\\n",
              "0  0000C359D63E      2.848783    2.757720        2.916774         2.689219   \n",
              "1  000BAD50D026      3.459145    3.497568        3.576564         3.321732   \n",
              "2  00367BB2546B      3.021687    2.954948        3.085000         2.853040   \n",
              "\n",
              "   old_grammar  old_conventions  cohesion    syntax  vocabulary  phraseology  \\\n",
              "0     2.842958         2.843955  3.052899  2.994851    3.121371     2.888386   \n",
              "1     3.509279         3.565041  3.460105  3.498620    3.577665     3.322809   \n",
              "2     3.020295         3.025939  3.021687  2.954948    3.085000     2.853040   \n",
              "\n",
              "    grammar  conventions  \n",
              "0  3.057064     3.065004  \n",
              "1  3.510040     3.566659  \n",
              "2  3.020295     3.025939  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-6771f561-b23b-4bcc-b2d4-adf1298b2e0c\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <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>text_id</th>\n",
              "      <th>old_cohesion</th>\n",
              "      <th>old_syntax</th>\n",
              "      <th>old_vocabulary</th>\n",
              "      <th>old_phraseology</th>\n",
              "      <th>old_grammar</th>\n",
              "      <th>old_conventions</th>\n",
              "      <th>cohesion</th>\n",
              "      <th>syntax</th>\n",
              "      <th>vocabulary</th>\n",
              "      <th>phraseology</th>\n",
              "      <th>grammar</th>\n",
              "      <th>conventions</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0000C359D63E</td>\n",
              "      <td>2.848783</td>\n",
              "      <td>2.757720</td>\n",
              "      <td>2.916774</td>\n",
              "      <td>2.689219</td>\n",
              "      <td>2.842958</td>\n",
              "      <td>2.843955</td>\n",
              "      <td>3.052899</td>\n",
              "      <td>2.994851</td>\n",
              "      <td>3.121371</td>\n",
              "      <td>2.888386</td>\n",
              "      <td>3.057064</td>\n",
              "      <td>3.065004</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>000BAD50D026</td>\n",
              "      <td>3.459145</td>\n",
              "      <td>3.497568</td>\n",
              "      <td>3.576564</td>\n",
              "      <td>3.321732</td>\n",
              "      <td>3.509279</td>\n",
              "      <td>3.565041</td>\n",
              "      <td>3.460105</td>\n",
              "      <td>3.498620</td>\n",
              "      <td>3.577665</td>\n",
              "      <td>3.322809</td>\n",
              "      <td>3.510040</td>\n",
              "      <td>3.566659</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>00367BB2546B</td>\n",
              "      <td>3.021687</td>\n",
              "      <td>2.954948</td>\n",
              "      <td>3.085000</td>\n",
              "      <td>2.853040</td>\n",
              "      <td>3.020295</td>\n",
              "      <td>3.025939</td>\n",
              "      <td>3.021687</td>\n",
              "      <td>2.954948</td>\n",
              "      <td>3.085000</td>\n",
              "      <td>2.853040</td>\n",
              "      <td>3.020295</td>\n",
              "      <td>3.025939</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-6771f561-b23b-4bcc-b2d4-adf1298b2e0c')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-6771f561-b23b-4bcc-b2d4-adf1298b2e0c button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-6771f561-b23b-4bcc-b2d4-adf1298b2e0c');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 68
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "for col in ['cohesion',\t'syntax',\t'vocabulary',\t'phraseology',\t'grammar',\t'conventions']:\n",
        "  merge_df['diff_' + col] = lstm_submission_new[col] - lstm_submission[\"old_\"+ col]"
      ],
      "metadata": {
        "id": "7WeUaO-HfJ8H"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "lift_dict = {}\n",
        "\n",
        "for col in ['cohesion',\t'syntax',\t'vocabulary',\t'phraseology',\t'grammar',\t'conventions']:\n",
        "  lift_dict[col] = merge_df['diff_' + col].mean()"
      ],
      "metadata": {
        "id": "NmcF3Id4fJ53"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "lift_dict"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ou698Hf9g8AU",
        "outputId": "28927752-25c4-453d-a5f7-040c3288b971"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'cohesion': 0.0683590571085612,\n",
              " 'syntax': 0.07939418156941731,\n",
              " 'vocabulary': 0.06856584548950195,\n",
              " 'phraseology': 0.06674814224243164,\n",
              " 'grammar': 0.07162229220072429,\n",
              " 'conventions': 0.07422280311584473}"
            ]
          },
          "metadata": {},
          "execution_count": 74
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rnabk-j_iZrM"
      },
      "source": [
        "### Suggestion on weak paragraph"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "5JnaXg2YiTDT"
      },
      "outputs": [],
      "source": [
        "lstm_submission_para.to_csv(\"/content/gdrive/MyDrive/Colab Notebooks/NLP/Project/feedback-prize-english-language-learning/abc.csv\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Sr7PvcY3ig_l"
      },
      "outputs": [],
      "source": []
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "machine_shape": "hm",
      "provenance": [],
      "include_colab_link": true
    },
    "gpuClass": "premium",
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}