gru.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "authorship_tag": "ABX9TyO/qLOiLaCWssrP9Hgt8TpU",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/ahadAli102/covid-sentiment-analysis/blob/main/gru.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": "1ouRQdS8rhnf"
      },
      "outputs": [],
      "source": [
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "%matplotlib inline\n",
        "import pandas as pd\n",
        "import seaborn as sns\n",
        "import re,nltk,json\n",
        "import tensorflow as tf\n",
        "from tensorflow import keras\n",
        "from tensorflow.keras import regularizers\n",
        "from tensorflow.keras.preprocessing.sequence import pad_sequences\n",
        "from keras import models\n",
        "from keras import layers\n",
        "from tensorflow.keras.layers import LSTM,GRU\n",
        "from tensorflow.keras.models import load_model\n",
        "from sklearn.metrics import confusion_matrix\n",
        "from sklearn.metrics import classification_report \n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.metrics import accuracy_score,precision_score,recall_score,f1_score,roc_auc_score\n",
        "from sklearn.metrics import average_precision_score,roc_auc_score, roc_curve, precision_recall_curve\n",
        "from sklearn.preprocessing import LabelEncoder\n",
        "from tensorflow.keras.preprocessing.text import Tokenizer\n",
        "np.random.seed(42)\n",
        "class color: # Text style\n",
        "   PURPLE = '\\033[95m'\n",
        "   CYAN = '\\033[96m'\n",
        "   DARKCYAN = '\\033[36m'\n",
        "   BLUE = '\\033[94m'\n",
        "   GREEN = '\\033[92m'\n",
        "   YELLOW = '\\033[93m'\n",
        "   RED = '\\033[91m'\n",
        "   BOLD = '\\033[1m'\n",
        "   UNDERLINE = '\\033[4m'\n",
        "   END = '\\033[0m'\n",
        "# dataset path\n",
        "path = '/content/drive/My Drive/Colab Notebooks/NLP Projects/sentiment/'"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "data = pd.read_csv('/content/data-collect - Copy.csv',encoding='utf-8')\n",
        "data = data.drop(range(0,100))\n",
        "print(f'Total number of headlines: {len(data)}')\n",
        "sns.set(font_scale=1.4)\n",
        "data['sentiment'].value_counts().plot(kind='barh', figsize=(6, 4))\n",
        "plt.xlabel(\"Number of Headlines\", labelpad=12)\n",
        "plt.ylabel(\"Sentiment\", labelpad=12)\n",
        "plt.yticks(rotation = 45)\n",
        "plt.title(\"Dataset Distribution\", y=1.02);"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 344
        },
        "id": "-LcbhuDssOl-",
        "outputId": "39b4c7a7-bd75-4041-dda5-168c24c44413"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Total number of headlines: 1110\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "data.columns"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "SCsoB0BI4_V3",
        "outputId": "81c8055a-634d-447b-efcf-eb7f91118fd0"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Index(['comments', 'sentiment'], dtype='object')"
            ]
          },
          "metadata": {},
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Cleaning Data [Remove unncessary symbols]\n",
        "def cleaning_data(row):\n",
        "      comment = re.sub('[^\\u0980-\\u09FF]',' ',str(row)) #removing unnecessary punctuation\n",
        "      return comment\n",
        "# Apply the function into the dataframe\n",
        "data['cleaned'] = data['comments'].apply(cleaning_data)  \n",
        "\n",
        "# print some cleaned reviews from the dataset\n",
        "sample_data = [250,350]\n",
        "for i in sample_data:\n",
        "  print('Original: ',data.comments[i],'\\nCleaned:',\n",
        "           data.cleaned[i],'\\n','Category:-- ',data.sentiment[i],'\\n')  "
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "e5JCQRQm5A-I",
        "outputId": "87568afe-3936-4225-b9b9-bfe5acfa5a5e"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Original:  ভারত টিকা ১টানা ১০ বছর সুরক্ষা দেয়, তাই সবাই ভারতের টিকা দিন \n",
            "Cleaned: ভারত টিকা ১টানা ১০ বছর সুরক্ষা দেয়  তাই সবাই ভারতের টিকা দিন \n",
            " Category:--  positive \n",
            "\n",
            "Original:  টিকা নিতে মন চায়না \n",
            "Cleaned: টিকা নিতে মন চায়না \n",
            " Category:--  negative \n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "data.drop(['sentiment'], axis=1)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 424
        },
        "id": "81RVnGR5KHhH",
        "outputId": "69a62512-21c8-425b-99f0-f06e4c9b52db"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                               comments  \\\n",
              "100   আমার পরিবারের সবাই নিবন্ধনের ২ মাসের মধ্যেই টি...   \n",
              "101   সরকার বিনামূল্যে টিকা দিচ্ছে অথচ এসব উজবুকরা এ...   \n",
              "102   ১ম ডোজ দিয়েছেন আজ প্রায় ৪ মাস হলো। ২য় ডোজ দেয়া...   \n",
              "103   ষাটোর্ধ্বদের মৃত্যুঝুঁকি বেশি। তাঁদের বিরাট অং...   \n",
              "104   এই টিকা নিলে এই হবে সেই হবে... পেট ব্যথায় মরে ...   \n",
              "...                                                 ...   \n",
              "1205       এদিকে টিকে নেওয়ায় শিশুর মৃত্যুর নিউজ পেলাম!    \n",
              "1206        আমি ১ম ডোজ নিয়েছি।আলহামদুলিল্লাহ সুস্থ আছি।   \n",
              "1207      টীকা নেওয়ার পরে আমার শরীরে ব্যথা দেখা দিয়েছে।   \n",
              "1208  কোনো পার্শ্ব প্রতিক্রিয়াই দেখলাম না প্রথম ডোজ ...   \n",
              "1209  সিনোফার্ম প্রথম ডোজ নেওয়ার পর দুইদিন হালকা হা...   \n",
              "\n",
              "                                                cleaned  \n",
              "100   আমার পরিবারের সবাই নিবন্ধনের ২ মাসের মধ্যেই টি...  \n",
              "101   সরকার বিনামূল্যে টিকা দিচ্ছে অথচ এসব উজবুকরা এ...  \n",
              "102   ১ম ডোজ দিয়েছেন আজ প্রায় ৪ মাস হলো  ২য় ডোজ দেয়া...  \n",
              "103   ষাটোর্ধ্বদের মৃত্যুঝুঁকি বেশি  তাঁদের বিরাট অং...  \n",
              "104   এই টিকা নিলে এই হবে সেই হবে    পেট ব্যথায় মরে ...  \n",
              "...                                                 ...  \n",
              "1205       এদিকে টিকে নেওয়ায় শিশুর মৃত্যুর নিউজ পেলাম    \n",
              "1206        আমি ১ম ডোজ নিয়েছি আলহামদুলিল্লাহ সুস্থ আছি   \n",
              "1207      টীকা নেওয়ার পরে আমার শরীরে ব্যথা দেখা দিয়েছে   \n",
              "1208  কোনো পার্শ্ব প্রতিক্রিয়াই দেখলাম না প্রথম ডোজ ...  \n",
              "1209  সিনোফার্ম প্রথম ডোজ নেওয়ার পর দুইদিন হালকা হা...  \n",
              "\n",
              "[1110 rows x 2 columns]"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-a3c4c45d-5740-4c14-868c-f78c4b1c9c0d\">\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>comments</th>\n",
              "      <th>cleaned</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>100</th>\n",
              "      <td>আমার পরিবারের সবাই নিবন্ধনের ২ মাসের মধ্যেই টি...</td>\n",
              "      <td>আমার পরিবারের সবাই নিবন্ধনের ২ মাসের মধ্যেই টি...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>101</th>\n",
              "      <td>সরকার বিনামূল্যে টিকা দিচ্ছে অথচ এসব উজবুকরা এ...</td>\n",
              "      <td>সরকার বিনামূল্যে টিকা দিচ্ছে অথচ এসব উজবুকরা এ...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>102</th>\n",
              "      <td>১ম ডোজ দিয়েছেন আজ প্রায় ৪ মাস হলো। ২য় ডোজ দেয়া...</td>\n",
              "      <td>১ম ডোজ দিয়েছেন আজ প্রায় ৪ মাস হলো  ২য় ডোজ দেয়া...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>103</th>\n",
              "      <td>ষাটোর্ধ্বদের মৃত্যুঝুঁকি বেশি। তাঁদের বিরাট অং...</td>\n",
              "      <td>ষাটোর্ধ্বদের মৃত্যুঝুঁকি বেশি  তাঁদের বিরাট অং...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>104</th>\n",
              "      <td>এই টিকা নিলে এই হবে সেই হবে... পেট ব্যথায় মরে ...</td>\n",
              "      <td>এই টিকা নিলে এই হবে সেই হবে    পেট ব্যথায় মরে ...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1205</th>\n",
              "      <td>এদিকে টিকে নেওয়ায় শিশুর মৃত্যুর নিউজ পেলাম!</td>\n",
              "      <td>এদিকে টিকে নেওয়ায় শিশুর মৃত্যুর নিউজ পেলাম</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1206</th>\n",
              "      <td>আমি ১ম ডোজ নিয়েছি।আলহামদুলিল্লাহ সুস্থ আছি।</td>\n",
              "      <td>আমি ১ম ডোজ নিয়েছি আলহামদুলিল্লাহ সুস্থ আছি</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1207</th>\n",
              "      <td>টীকা নেওয়ার পরে আমার শরীরে ব্যথা দেখা দিয়েছে।</td>\n",
              "      <td>টীকা নেওয়ার পরে আমার শরীরে ব্যথা দেখা দিয়েছে</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1208</th>\n",
              "      <td>কোনো পার্শ্ব প্রতিক্রিয়াই দেখলাম না প্রথম ডোজ ...</td>\n",
              "      <td>কোনো পার্শ্ব প্রতিক্রিয়াই দেখলাম না প্রথম ডোজ ...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1209</th>\n",
              "      <td>সিনোফার্ম প্রথম ডোজ নেওয়ার পর দুইদিন হালকা হা...</td>\n",
              "      <td>সিনোফার্ম প্রথম ডোজ নেওয়ার পর দুইদিন হালকা হা...</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>1110 rows × 2 columns</p>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-a3c4c45d-5740-4c14-868c-f78c4b1c9c0d')\"\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-a3c4c45d-5740-4c14-868c-f78c4b1c9c0d 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-a3c4c45d-5740-4c14-868c-f78c4b1c9c0d');\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": 6
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def data_summary(dataset):\n",
        "    \n",
        "    \"\"\"\n",
        "    This function will print the summary of the headlines and words distribution in the dataset. \n",
        "    \n",
        "    Args:\n",
        "        dataset: list of cleaned sentences   \n",
        "        \n",
        "    Returns:\n",
        "        Number of documnets per class: int \n",
        "        Number of words per class: int\n",
        "        Number of unique words per class: int\n",
        "    \"\"\"\n",
        "    documents = []\n",
        "    words = []\n",
        "    u_words = []\n",
        "    total_u_words = [word.strip().lower() for t in list(dataset.cleaned) for word in t.strip().split()]\n",
        "    class_label= [k for k,v in dataset.sentiment.value_counts().to_dict().items()]\n",
        "  # find word list\n",
        "    for label in class_label: \n",
        "        word_list = [word.strip().lower() for t in list(dataset[dataset.sentiment==label].cleaned) for word in t.strip().split()]\n",
        "        counts = dict()\n",
        "        for word in word_list:\n",
        "                counts[word] = counts.get(word, 0)+1\n",
        "        # sort the dictionary of word list  \n",
        "        ordered = sorted(counts.items(), key= lambda item: item[1],reverse = True)\n",
        "        # Documents per class\n",
        "        documents.append(len(list(dataset[dataset.sentiment==label].cleaned)))\n",
        "        # Total Word per class\n",
        "        words.append(len(word_list))\n",
        "        # Unique words per class \n",
        "        u_words.append(len(np.unique(word_list)))\n",
        "       \n",
        "        print(\"\\nClass Name : \",label)\n",
        "        print(\"Number of Documents:{}\".format(len(list(dataset[dataset.sentiment==label].cleaned))))  \n",
        "        print(\"Number of Words:{}\".format(len(word_list))) \n",
        "        print(\"Number of Unique Words:{}\".format(len(np.unique(word_list)))) \n",
        "        print(\"Most Frequent Words:\\n\")\n",
        "        for k,v in ordered[:10]:\n",
        "              print(\"{}\\t{}\".format(k,v))\n",
        "    print(\"Total Number of Unique Words:{}\".format(len(np.unique(total_u_words))))           \n",
        "   \n",
        "    return documents,words,u_words,class_label\n",
        "\n",
        "#call the fucntion\n",
        "documents,words,u_words,class_names = data_summary(data)   "
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ccxWGFCv5YRV",
        "outputId": "164183f9-af0a-4652-f4ca-e4512778b034"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "Class Name :  negative\n",
            "Number of Documents:601\n",
            "Number of Words:5843\n",
            "Number of Unique Words:1728\n",
            "Most Frequent Words:\n",
            "\n",
            "টিকা\t530\n",
            "না\t150\n",
            "করোনা\t114\n",
            "নেওয়ার\t74\n",
            "টিকার\t68\n",
            "পর\t65\n",
            "আক্রান্ত\t61\n",
            "এই\t56\n",
            "আমি\t44\n",
            "মারা\t42\n",
            "\n",
            "Class Name :  positive\n",
            "Number of Documents:394\n",
            "Number of Words:4189\n",
            "Number of Unique Words:1390\n",
            "Most Frequent Words:\n",
            "\n",
            "টিকা\t363\n",
            "করোনা\t73\n",
            "হবে\t61\n",
            "না\t45\n",
            "আমি\t39\n",
            "টিকার\t37\n",
            "এই\t34\n",
            "কোন\t31\n",
            "নিলে\t29\n",
            "আমার\t28\n",
            "\n",
            "Class Name :  neutral\n",
            "Number of Documents:115\n",
            "Number of Words:1272\n",
            "Number of Unique Words:683\n",
            "Most Frequent Words:\n",
            "\n",
            "টিকা\t97\n",
            "না\t30\n",
            "করোনা\t23\n",
            "কি\t16\n",
            "টিকার\t14\n",
            "আর\t11\n",
            "করে\t10\n",
            "নিয়ে\t9\n",
            "নাকি\t9\n",
            "তা\t9\n",
            "Total Number of Unique Words:2798\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "data_matrix = pd.DataFrame({'Total Documents':documents,\n",
        "                            'Total Words':words,\n",
        "                            'Unique Words':u_words,\n",
        "                            'Class Names':class_names})\n",
        "df = pd.melt(data_matrix, id_vars=\"Class Names\", var_name=\"sentiment\", value_name=\"Values\")\n",
        "plt.figure(figsize=(8, 6))\n",
        "ax = plt.subplot()\n",
        "\n",
        "sns.barplot(data=df,x='Class Names', y='Values' ,hue='sentiment')\n",
        "ax.set_xlabel('Class Names') \n",
        "ax.set_title('Data Statistics')\n",
        "\n",
        "ax.xaxis.set_ticklabels(class_names, rotation=45);"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 467
        },
        "id": "4zO0MqKF6QaG",
        "outputId": "97334716-d5dc-4ae3-c8de-d14b02bd91c4"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 576x432 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Calculate the Review of each of the Review\n",
        "data['comments_length'] = data.cleaned.apply(lambda x:len(x.split()))\n",
        "frequency = dict()\n",
        "for i in data.comments_length:\n",
        "    frequency[i] = frequency.get(i, 0)+1\n",
        "\n",
        "plt.bar(frequency.keys(), frequency.values(), color =\"b\")\n",
        "plt.xlim(1, 20)\n",
        "# in this notbook color is not working but it should work.\n",
        "plt.xlabel('Length of the Headlines')\n",
        "plt.ylabel('Frequency')\n",
        "plt.title('Length-Frequency Distribution')\n",
        "plt.show()  \n",
        "print(f\"Maximum Length of a headline: {max(data.comments_length)}\")\n",
        "print(f\"Minimum Length of a headline: {min(data.comments_length)}\")\n",
        "print(f\"Average Length of a headline: {round(np.mean(data.comments_length),0)}\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 367
        },
        "id": "-uraPBNd6Wmn",
        "outputId": "46405228-a244-4497-faa3-ba377a3c7f1c"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAaMAAAEqCAYAAAC4Kj/kAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3de1xNaf8//tcunUSlA0bE3WFHJ+d0GBo5RlFh3CjGcRD5OJbj/XXKYW7HchriQSORmga3uGeMO2McbzHORoWIMSmdRKnW7w+/vW97dtGuXXuX1/Px8HjY17rWWu91aL/3uq5rrSUSBEEAERGRCmmoOgAiIiImIyIiUjkmIyIiUjkmIyIiUjkmIyIiUjkmIyIiUjkmI6p1oaGhcHR0VHUYVIMCAwMRGBhYK+sKDQ2Fp6enTJmtrS2WLFlSK+t/8uQJbG1tER8fXyvrq6+YjOqo+Ph42Nra4tq1a6oOpVzPnz9HeHg47ty5UyPL9/T0hK2tbbn/srOza2Sdn6rQ0FCZ/duxY0f06tULwcHBOHnyJMrKypSynpSUFISHh+PJkydKWZ6y7d+/nwmnBjVQdQBUP/3555+IiIiAubk52rVrVyPrsLW1xfjx4+XKGzVqVCPr+5RpaWlh5cqVAICioiJkZGTg9OnTCA4OhrOzM7Zt2yaz3yMjIxVeR0pKCiIiIuDs7IyWLVtWer7ly5ejNu7dP3DgAJo0aQJ/f3+ZcnNzc1y/fh0NGvDrtDq496jOMjMzw+DBgytd//Xr19DT06vBiOovkUgkt69nzpyJb7/9FuvWrcOiRYuwceNG6TRtbe0aj0lyPLW0tGp8XR8iEomgo6Oj0hjqAzbT1XN//vknFi5cCHd3dzg4OMDLywvR0dEydS5evAhbW1scO3YM27dvR48ePeDo6IgxY8bg0aNHcsvcv38/evXqBScnJwwdOhSXL1+W6SO4ePEihg4dCgCYP3++tHknPDxcZjnPnz/H1KlT0bFjR7i4uGDNmjUoLS1VynYHBgaif//+uHPnDgIDA9GhQwcsXboUACAIAqKiouDj4wNHR0e4urpiwYIFcs17giBg69at6NGjB9q3b4/AwEDcv38fnp6eCA0NldYLDw+Hra2tXAySptS/NjudPXsWAQEB6NixIzp27Ijx48fLNWdK+tUqs48EQcD+/fsxePBgODk5oVu3bhg7diz++9//AgD+/ve/w8fHp9z95Ofnh2HDhlVyr8qbNGkSPv/8c5w4cQIPHjyQlpfXZ3T8+HEMGTIEnTp1QseOHeHl5YUtW7YAeLevZsyYAQAYPXq09JyRNIt96HiW12f0/joHDBgAR0dHDBo0CGfOnJGZXtlj5+npifv37+PSpUvS2CTrrKjP6O7du5g4cSI6deqEDh06IDAwUHpM/rqeS5cuYdWqVXBxcUGHDh0QFBT0yTU388qoHsvKysLw4cNRWlqKESNGwMTEBOfPn8fSpUuRk5ODqVOnytTftWsXNDQ0MG7cOBQUFGDXrl2YM2cOYmNjpXWio6OxbNkydO7cGV999RWePn2KoKAgGBoaonnz5gAAKysrBAcHY/PmzRg+fDg6d+4MADJ/9IIgYMKECXB0dMS8efNw/vx57N69G61atcLIkSMrtX0lJSVyf7B6enrSq5/8/HyMHz8effv2hbe3Nxo3bgwA+Mc//oG4uDj4+vpi1KhRePbsGb777jvcuHEDhw8flv7K3bRpE7Zt2wYPDw94eHjgzp07GD9+PIqLixU5DDKOHj2KuXPnwt3dHbNmzUJxcTEOHTqEkSNH4vDhw7CyslJ4Hy1evBixsbH4/PPP4efnBwBITk7G5cuX0aVLF/j5+WHJkiW4e/cu2rZtK50vNTUVt2/frnZH/6BBg3D27FmcO3cOf/vb38qtc+7cOcyaNQsuLi6YPXs2NDQ08ODBAyQnJwMAunbtisDAQERFRWHy5MmwtLQEAHTq1Em6jIqOZ0WuXr2KEydOIDAwEPr6+jh48CCmTJmCvXv3okuXLgpt44IFC7B8+XI0bNgQkydPBgDo6+tXWD81NRUjR45Ew4YNMX78eOjo6CA2NhZfffUV9uzZg65du8rUDwsLg6GhIaZNm4aMjAzs3bsXy5Ytk7narPcEqpPi4uIEsVgsXL16tcI6ixYtEtzc3ISsrCyZ8oULFwpOTk5Cbm6uIAiCcOHCBUEsFgv9+/cXioqKpPX27t0riMVi4d69e4IgCEJRUZHg7Ows+Pr6CsXFxdJ68fHxglgsFgICAqRl169fF8RisRAXFycXV0hIiCAWi4Xw8HCZcl9fX8HPz69S29+zZ09BLBbL/Vu9erUgCIIQEBAgiMViYe/evTLzXblyRRCLxcL3338vU3758mVBLBYLMTExgiAIQlZWlmBvby9MmjRJKCsrk9bbuHGjIBaLhZCQEGnZ5s2bBbFYLBej5Bg9fvxYEARBePXqldC1a1chNDRUpl5OTo7g4uIizJo1S+F9JDl2/+///T+59Uvizs3NFRwdHYU1a9bITF+3bp1gb28vd378VUhIiODg4FDh9Nu3bwtisVgICwuTlgUEBMicDytXrhQ6deoklJSUVLicxMREQSwWCxcuXJCbVtHxlMTXs2dPmTLJ+ZCcnCwty87OFrp27Sr8/e9/l5ZV9tgJgiAMHDhQZpskHj9+LHeuBwUFCfb29sKDBw+kZVlZWYKzs7PM8ZOsZ8yYMTLnWVhYmNCuXTshLy9Pbn31FZvp6ilBEHDy5El4eHgAALKzs6X/3N3d8ebNG/z2228y8/j6+sq09Ut+PT5+/BgAcPPmTeTk5GDYsGEy7fQ+Pj4wNDRUOMYvv/xS5nPnzp0VGknl4OCAPXv2yPwbPny4dHqDBg1kPgNAYmIiGjZsiO7du8vsE0tLS5iamuLixYsA3v2Sf/v2LUaOHAmRSCSdvzrDlc+dO4fc3Fz4+PjIrLu0tBRdunSRrvt9H9tHJ0+eBABMnz5dbl5J3AYGBvD09MSxY8ekI98EQcCxY8fQvXt3GBsbV3mbAKBhw4YAgFevXlVYp3Hjxnj9+jXOnj1b5fWUdzw/xMHBAR07dpR+btKkCby9vZGcnIzc3Nwqx/ExpaWlOHv2LHr27Ik2bdpIy42NjeHv749bt27hxYsXMvMMHTpU5jzr0qULSktLkZGRUWNxqhs209VT2dnZyM3NRVxcHOLi4sqtk5WVJfP5s88+k/lsYGAAAMjLywMAPH36FADQunVrmXoNGjSAubm5QvFpaWmhadOmMmWGhoYyXxI5OTl4+/at9LOurq5M04yRkRHc3NwqXEfTpk3lOpYfPnyIwsLCCueT7JOKttXY2LhKiReAtE9l7Nix5U7X0JD9bViZfZSeng5TU9OPJhQ/Pz8kJibi4sWLcHV1xX//+19kZGRg3rx5VdkUGYWFhQA+3Gw1cuRInDhxApMmTULTpk3h5uaGvn37wtPTU+ZL+EPKO54f8n4i+GvZ06dPq3wcPyY7OxuvX78ut8lS0vyYkZEBU1NTafnH/vY+BUxG9ZTkF7C3tzeGDBlSbh1ra2uZz5qamuXWE2pg2GxlvoCmT5+OS5cuST/7+flh9erVlV5HeV9cZWVlMDIywoYNG8qdR/IloIiKtqW8gQYAsHr1ajRr1qzKy62Kzz//HKampjhy5AhcXV1x5MgR6RVTdf3+++8AAAsLiwrrmJiY4Pvvv8f58+dx5swZ/PLLL0hISEDPnj2xbdu2Sm1rTYxYq+yxq2m1+benrpiM6iljY2Po6+ujpKTkg1cPimjRogUA4NGjR3B3d5eWl5SUICMjQ2aAgjK+SENCQmR+Gf71KqEqLCwscO7cObRv3/6Dv+Tf39b3f2FLrjjf9/6v2PeTmeTqSqJVq1YA3h0bZR0TCwsL/PLLL8jOzv7g1ZGmpiZ8fHxw6NAhzJ8/HydPnkT//v2VMgT7yJEjEIlEMudEebS1taWDQQRBwLp167Bz504kJyejc+fOSk2+wLur4IrKJMe3sscOqPw5bWxsDD09PZnRhRJpaWkAoHBLwqeAfUb1lKamJvr164dTp07h7t27ctOrMmzUwcEBRkZGiI2NlWk+O3r0qNwXtGREW3WaGRwcHODm5ib999cruaoYMGAAysrKpEOK31daWirdDjc3N2hpaSE6Olrm12lUVJTcfJIrgsuXL0vLCgsLkZCQIFOve/fuMDAwwPbt28sdkVeVY9KvXz8AkBs2D8j/qvb19cWrV6+wZMkS5ObmKnSPVkW+/fZbnD17FgMGDCi3WUzi5cuXMp9FIhHs7OwAvBslByjnnHnfzZs3cfXqVZkYjh07ho4dO0qb6Cp77CTxVaavSVNTE927d8fp06eRnp4uLc/JyUFCQgIcHBxkmujoHV4Z1XHx8fE4d+6cXPmIESMwZ84cXLp0CcOHD8ewYcNgY2OD3Nxc3L17Fz/++CNu3Lih0Lq0tbUxffp0LF++HGPGjIGXlxeePn2K+Ph4uSYaCwsLGBoa4sCBA2jYsCH09fVhY2MDsVhcre2trq5du2LUqFGIjIzEvXv30L17d2hpaSE9PR0nT55EcHAw/P39YWxsjHHjxmHHjh34+uuv4eHhgbt37yIpKQlNmjSRWaa7uztatGiBhQsXIi0tDZqamoiLi0OTJk1kfmE3atQIS5cuxZw5c+Dn54eBAwfC1NQUT58+xS+//AIbGxuFmiEBoFu3bvD390d0dDTS09PRo0cPAMC1a9dga2srHYYMAG3btkXbtm2RmJiIli1bSofcV4YgCPjhhx8AAMXFxcjIyMDPP/+Me/fuoVu3bli2bNkH51+0aBFevnwJV1dXNG/eHM+fP8f+/fthZmYmHShjZ2cHTU1N7NixA3l5edDV1YWTk5P0ilJRYrEYkydPRkBAgHRo96tXrzB79mxpncoeO+Ddj6Po6GhERESgTZs2aNiwYYXNnP/3f/+Hs2fPYuTIkRg5ciR0dHRw6NAh5Ofny9yjRv/DZFTHHTx4sNzyfv36wcrKCrGxsdi6dStOnTqFmJgYGBoawtLSssp/EAEBARAEAXv27MHatWvRtm1bbNu2DStWrJBp09fS0sLatWuxbt06LFu2DG/fvsW0adNUnowAYMmSJbCzs0NMTAw2bNgATU1NtGjRAl5eXnBxcZHW+7//+z9oa2sjJiYGFy9ehJOTEyIjI/H111/LLE9LSwsRERFYunQpNm3aBDMzM4wZMwYGBgaYP3++TN0BAwagadOm2L59O/bs2YOioiI0bdoUnTp1Umik2PtWrlwJW1tbxMbG4ptvvoG+vj7s7e3l7mUB3vW7rVq1CoMGDVKoWezt27fSwQ56enowNjaGg4MDgoKC0KdPH7nBF381aNAgxMbGIiYmBrm5uTA1NYWHhweCgoKkjxEyNTXF8uXLsWPHDixevBilpaVYtWpVlZOR5Ebh8PBwPH78GH/729+wZcsWmf2iyLELCgrCs2fPsGfPHhQUFMDc3LzCZGRlZYUDBw5ImyIFQYCDgwNWrFhR7nEhQCR8Sj1kVCPKysrg6uqKPn36YMWKFaoOp8Z5enrC2dlZ4asYdbB//34sW7YMJ06cqPAGVSJVYJ8RKaSoqEiuLyIhIQE5OTlwdnZWUVRUWbGxsejYsSMTEakdNtORQq5du4ZVq1ahf//+MDIywu3bt3H48GGIxWL0799f1eFROQoLC/Hzzz/j0qVLuHPnDjZv3qzqkIjkMBmRQszNzdG8eXNERUUhNzcXhoaGGDx4MObMmVMrT2omxWVnZ2P27NkwMDDAhAkTpCPwiNQJ+4yIiEjl2GdEREQqx2REREQqxz6jcrx8+QplZWy9fJ+JSSNkZRWoOox6jfu4ZnH/1hwNDRGaNKn48VqVwWRUjrIygcmoHNwnNY/7uGZx/6ovNtMREZHKMRkREZHKMRkREZHKMRkREZHKMRkREZHKMRkREZHKMRkREZHK8T4jolrS2EAPujof/pMzM2tcbvmbohLk572uibCI1AKTEVEt0dVpAJ/ZP1Rp3qPrBiNfyfEQqRM20xERkcqp9Mro0aNHiIyMxG+//Yb79+/D0tISx44dk04vLS3F7t27kZSUhJSUFJSWlkIsFmPatGlwdXWVW15kZCT279+PFy9ewNraGnPnzi23HhERqReVXhndv38fSUlJaN26NaysrOSmv3nzBjt27EDbtm2xatUqrF+/Hs2aNcPYsWNx+vRpmbqRkZHYsGEDRo0ahR07dqBNmzaYNGkS7t69W1ubQ0REVaTSl+uVlZVBQ+NdPgwNDcXNmzflrowKCgpgaGgoLRMEAUOGDIG+vj6ioqIAAMXFxXBzc8OXX36JefPmSef18fGBjY0NNm3apFBcWVkFfKDiX5iZNUZmJnstqsPMrHG1+oy4/6uH53DN0dAQwcSkUfWWoaRYqrZyjQ+vXlNTUyYRAYBIJELbtm3x559/SsuSk5ORn5+PgQMHyszr5eWFM2fOgC+zJSJSb3VuAENZWRmuXr0q06yXmpoKAHJNfdbW1igsLMTz589rNUYiIlJMnRvaHRUVhQcPHmD58uXSsry8PGhra0NXV1emruSqKicnB82bN6/0Oqp7uVlfVXQPDNUO7v/q4z5UX3UqGV26dAnffPMNxo0bhy5dutTYethnJI/t7dVX3S9C7v/q4Tlcc+p8n5Ei7t69i6lTp6J3796YO3euzDQDAwMUFxejqKhIpjw3NxcAYGRkVGtxEhGR4upEMkpPT8eECRNgZ2eHtWvXQiQSyUyX9BVJ+o4kUlNToa+vj2bNmtVarEREpDi1T0aZmZkYN24cTE1NsXXrVmhra8vV6dSpExo3bozjx49Ly0pLS5GYmIju3bvLJS8iIlIvKu0zev36NZKSkgAAGRkZKCgowIkTJwAAjo6OMDExwYQJE5CVlYXQ0FCkpKTIzN+hQwcAgLa2NqZMmYINGzbA2NgYdnZ2iI2NRXp6OtatW1e7G0VERApTaTLKysrCjBkzZMokn1etWgVnZ2fpExSCgoLk5r937570/+PHjwfwbrTdixcvYGNjg2+//RZt27atqfCJiEhJVJqMWrZsKZNQyvOx6e8bP368NCkREVHdofZ9RkREVP8xGRERkcoxGRERkcoxGRERkcoxGRERkcoxGRERkcoxGRERkcoxGRERkcoxGRERkcrVqfcZ0aelsYEedHUUP0XfFJUgP+91DURERDWFyYjUlq5OA/jM/kHh+Y6uGwy+Qo2obmEzHRERqRyvjKheY1MfUd3AZET1Gpv6iOoGNtMREZHKMRkREZHKMRkREZHKMRkREZHKMRkREZHKMRkREZHKqTQZPXr0CEuWLMHgwYNhZ2cHb2/vcuslJSXBz88Pjo6O6N27N6KiosqtFxkZCU9PTzg5OcHf3x/nz5+vyfCJiEhJVJqM7t+/j6SkJLRu3RpWVlbl1rl69SqmTp2Kdu3aYefOnfD390dYWBgOHDggUy8yMhIbNmzAqFGjsGPHDrRp0waTJk3C3bt3a2NTiIioGlR606unpyd69+4NAAgNDcXNmzfl6mzZsgV2dnYICwsDALi4uODZs2fYsmULhg8fDg0NDRQXF2Pbtm0YPXo0xo8fDwBwdnaGj48Ptm3bhk2bNtXeRhERkcJUemWkofHh1RcXF+PChQsYMGCATLm3tzcyMzNx69YtAEBycjLy8/MxcOBAaR1NTU14eXnhzJkzEARB+cETEZHSqPUAhvT0dLx9+1auCc/GxgYAkJaWBgBITU0FALl61tbWKCwsxPPnz2shWiIiqiq1Tka5ubkAAAMDA5lyyWfJ9Ly8PGhra0NXV1emnqGhIQAgJyenpkMlIqJq4INSy2Fi0kjVIaglM7PGqg6h0pQRq7ptr7rFUxdxH6ovtU5GkiubvLw8mXLJZ8l0AwMDFBcXo6ioCDo6OtJ6kisnIyMjhdablVWAsjL2M73PzKwxMjNr9znW1fnikMSqjGUoS3W/CGt7/9c3qjiHPxUaGqJq/4hX62Y6CwsLaGlpSfuGJFJSUgAAlpaWAP7XVyTpO5JITU2Fvr4+mjVrVgvREhFRVal1MtLW1oaLiwsSExNlyo8dOwYzMzPY29sDADp16oTGjRvj+PHj0jqlpaVITExE9+7dIRKJajVuIiJSjEqb6V6/fo2kpCQAQEZGBgoKCnDixAkAgKOjI8zNzREUFISAgAAsWrQIPj4+SE5ORmxsLJYsWSIdGq6trY0pU6Zgw4YNMDY2hp2dHWJjY5Geno5169apbPuIiKhyVJqMsrKyMGPGDJkyyedVq1bB398fHTt2xNatW7F+/XokJCSgadOmmD9/PkaMGCEzn+Rm16ioKLx48QI2Njb49ttv0bZt29rZGCIiqjKVJqOWLVvi3r17H63n4eEBDw+Pj9YbP368NCkREVHdodZ9RkRE9GlgMiIiIpVjMiIiIpVT65teqW5qbKAHXR3FT603RSXIz3tdAxERkbpjMiKl09VpAJ/ZPyg839F1g6Gu98czwRLVLCYjokqojwmWSJ2wz4iIiFSOyYiIiFSOyYiIiFSOyYiIiFRO4WT0yy+/QBD4rh8iIlIehUfTTZw4EWZmZvD29sbgwYP5IFIiIqo2ha+MtmzZgk6dOiE6Ohp+fn7w8fFBZGQknj9/XhPxERHRJ0DhK6NevXqhV69e0ncPHT16FOvWrcP69evRrVs3DB48GH379oWenl5NxEtERPVQlQcwNGrUCEOHDsXevXtx+vRpzJw5E1lZWQgNDYWbmxvmzZuH8+fPKzNWIiKqp5Qymq60tBQlJSUoLi6GIAjQ1dXFuXPnMHbsWPj6+uL3339XxmqIiKieqvLjgPLz85GYmIgjR44gOTkZmpqa+OKLLzB79mx88cUX0NDQwM8//4ywsDDMnz8fcXFxyoybiIjqEYWT0U8//YQjR44gKSkJRUVFcHR0xMKFCzFw4EAYGRnJ1O3duzdevnyJpUuXKi1gIiKqfxRORtOmTUOzZs0wevRo+Pr6wsrK6oP1bW1t4ePjU+UAiYio/lM4Ge3evRuurq4QiUSVqu/k5AQnJyeFAyMiok+HwgMY3NzcKp2IlOWnn37C0KFD0bFjR7i7u2P69Ol4+PChXL2EhAT0798fjo6OGDhwII4fP16rcRIRUdUonIzCwsLQt2/fCqf369cPa9asqVZQ7zt//jymTZsGS0tLREREYNGiRUhLS8PYsWNRUFAgrXfixAmEhISgT58+2LlzJ1xdXTFr1iwkJSUpLRYiIqoZCjfTJSUlwcvLq8LpXl5e0sSgDMeOHUOLFi2wZs0a6RWZubk5hg0bhitXrsDDwwMAsGnTJvTv3x+zZ88GALi4uCAtLQ3h4eHSOkREpJ4UvjJ69uwZWrZsWeF0c3Nz/PHHH9UK6n0lJSXQ19eXaRps3LixTJ3Hjx8jLS0NAwcOlCn39vbGjRs3kJ2drbR4iIhI+RRORo0aNcKTJ08qnP748WPo6OhUK6j3+fn5IS0tDVFRUcjLy8OTJ0+wZs0aWFlZwdXVFQCQlpYGAHIj+6ytrWWmExGRelI4Gbm4uCAmJgZPnz6Vm/bkyRMcPHgQLi4uSglOsr7w8HBs2LABXbt2Ra9evZCRkYE9e/ZAW1sbAJCbmwsAMDAwkJnX0NBQZjoREaknhfuMgoODcebMGXh7e8Pf3x82NjYAgPv37yM+Ph6ampqYMWOG0gJMTk5GSEgIhg4dCk9PT+Tk5GDr1q2YMmUKoqOjoaurq7R1SZiYNFL6MusDM7PGH6+kJutQxnLUKRZlLudTxn2ovhRORm3atMGBAwewbNkyfPfddzLTunbtisWLF8PS0lJpAa5YsQLdunXDggULpGUdOnTAF198gR9++AHDhw+XXgHl5eXBzMxMWk9yRSSZXllZWQUoK+MLBN9nZtYYmZn5la5bVe+vQxnLqS+x/DUeUpwi5zApRkNDVO0f8VV6Np2NjQ2ioqKQnZ0t7T9q1aoVmjRpUq1gypOamgpPT0+ZsubNm6NJkyZIT08HAGnyS0tLk+k3Sk1NlZlORETqqcoPSgUAY2NjGBsbKyuWcrVo0QK3bt2SKcvIyMDLly9hbm4O4F0itLS0xPHjx9GnTx9pvWPHjsHR0bHGYySqTY0N9KCro/if7puiEuTnva6BiIiqr0rJqLS0FGfPnsXjx4+Rm5sLQZBt0hKJRAgKClJKgKNGjcLy5cuxfPly9OrVCzk5Odi2bRtMTExk7ncKDg7GzJkzYWFhATc3N5w6dQq//vorduzYoZQ4iNSFrk4D+Mz+QeH5jq4bDDZSkbpSOBnduHEDwcHB+OOPP+SSkISyk5GWlhaio6MRHx8PfX19tG/fHhs3bpRpFvTy8sKbN2+wfft2REZGwsLCAuvWreMNr0REdYDCyWjp0qV48+YNtmzZgi5dusgNp1Y2kUiE4cOHY/jw4R+t6+fnBz8/vxqNh4iIlE/hZHTv3j3MnDlTblABERFRVSl802vz5s0rbJ4jIiKqCoWT0aRJk3Do0CGZJ2YTERFVh8LNdDk5OWjYsCH69OmDfv364bPPPoOGhmxOE4lEmDBhgtKCJCKi+k3hZLRu3Trp/2NiYsqtw2RERESKUDgZnTp1qibiICKiT5jCyUjy1AMiIiJlqfLjgB4+fIhLly4hKysLPj4+aNmyJYqLi/HixQuYmppKX+9ARET0MQono7KyMvzjH//A4cOHIQgCRCIROnTogJYtW+Lt27fw8fFBUFAQxo0bVxPxEhFRPaTw0O7t27cjLi4OM2bMwMGDB2XuOdLX10ffvn3x73//W6lBEhFR/aZwMoqPj8eQIUMwefJkWFhYyE23tbXFo0ePlBIcERF9GhRORn/88QecnJwqnK6jo4NXr15VKygiIvq0KJyMzMzMkJGRUeH0W7duoUWLFtUKioiIPi0KD2Do27cvDhw4AF9fX+nrvEUiEQAgKSkJCQkJmDhxonKjpFrxsZe2VfTabL60jYiqS+FkNH36dFy6dAl+fn7o1KkTRCIRduzYgfXr1+PGjRuwt7fH119/XROxUg3jS9s+LXxjLKkThc/ERo0aISYmBnv27MGJEyego6ODK1euwMLCAtOmTcOECROgo6NTE7ESkRLxxwepkyrd9Kqjo4PJkydj8uTJyo6HiISjTOEAAB/+SURBVIg+QQoPYCAiIlI2ha+M5s+f/9E6IpEIYWFhVQqIiOoO9juRsih8Fl28eFGurKysDJmZmSgtLYWxsTH09PSUEhwRqTf2O5GyKJyMfv7553LL3759i4MHD2Lv3r3YvXt3tQP7q4SEBOzbtw8pKSnQ09ODnZ0d1q1bB2NjYwDvhpVv3LgRKSkpaNasGcaMGYPAwEClx0FERMqntD4jLS0tBAQEwN3dHcuXL1fWYgEA27Ztw9KlS9GnTx/s3LkTK1euhLW1Nd6+fQsAuHr1KqZOnYp27dph586d8Pf3R1hYGA4cOKDUOIiIqGZU+RUSFWnbti1++EHxy/aKpKWlISIiAhEREejZs6e0vHfv3tL/b9myBXZ2dtJ+KhcXFzx79gxbtmzB8OHD5V6LTkRE6kXp39Lnzp1Tap9RfHw8WrRoIZOI3ldcXIwLFy5gwIABMuXe3t7IzMzErVu3lBYLERHVDIWvjCIiIsotz8/Px+XLl3H79m1MmjSp2oFJ/Pbbb7C1tcXWrVuxf/9+5OTkoF27dpg3bx6cnZ2Rnp6Ot2/fwsrKSmY+GxsbAO+urBwdHZUWDxERKZ/SkpGhoSFatWqFpUuX4ssvv6x2YBKZmZm4efMm7t69i4ULF6JRo0bYvXs3JkyYgOPHjyM3NxcAYGBgIDOf5LNkOhERqS+Fk9Hdu3drIo4KCYKAwsJCREdHo127dgCArl27olevXoiMjIS3t7fS12li0kjpy6zvKnqIal1ejjrFoqzl1MdY1H2dVDlKH8CgbAYGBjAyMpImIgDQ09ND+/btcf/+femTw/Py8mTmk3yWTFdEVlYBysqEj1esZ6rzh5qZ+b+7RtRpOfUlFmUtR5226f1YaoOZWeNaX+enQkNDVO0f8Qono6dPn1ZpRVV9x5G1tTXS09PLnVZUVAQLCwtoaWkhLS0NPXr0kE5LSUkBAFhaWlZpvUREVHsUTkaenp7S9xcp4s6dOwrPAwA9e/ZEfHw8bt26BXt7ewBAYWEhrl27hn79+kFbWxsuLi5ITEzEV199JZ3v2LFjMDMzk85DRETqS+FktGLFCkRFReGPP/6At7c32rRpAwB48OAB/vWvf+Gzzz5DYGCg0u7t6d27N5ycnBAcHIyZM2dCX18fu3fvxps3bzB27FgAQFBQEAICArBo0SL4+PggOTkZsbGxWLJkCe8xIiKqAxRORpmZmSguLsa///1vuf6Y6dOnY8SIEXjx4oXSXrCnoaGBHTt2YO3atVi6dCmKiorQvn177Nu3D61btwYAdOzYEVu3bsX69euRkJCApk2bYv78+RgxYoRSYiAiopqlcDKKiYnB2LFjyx0Y0KRJE3z55ZfYt2+fUt/2amxsjNWrV3+wjoeHBzw8PJS2TiIiqj0KJ6OcnBy8evWqwumFhYXIycmpVlBE9GnhqyhI4aMvaSLz8PCAg4ODzLQbN24gKioKTk5OSguQiOo/voqCFE5GS5YsQWBgIIYNGwZHR0dpv82jR49w48YNGBoaYvHixUoPlIiI6i+Fh5pZW1vj6NGjCAwMRF5eHk6ePImTJ08iLy8Po0ePxtGjR6XPhSMiIqqMKj2BwdTUFAsWLMCCBQuUHQ8REX2CqnUTzsOHD3HlyhXk57PVloiIqq5Kyejo0aP44osv4OXlhYCAANy8eRMAkJ2djX79+uH48eNKDZKIiOo3hZPRyZMnMXfuXFhZWWHevHkQhP89UNTY2BiWlpZKfdMrERHVfwono+3bt8PNzQ2RkZHw9fWVm96+fftaf80EERHVbQoPYEhNTUVoaGiF001MTJCdnV2toEhxvGmQiOoyhb+99PT0UFhYWOH09PR0NGnSpFpBkeJ40yAR1WUKN9O5uLjg+++/R3Fxsdy058+f49ChQ/j888+VEhwREX0aFL4ymjlzJr788ksMGTIE/fv3h0gkwpkzZ/Drr78iNjYWGhoamDZtWk3ESkRE9ZTCV0Zt2rTBgQMHYGZmhoiICAiCgD179mDXrl1o27YtoqOjq/xWVyIi+jQpdGVUWlqK58+fw8TEBLt370Zubi4ePXoEQRDQqlUrGBsb11ScREQf9bGBPGZmjcst50Ae1VMoGZWVlaFPnz6YM2eO9J1GfEI3EakLDuSpuxRqptPS0kLTpk0hEolqKh4iIvoEKdxn5O/vj/j4eBQVFdVEPERE9AlSeDRdmzZtUFZWBi8vL/j6+qJVq1bQ0dGRqzdgwAClBEhERPWfwslo7ty50v9v3bq13DoikYjJiIiIKq1SyWjFihXw9fWFg4MD9u3bBwAoLCyEjo4ONDU1azTAv3r16hW8vLzw/PlzHD58GI6OjtJpCQkJ2L59OzIyMmBhYYGgoCAmRSKiOqBSyei7775D+/bt4eDgAGdnZ7x8+RJubm7YvXs3nJ2dazpGGRERESgtLZUrP3HiBEJCQjBp0iS4u7vjp59+wqxZs6Cvrw8PD49ajZGIiBRT5Zfrvf/qiNry+++/IyYmBsHBwXLTNm3ahP79+2P27NlwcXHBokWL4ObmhvDw8FqPk4iIFFOtN73WtmXLlmHUqFFo06aNTPnjx4+RlpaGgQMHypR7e3vjxo0bfIo4EZGaqzPJKCEhAY8ePcKUKVPkpqWlpQEArKysZMqtra1lphMRkXqq9Gi6x48f4/r16wCA/Px39yqnpaVBX1+/3PrKfDJDfn4+vvnmG4SEhJS7vtzcXACAgYGBTLmhoaHMdCIiUk+VTkbh4eFy/S8rVqyQqycIAkQiEe7cuVP96P5/GzduROvWrTFo0CClLfNDTEwa1cp61EVFz+uq7WWo23LUKRZlLYex1PxyqGoqlYxWrVpV03FU6P79+4iJicHu3buRl5cHANKX+xUWFqKgoEB6BZSXlwczMzPpvJIrIsn0ysrKKkBZWe0P0KiO6vwhZWbmK20Z6rac+hKLspajTtukrucMKU5DQ1TtH/GVSkZ+fn7VWkl1PHr0CCUlJRg9erTctNGjR6Nt27aIiIgA8K7Z8P1+o9TUVACApaVl7QRLRERVovATGGpbp06dpDfaSty5cwerVq3C0qVLYW9vj1atWsHS0hLHjx9Hnz59pPWOHTsGR0dHvtqCiEjNqX0yMjY2Rrdu3cqdZm9vL30CQ3BwMGbOnAkLCwu4ubnh1KlT+PXXX7Fjx47aDJeIiKpA7ZNRZXl5eeHNmzfYvn07IiMjYWFhgXXr1vHpC0REdUCdTEbdunXDvXv35Mr9/PxU2r9FRERVUyeTERFRTfnYq8srwleXVw+TERHRe/jqctWoM48DIiKi+ovJiIiIVI7JiIiIVI7JiIiIVI7JiIiIVI7JiIiIVI7JiIiIVI7JiIiIVI7JiIiIVI7JiIiIVI7JiIiIVI7JiIiIVI7JiIiIVI7JiIiIVI7JiIiIVI7JiIiIVI7JiIiIVI7JiIiIVI6vHSciqgGNDfSgq6P4V+ybohLk572ugYjUm9ono8TERBw9ehS3bt1Cbm4uWrVqhREjRuDvf/87NDT+d2GXlJSEjRs3IiUlBc2aNcOYMWMQGBiowsiJ6FOmq9MAPrN/UHi+o+sGI78G4lF3ap+M9uzZgxYtWmDevHkwMTHBxYsXsXLlSjx+/BghISEAgKtXr2Lq1KkYPHgwQkJCkJycjLCwMDRo0AAjRoxQ8RYQEdHHqH0y2r59O4yNjaWfXVxcUFhYiP3792PmzJnQ1tbGli1bYGdnh7CwMGmdZ8+eYcuWLRg+fLjMFRQREakftf+Wfj8RSbRr1w5FRUXIyclBcXExLly4gAEDBsjU8fb2RmZmJm7dulVboRIRURWpfTIqz5UrV2BkZAQTExOkp6fj7du3sLKykqljY2MDAEhLS1NFiEREpAC1b6b7qxs3biA+Ph5BQUHQ1NREbm4uAMDAwECmnuSzZLoiTEwaVT/QOsTMrLFaLEPdlqNOsShrOYylZpejTrHUNXUqGWVmZiI4OBiOjo6YOHFija0nK6sAZWVCjS2/JlTn5M3MzFfaMtRtOfUlFmUtR522iedM5ZZTF2hoiKr9I77ONNPl5+dj4sSJ0NXVxbZt26ClpQUAMDQ0BADk5eXJ1Jd8lkwnIiL1VSeSUVFREaZMmYKsrCzs2rULTZo0kU6zsLCAlpaWXN9QSkoKAMDS0rJWYyUiIsWpfTIqKSnBjBkzcO/ePezcuRPm5uYy07W1teHi4oLExESZ8mPHjsHMzAz29va1GS4REVWB2vcZLVu2DKdPn8bcuXPx5s0bXLt2TTrN2toajRo1QlBQEAICArBo0SL4+PggOTkZsbGxWLJkCe8xIiKqA9Q+GZ09exYA8M0338hN27dvH7p164aOHTti69atWL9+PRISEtC0aVPMnz+fT18gIqoj1D4Z/fzzz5Wq5+HhAQ8PjxqOhoiIaoLaJ6P6jk/2JSJiMlI5PtmXiKgOjKYjIqL6j8mIiIhUjs10RERqqqp9ykDN9CtXJ56PYTIiIlJTVe1TBmT7lZU1UKqieJo20UPkor5VilOCyYiIqJ6rCwOl2GdEREQqx2REREQqx2REREQqx2REREQqx2REREQqx2REREQqx2REREQqx2REREQqx5teq4GvfyAiUg4mo2qoC3c1ExHVBWymIyIilWMyIiIilWMyIiIilatXyejhw4cYP348OnbsCBcXFyxfvhyvX3OgABGRuqs3Axjy8vIwevRotGjRAps2bUJ2djZWrVqF7OxsbNiwQdXhERHRB9SbZBQTE4O8vDwkJCTA2NgYAKCpqYk5c+Zg6tSpsLGxUXGERERUkXrTTHfmzBm4uLhIExEA9OvXD9ra2jhz5owKIyMioo+pN1dGqampGDJkiEyZtrY2LCwskJaWptCyNDREla7btImeQssubx3KWEZ9jEVZy6kPsShrOeq0TTxnajYWZS2nMttkYqhbpWW/TyQIglDtpagBe3t7zJgxA5MmTZIpHzFiBExMTBAREaGiyIiI6GPqTTMdERHVXfUmGRkYGCAvL0+uPC8vD4aGhiqIiIiIKqveJCMrKyukpqbKlBUXFyM9PR2WlpYqioqIiCqj3iSjHj164MKFC3j58qW07Mcff0RxcTE8PDxUGBkREX1MvRnAkJeXB29vb5ibm2Pq1KnIysrC6tWr4erqypteiYjUXL1JRgDw4MEDrFixAleuXIGOjg4GDhyIuXPnQk+v6kMjiYio5tWrZERERHVTvekzIiKiuovJiIiIVI7JiCoUHx8PW1tbuX/Lli1TdWh1zqNHj7BkyRIMHjwYdnZ28Pb2LrdeUlIS/Pz84OjoiN69eyMqKqqWI627KrOPQ0NDyz2nT5w4oYKI647ExERMnToVHh4e6NChA3x8fBAdHY2ysjKZetU5f+vNs+mo5uzatQuNGzeWfjY1NVVhNHXT/fv3kZSUhPbt26OsrAzlddVevXoVU6dOxeDBgxESEoLk5GSEhYWhQYMGGDFihAqirlsqs48BoFWrVvjnP/8pU9amTZtaiLDu2rNnD1q0aIF58+bBxMQEFy9exMqVK/H48WOEhIQAUML5KxBVIC4uThCLxUJWVpaqQ6nzSktLpf8PCQkRBg4cKFdn/PjxwtChQ2XKFi1aJLi7u8vMT+WrzD6uqJw+rLzvgLCwMMHR0VEoKioSBKH65y+b6YhqgYbGh//UiouLceHCBQwYMECm3NvbG5mZmbh161ZNhlcvfGwfU9W9/2oeiXbt2qGoqAg5OTlKOX959OijfHx80K5dO3h6eiIiIgIlJSWqDqneSU9Px9u3b2FlZSVTLnkppKKvQaGKpaeno0uXLrC3t4evry+OHz+u6pDqpCtXrsDIyAgmJiZKOX/ZZ0QVMjMzw/Tp0+Hk5ARNTU2cOXMGW7duxZMnT7B69WpVh1ev5ObmAnj3wN/3ST5LplP1tGvXDo6OjrC2tkZ+fj4OHz6MmTNn4s2bN/D391d1eHXGjRs3EB8fj6CgIGhqairl/GUyogp1794d3bt3l352d3dH48aNER4ejqlTp8LCwkKF0REpbsyYMTKfe/fujdGjRyM8PJzJqJIyMzMRHBwMR0dHTJw4UWnLZTMdKcTLywsA2IehZJLXnPz1NSiSz3wNSs3p378/nj59iuzsbFWHovby8/MxceJE6OrqYtu2bdDS0gKgnPOXyYhIDVhYWEBLS0uubT0lJQUA+BoUUrmioiJMmTIFWVlZ2LVrF5o0aSKdpozzl8mIFPKvf/0LIpEIDg4Oqg6lXtHW1oaLiwsSExNlyo8dOwYzMzPY29urKLL6TRAEJCYmwtzcvNwRY/ROSUkJZsyYgXv37mHnzp0wNzeXma6M85d9RlSh8ePHo1u3bhCLxRCJRPjll18QHR2NoUOHolWrVqoOr055/fo1kpKSAAAZGRkoKCiQ3vXv6OgIc3NzBAUFISAgAIsWLYKPjw+Sk5MRGxuLJUuWcNhyJXxsHwPvnsAwcOBAtG7dGnl5eYiNjcWlS5ewdu1alcVdFyxbtgynT5/G3Llz8ebNG1y7dk06zdraGo0aNar2+cundlOFVq5ciTNnzuD58+coKSlBmzZt4O/vjzFjxkBTU1PV4dUpT548Qa9evcqdtmrVKmnneVJSEtavX4/U1FQ0bdoUX331FUaPHl2bodZZH9vHnp6emD9/Pm7fvo2srCxoaWnBzs4O48ePh6enZy1HW7d4enoiIyOj3Gn79u1Dt27dAFTv/GUyIiIileO1PxERqRyTERERqRyTERERqRyTERERqRyTERERqRyTERERqRyTEdVroaGh0hsea9r58+fh5+cHJycn2Nrayj2nqzICAwPRv3//Goiu7ggNDZW778fW1hbh4eHSzxcvXoStrS0uXrxY2+FRDWEyoo+Kj4+Hra2tzF3X6uT58+cIDw/HnTt3VBZDQUEBZsyYAZFIhMWLF2Pt2rXQ09Mrt25KSgrCw8Px5MmTWo7yf8LDw2Fra4vMzMxyp3t7eyMwMLCWo6JPGR8HRHXen3/+iYiICJibm6Ndu3YqieHu3bvIzc1FWFgYevfu/cG6KSkpiIiIgLOzM1q2bFlLEdYvXbt2xfXr16VPjaa6j8mISAkkrx9o3LixiiP5NGhoaEBHR0fVYZASsZmOlObPP//EwoUL4e7uDgcHB3h5eSE6OlqmjqSt/9ixY9i+fTt69OgBR0dHjBkzBo8ePZJb5v79+9GrVy84OTlh6NChuHz5MgIDA6VNSBcvXsTQoUMBAPPnz4etra1c/wLwrilv6tSp6NixI1xcXLBmzRqUlpZWartOnjwJf39/ODk5oVu3bpg1axaePn0qnR4YGIjp06cDAEaPHg1bW1uEhoaWu6z4+HjMmDFDpq6trS3i4+Nl6qWkpGD06NFo3749unfvjp07d8otq7i4GBEREejbty8cHBzQvXt3hIWF4fXr15XaLkUJgoCoqCj4+PjA0dERrq6uWLBggdx7gE6dOoXJkyejR48ecHBwQM+ePbFmzRoUFRXJLfOnn36Ct7c3HB0d4e3tjR9//LFSsZTXZyTpH6zMsa7stty6dQsTJ06Ei4sLHB0d0bNnT+nDQkm5eGVESpGVlYXhw4ejtLQUI0aMgImJCc6fP4+lS5ciJycHU6dOlam/a9cuaGhoYNy4cSgoKMCuXbswZ84cxMbGSutER0dj2bJl6Ny5M7766is8ffoUQUFBMDQ0RPPmzQEAVlZWCA4OxubNmzF8+HB07twZwLsObwlBEDBhwgQ4Ojpi3rx5OH/+PHbv3o1WrVph5MiRH9yuH374AfPmzYO9vT1mzZqFly9fYt++fbhy5Qq+//57GBsbY/LkybC1tUVUVBQmT54MS0vLCt+C27VrVwQGBsrUBYBOnTpJ6xQUFGDixIno3bs3+vfvj5MnT+Kf//wnxGIxPDw8pNsUFBSEy5cvY9iwYbC2tkZqaiqio6ORkpKCyMhIiESijx633Nzcch96W16i/sc//oG4uDj4+vpi1KhRePbsGb777jvcuHEDhw8fll6pxMfHQ1tbG4GBgWjcuDF+++037N27F3/88Qc2bNggXd7Zs2cxffp0WFlZYdasWcjNzcWCBQvQrFmzj8Zdkcoe68psS3Z2NsaNGwcjIyNMmDABhoaGePbsGX7++WcUFhZCV1e3ynFSOQSij4iLixPEYrFw9erVCussWrRIcHNzE7KysmTKFy5cKDg5OQm5ubmCIAjChQsXBLFYLPTv318oKiqS1tu7d68gFouFe/fuCYIgCEVFRYKzs7Pg6+srFBcXS+vFx8cLYrFYCAgIkJZdv35dEIvFQlxcnFxcISEhglgsFsLDw2XKfX19BT8/vw9ud3FxseDm5iZ4eXkJr1+/lpZLtmH16tXSssTEREEsFgsXLlz44DI/VjcgIEAQi8XC999/Ly0rKioS3N3dhenTp0vLjhw5Itja2goXL16Umf+HH34QxGKx8Msvv3wwhs2bNwtisfiD/97fx1euXJGLSxAE4fLly4JYLBZiYmKkZYWFhXLr27Ztm2Brays8ffpUWjZ48GDB3d1dyMvLk5adO3dOEIvFQs+ePWXmF4vFwubNm6WfJcfg/X1Y2WNd2W358ccfBbFYLFy/fr2cPUjKxmY6qjZBEHDy5Enpr/bs7GzpP3d3d7x58wa//fabzDy+vr7Q1taWfu7SpQsA4PHjxwCAmzdvIicnB8OGDZPppPbx8anSK7i//PJLmc+dO3f+6Gi2mzdv4sWLFxgxYoTMr+Bu3brB3t4e//nPfxSOozJ0dXUxaNAg6WdtbW04OjpK9w0AJCYmok2bNrC2tpbZ387OzhCJRJUe8rxx40bs2bNH7t9fX56WmJiIhg0bonv37jLrs7S0hKmpqcz6JKMIy8rKkJ+fj+zsbHTu3BmCIEhfV//nn3/izp07GDRokEw/m6urK2xsbBTfae/52LGu7LZI4vrPf/6Dt2/fVism+jg201G1ZWdnIzc3F3FxcYiLiyu3TlZWlsznzz77TOazgYEBAEjvzZH0ybRu3VqmXoMGDeS+KD9GS0sLTZs2lSkzNDREbm7uB+eTxPC3v/1NbpqVlRVOnjypUByV1axZM7mXkRkaGuLevXvSzw8fPsSDBw/g6upa7jL+ur8r0qVLF5iZmcmVN2zYUObzw4cPUVhYCDc3t4+u7/fff8c333yDS5cuyfWtFBQUAPjfvm3Tpo3cstq0aYPbt29XKv6/qsyxruy2ODs7o1+/foiIiMCePXvg7OwMT09PeHt7y+0fqj4mI6q2srIyAO/uTRkyZEi5daytrWU+V/RyPqEGXq9Vmb4TdVKZt2KWlZXB2toaCxcuLHf6X7+Qq6usrAxGRkYyfT7vk/yYyM/Px+jRo9GwYUPMnDkTFhYW0NXVxfPnzxEaGio9V2pKZY51ZbdFJBJh8+bNuH79Ok6fPo1z585h8eLF2LFjBw4dOgQTExOlxv6pYzKiajM2Noa+vj5KSkoq/LWpqBYtWgAAHj16BHd3d2l5SUkJMjIyZAYo1FSykcTw4MEDfP755zLT0tLSFL5Ck1BGvBYWFrh16xZcXV1rJdlaWFjg3LlzaN++PfT19Susd/HiRbx8+RKbN2+Gs7OztPzXX3+VqSfZtw8fPpRbRnllylTZbZFwcnKCk5MTZsyYgaSkJEyaNAmHDh3ClClTajTOTw37jKjaNDU10a9fP5w6dQp3796Vm/7X4bKV4eDgACMjI8TGxsq01x89elSueU3SR1GVx+98LAZTU1McPHhQZljyf//7X9y8eRNffPFFlZarjHgHDBiAFy9eyA2dB94N+ZY0hynLgAEDUFZWhi1btshNKy0tlR4TyVXd+1e4ZWVl2LNnj8w8TZs2Rbt27XDkyBHk5+dLy8+fP4/79+8rNfa/quy25Obmyl2p29vbA1D+uUa8MiIFxMfH49y5c3LlI0aMwJw5c3Dp0iUMHz4cw4YNg42NDXJzc3H37l38+OOPuHHjhkLr0tbWxvTp07F8+XKMGTMGXl5eePr0KeLj4+WGTVtYWMDQ0BAHDhxAw4YNoa+vDxsbG4jF4mptr5aWFubOnYuQkBCMHDkSgwYNQnZ2NqKiotCsWTNMnDixSsu1s7ODpqYmduzYgby8POjq6sLJyQmtWrWq9DIGDRqEEydOYNmyZbh8+bJ0gMCDBw+QmJiITZs2oVu3blWKrzxdu3bFqFGjEBkZiXv37qF79+7Q0tJCeno6Tp48ieDgYPj7+6NTp04wMjJCaGgoAgIC0KBBA5w8eRKFhYVyy5w1axa+/vprjBw5Ev7+/sjLy8N3330HGxubcuvX9rZ8//33iI6ORu/evWFhYYE3b94gPj4empqan/zzA2sCkxFV2sGDB8st79evH6ysrBAbG4utW7fi1KlTiImJgaGhISwtLSu8AfRjAgICIAgC9uzZg7Vr16Jt27bYtm0bVqxYIXP3vZaWFtauXYt169Zh2bJlePv2LaZNm1btZAS8G/Wnp6eHHTt24J///Cf09PTg4eGBOXPmwNjYuErLNDU1xfLly7Fjxw4sXrwYpaWlWLVqlULJSENDAxEREdi7dy8SEhJw6tQp6OrqSu+neb8ZU1mWLFkCOzs7xMTEYMOGDdDU1ESLFi3g5eUFFxcXAICRkRG+/fZbrF69GuHh4WjYsCH69u2LESNGyIwQBIAePXpg06ZN2LhxI9avXw8LCwuEhYXh1KlTuHTpktLjV3RbnJ2dcePGDZw4cQKZmZlo1KgR7OzssHjxYrRv375G4/sUiYSa6DEmqiFlZWVwdXVFnz59sGLFClWHQ0RKwj4jUltFRUVybfYJCQnIycmR6RwnorqPzXSktq5du4ZVq1ahf//+MDIywu3bt3H48GGIxWK22RPVM0xGpLbMzc3RvHlzREVFITc3F4aGhhg8eDDmzJkj8/QGIqr72GdEREQqxz4jIiJSOSYjIiJSOSYjIiJSOSYjIiJSOSYjIiJSOSYjIiJSuf8Phghm4KLDVjUAAAAASUVORK5CYII=\n"
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Maximum Length of a headline: 34\n",
            "Minimum Length of a headline: 2\n",
            "Average Length of a headline: 10.0\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def label_encoding(sentiment,bool):\n",
        "    \"\"\"\n",
        "    This function will return the encoded labels in array format. \n",
        "    \n",
        "    Args:\n",
        "        sentiment: series of class names(str)\n",
        "        bool: boolean (True or False)\n",
        "        \n",
        "    Returns:\n",
        "        labels: numpy array \n",
        "    \"\"\"\n",
        "    le = LabelEncoder()\n",
        "    le.fit(sentiment)\n",
        "    encoded_labels = le.transform(sentiment)\n",
        "    labels = np.array(encoded_labels) # Converting into numpy array\n",
        "    class_names =le.classes_ ## Define the class names again\n",
        "    if bool == True:\n",
        "        print(\"\\n\\t\\t\\t===== Label Encoding =====\",\"\\nClass Names:-->\",le.classes_)\n",
        "        for i in sample_data:\n",
        "            print(sentiment[i],' ', encoded_labels[i],'\\n')\n",
        "\n",
        "    return labels"
      ],
      "metadata": {
        "id": "jHFYjynp7Mcm"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def dataset_split(headlines,category):\n",
        "    \"\"\"\n",
        "    This function will return the splitted (90%-10%-10%) feature vector . \n",
        "    \n",
        "    Args:\n",
        "        headlines: sequenced headlines \n",
        "        category: encoded lables (array) \n",
        "        \n",
        "    Returns:\n",
        "        X_train: training data \n",
        "        X_valid: validation data\n",
        "        X_test : testing feature vector \n",
        "        y_train: training encoded labels (array) \n",
        "        y_valid: training encoded labels (array) \n",
        "        y_test : testing encoded labels (array) \n",
        "    \"\"\"\n",
        "    X,X_test,y,y_test = train_test_split(headlines,category,train_size = 0.9,\n",
        "                                                  test_size = 0.1,random_state =0)\n",
        "    X_train,X_valid,y_train,y_valid = train_test_split(X,y,train_size = 0.8,\n",
        "                                                  test_size = 0.2,random_state =0)\n",
        "    print(color.BOLD+\"\\nDataset Distribution:\\n\"+color.END)\n",
        "    print(\"\\tSet Name\",\"\\t\\tSize\")\n",
        "    print(\"\\t========\\t\\t======\")\n",
        "\n",
        "    print(\"\\tFull\\t\\t\\t\",len(headlines),\n",
        "        \"\\n\\tTraining\\t\\t\",len(X_train),\n",
        "        \"\\n\\tTest\\t\\t\\t\",len(X_test),\n",
        "        \"\\n\\tValidation\\t\\t\",len(X_valid))\n",
        "  \n",
        "    return X_train,X_valid,X_test,y_train,y_valid,y_test\n"
      ],
      "metadata": {
        "id": "iB-oJ6e47xyY"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "labels = label_encoding(data.sentiment,True)"
      ],
      "metadata": {
        "id": "tCr-wAoF740o",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "1ee26244-c3fb-4311-947c-8906ce3f8e4f"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "\t\t\t===== Label Encoding ===== \n",
            "Class Names:--> ['negative' 'neutral' 'positive']\n",
            "positive   0 \n",
            "\n",
            "negative   0 \n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "X_train,X_valid,X_test,y_train,y_valid,y_test = dataset_split(data.comments,labels)"
      ],
      "metadata": {
        "id": "ttbtcB2j7-IP",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "b068e0fe-a18f-43ad-fe0a-5ffddd73e9c2"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m\n",
            "Dataset Distribution:\n",
            "\u001b[0m\n",
            "\tSet Name \t\tSize\n",
            "\t========\t\t======\n",
            "\tFull\t\t\t 1110 \n",
            "\tTraining\t\t 799 \n",
            "\tTest\t\t\t 111 \n",
            "\tValidation\t\t 200\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "vocab_size = 57000\n",
        "embedding_dim = 64\n",
        "max_length = 21\n",
        "trunc_type='post'\n",
        "padding_type='post'\n",
        "oov_tok = \"\"\n",
        "\n",
        "def padded_comment(original,encoded,padded):\n",
        "  '''\n",
        "  print the samples padded comment\n",
        "  '''\n",
        "  print(color.BOLD+\"\\n\\t\\t\\t====== Encoded Sequences ======\"+color.END,\"\\n\")  \n",
        "  print(original,\"\\n\",encoded) \n",
        "  print(color.BOLD+\"\\n\\t\\t\\t====== Paded Sequences ======\\n\"+color.END,original,\"\\n\",padded)  "
      ],
      "metadata": {
        "id": "NnWCF72N8H_o"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Train Data Tokenization\n",
        "tokenizer = Tokenizer(num_words = vocab_size, oov_token=oov_tok)\n",
        "tokenizer.fit_on_texts(X_train)\n",
        "word_index = tokenizer.word_index\n",
        "train_sequences = tokenizer.texts_to_sequences(X_train)\n",
        "train_padded = pad_sequences(train_sequences, padding=padding_type, maxlen=max_length)"
      ],
      "metadata": {
        "id": "kzYg6-G28Msd"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        " #============================== Tokenizer Info =================================\n",
        "(word_counts,word_docs,word_index,document_count) = (tokenizer.word_counts,\n",
        "                                                       tokenizer.word_docs,\n",
        "                                                       tokenizer.word_index,\n",
        "                                                       tokenizer.document_count)\n",
        "def tokenizer_info(mylist,bool):\n",
        "  ordered = sorted(mylist.items(), key= lambda item: item[1],reverse = bool)\n",
        "  for w,c in ordered[:10]:\n",
        "    print(w,\"\\t\",c)\n",
        "  #=============================== Print all the information =========================\n",
        "print(color.BOLD+\"\\t\\t\\t====== Tokenizer Info ======\"+color.END)   \n",
        "print(\"Words --> Counts:\")\n",
        "tokenizer_info(word_counts,bool =True )\n",
        "print(\"\\nWords --> Documents:\")\n",
        "tokenizer_info(word_docs,bool =True )\n",
        "print(\"\\nWords --> Index:\")\n",
        "tokenizer_info(word_index,bool =True )    \n",
        "print(\"\\nTotal Documents -->\",document_count)\n",
        "print(f\"Found {len(word_index)} unique tokens\")"
      ],
      "metadata": {
        "id": "FTdM8dAn8P_L",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "5e8c052a-3b6b-4194-d736-bee4e92ed949"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m\t\t\t====== Tokenizer Info ======\u001b[0m\n",
            "Words --> Counts:\n",
            "টিকা \t 694\n",
            "করোনা \t 144\n",
            "না \t 135\n",
            "টিকার \t 81\n",
            "পর \t 74\n",
            "নেওয়ার \t 69\n",
            "এই \t 68\n",
            "আমি \t 65\n",
            "হবে \t 61\n",
            "আক্রান্ত \t 56\n",
            "\n",
            "Words --> Documents:\n",
            "টিকা \t 588\n",
            "করোনা \t 140\n",
            "না \t 120\n",
            "টিকার \t 79\n",
            "পর \t 73\n",
            "নেওয়ার \t 65\n",
            "এই \t 64\n",
            "আমি \t 60\n",
            "আক্রান্ত \t 54\n",
            "হবে \t 53\n",
            "\n",
            "Words --> Index:\n",
            "এসেই \t 2400\n",
            "দিব \t 2399\n",
            "দিয়েই \t 2398\n",
            "আমিতো \t 2397\n",
            "জনসন \t 2396\n",
            "আমেরিকার \t 2395\n",
            "রাশিয়ার \t 2394\n",
            "উঠিয়ে \t 2393\n",
            "উছিলায় \t 2392\n",
            "হয়ত \t 2391\n",
            "\n",
            "Total Documents --> 799\n",
            "Found 2400 unique tokens\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "X_train"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "oMt74KlCMeim",
        "outputId": "242284c1-c6a2-4ac6-d7b9-fb3203b36f27"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "436     ছয় মাস পরে নাকি টিকার কার্যকারিতা কমে যায় তা...\n",
              "988     মানুষের হায়াত /রিযিক যতদিন আছে ততদিন ই বাঁচে। ...\n",
              "194     টিকা আসলে সেটার ব্যবস্থাপনাও ত্রাণের মত হয়ে যাবে।\n",
              "607     প্রথম ডোজ ফাইজার আর দ্বিতীয় ডোজ এস্ট্রাজেনিকা/...\n",
              "104     এই টিকা নিলে এই হবে সেই হবে... পেট ব্যথায় মরে ...\n",
              "                              ...                        \n",
              "1025                            আমিতো ভ্যাক্সিন নিবো ই না\n",
              "1175                      করোনা ভ্যাকসিন ওইটা দিয়েই দিব!!\n",
              "966                               টিকার উপর কোনো ভরসা নেই\n",
              "1083                                  টিকা না নিলে সমস্যা\n",
              "1128             টিকে নিতে এসেই করোনার সংক্রমণ নিয়ে যাবে!\n",
              "Name: comments, Length: 799, dtype: object"
            ]
          },
          "metadata": {},
          "execution_count": 17
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# for /content/Sentimentdata - Copy.csv\n",
        "index = 684\n",
        "#padded_comment(X_train[index],train_sequences[index],train_padded[index]) "
      ],
      "metadata": {
        "id": "A1iX6J0X8ULP"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "X_valid"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZMbKz81CJs-U",
        "outputId": "bb196ece-99da-4d97-f92b-1bf963b10a1d"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "904     টিকা নিলে যে করোনা হবে না তাতো না।টিকা তো করোন...\n",
              "1039                               আম জনতার টিকা লাগবে না\n",
              "375              টিকা নামে শরীরে জোর করে কী ঢুকিয়ে দিচ্ছে\n",
              "736            করোনা টিকার মূল চাবিকাঠি কাঁকড়ার নীল রক্ত\n",
              "223     জনগণকে টিকা দিতে হলে সেনাবাহিনী নামিয়ে দিতে বল...\n",
              "                              ...                        \n",
              "710     অল্প টিকা এগুলো প্রথমে সিনিয়র রাজনৈতিক নেতাদের...\n",
              "765     আমার ছেলেকে করোনা টিকা দেয়ালাম আমার অফিসের কয়ে...\n",
              "147     চিনা টিকার কার্যকারিতা কম।শেষে\\nদেশে দেখা যাবে...\n",
              "172     টিকা আনবে beximco সেই টিকা ১০০ গুন বেশি দাম দি...\n",
              "209              কারখানা হই অথবা গোডাউন তবুও টিকা নিব না।\n",
              "Name: comments, Length: 200, dtype: object"
            ]
          },
          "metadata": {},
          "execution_count": 19
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Validation Data Tokenization for /content/Sentimentdata - Copy.csv\n",
        "index = 172\n",
        "validation_sequences = tokenizer.texts_to_sequences(X_valid)\n",
        "validation_padded = pad_sequences(validation_sequences, padding=padding_type , maxlen=max_length)\n",
        "#padded_comment(X_valid[index],validation_sequences[index],validation_padded[index]) "
      ],
      "metadata": {
        "id": "C0nIidri8YJ5"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "X_test"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "1GgmCyqBM4Yk",
        "outputId": "9320f8da-1497-42d6-fcfe-b8d6f0bb6fd5"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "713     প্রধানমন্ত্রী সহ সকল মন্ত্রীদের প্রকাশ্যে টিকা...\n",
              "1084                        টিকা নিয়েও বা কি বাল হচ্ছে? \n",
              "1195                         সিনোফার্মেরটাই বেশি কার্যকর।\n",
              "594     টিকার সাইডএফেক্টে অনেকে মারা যাচ্ছে, ফ্রান্স, ...\n",
              "841     আগে করোনার সকল ধরনের বেরিয়ে নিক , তারপর আমি ফ...\n",
              "                              ...                        \n",
              "543                 সারা পৃথিবী ভাপছে এই টিকার সফলতা নিয়ে\n",
              "786     এই করোনা টিকা সম্পূর্ণ নিরাপদ।যদি কারো মৃত্যু ...\n",
              "525     টিকা নিয়ে নিজে সুস্থ থাকি বাকি সব কিছু আল্লাহর...\n",
              "932     স্যালাইনের পানি টিকা হিসাবে ব্যবহার করলে এমন হবেই\n",
              "368     প্রতিটি টীকা সফল হউক।।বৈশ্বিক এই মহামারীতে তড়ি...\n",
              "Name: comments, Length: 111, dtype: object"
            ]
          },
          "metadata": {},
          "execution_count": 21
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Test Data Tokenization\n",
        "index = 70\n",
        "test_sequences = tokenizer.texts_to_sequences(X_test)\n",
        "test_padded = pad_sequences(test_sequences, padding=padding_type , maxlen=max_length)\n",
        "#padded_comment(X_test[index],test_sequences[index],test_padded[index]) "
      ],
      "metadata": {
        "id": "BdXdKaSAJaTc"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "train_label_seq = y_train\n",
        "valid_label_seq = y_valid\n",
        "testing_label_seq = y_test"
      ],
      "metadata": {
        "id": "JnwLCU2j_0sa"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "keras.backend.clear_session()\n",
        "accuracy_threshold = 1.0\n",
        "vocab_size = 57000\n",
        "embedding_dim = 64\n",
        "max_length = 21\n",
        "num_category = 6\n",
        "\n",
        "class myCallback(keras.callbacks.Callback):\n",
        "  def on_epoch_end(self, epoch, logs={}):\n",
        "      if(logs.get('accuracy')>accuracy_threshold):\n",
        "        print(\"\\nReached %2.2f%% accuracy so we will stop trianing\" % (accuracy_threshold*100))\n",
        "        self.model.stop_training = True\n",
        "\n",
        "acc_callback = myCallback()\n",
        "# Saved the Best Model\n",
        "filepath = path+\"Model.h5\"\n",
        "checkpoint = keras.callbacks.ModelCheckpoint(filepath, monitor='val_accuracy', verbose=2, save_best_only=True, \n",
        "                                             save_weights_only=False, mode='max')\n",
        "callback_list = [acc_callback, checkpoint] \n",
        "model = tf.keras.Sequential([\n",
        "    tf.keras.layers.Embedding(vocab_size, embedding_dim, input_length=max_length),\n",
        "    tf.keras.layers.Bidirectional(GRU(64,dropout=0.2)),\n",
        "    tf.keras.layers.Dense(24, activation='relu'),\n",
        "    tf.keras.layers.Flatten(),\n",
        "    tf.keras.layers.Dense(num_category, activation='softmax')\n",
        "])\n",
        "model.compile(loss='sparse_categorical_crossentropy',optimizer='adam',metrics=['accuracy'])\n",
        "model.summary()"
      ],
      "metadata": {
        "id": "E4v_OjEy_5J0",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "a0539e90-a8cb-483c-a6cc-f068dc37dcda"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " embedding (Embedding)       (None, 21, 64)            3648000   \n",
            "                                                                 \n",
            " bidirectional (Bidirectiona  (None, 128)              49920     \n",
            " l)                                                              \n",
            "                                                                 \n",
            " dense (Dense)               (None, 24)                3096      \n",
            "                                                                 \n",
            " flatten (Flatten)           (None, 24)                0         \n",
            "                                                                 \n",
            " dense_1 (Dense)             (None, 6)                 150       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 3,701,166\n",
            "Trainable params: 3,701,166\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "num_epochs = 10\n",
        "batch = 64\n",
        "history = model.fit(train_padded, train_label_seq, \n",
        "                    epochs=num_epochs,\n",
        "                    batch_size = batch,\n",
        "                    validation_data=(validation_padded, valid_label_seq), \n",
        "                    verbose=1,\n",
        "                    callbacks = callback_list)"
      ],
      "metadata": {
        "id": "dDfj9qGAACa8",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "8aa8eeee-7ea0-495f-de14-3ff7bc872af6"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/10\n",
            "13/13 [==============================] - ETA: 0s - loss: 1.7213 - accuracy: 0.5332\n",
            "Epoch 1: val_accuracy improved from -inf to 0.50000, saving model to /content/drive/My Drive/Colab Notebooks/NLP Projects/sentiment/Model.h5\n",
            "13/13 [==============================] - 11s 148ms/step - loss: 1.7213 - accuracy: 0.5332 - val_loss: 1.6145 - val_accuracy: 0.5000\n",
            "Epoch 2/10\n",
            "13/13 [==============================] - ETA: 0s - loss: 1.4237 - accuracy: 0.5544\n",
            "Epoch 2: val_accuracy did not improve from 0.50000\n",
            "13/13 [==============================] - 1s 68ms/step - loss: 1.4237 - accuracy: 0.5544 - val_loss: 1.2768 - val_accuracy: 0.5000\n",
            "Epoch 3/10\n",
            "13/13 [==============================] - ETA: 0s - loss: 1.0905 - accuracy: 0.4869\n",
            "Epoch 3: val_accuracy improved from 0.50000 to 0.50500, saving model to /content/drive/My Drive/Colab Notebooks/NLP Projects/sentiment/Model.h5\n",
            "13/13 [==============================] - 1s 78ms/step - loss: 1.0905 - accuracy: 0.4869 - val_loss: 1.0460 - val_accuracy: 0.5050\n",
            "Epoch 4/10\n",
            "13/13 [==============================] - ETA: 0s - loss: 0.9849 - accuracy: 0.5557\n",
            "Epoch 4: val_accuracy did not improve from 0.50500\n",
            "13/13 [==============================] - 1s 69ms/step - loss: 0.9849 - accuracy: 0.5557 - val_loss: 1.0187 - val_accuracy: 0.5000\n",
            "Epoch 5/10\n",
            "13/13 [==============================] - ETA: 0s - loss: 0.8842 - accuracy: 0.5594\n",
            "Epoch 5: val_accuracy improved from 0.50500 to 0.54000, saving model to /content/drive/My Drive/Colab Notebooks/NLP Projects/sentiment/Model.h5\n",
            "13/13 [==============================] - 1s 79ms/step - loss: 0.8842 - accuracy: 0.5594 - val_loss: 0.9265 - val_accuracy: 0.5400\n",
            "Epoch 6/10\n",
            "13/13 [==============================] - ETA: 0s - loss: 0.7234 - accuracy: 0.7121\n",
            "Epoch 6: val_accuracy improved from 0.54000 to 0.64500, saving model to /content/drive/My Drive/Colab Notebooks/NLP Projects/sentiment/Model.h5\n",
            "13/13 [==============================] - 1s 78ms/step - loss: 0.7234 - accuracy: 0.7121 - val_loss: 0.8968 - val_accuracy: 0.6450\n",
            "Epoch 7/10\n",
            "13/13 [==============================] - ETA: 0s - loss: 0.5442 - accuracy: 0.8073\n",
            "Epoch 7: val_accuracy improved from 0.64500 to 0.65000, saving model to /content/drive/My Drive/Colab Notebooks/NLP Projects/sentiment/Model.h5\n",
            "13/13 [==============================] - 1s 78ms/step - loss: 0.5442 - accuracy: 0.8073 - val_loss: 0.8319 - val_accuracy: 0.6500\n",
            "Epoch 8/10\n",
            "13/13 [==============================] - ETA: 0s - loss: 0.4470 - accuracy: 0.8385\n",
            "Epoch 8: val_accuracy improved from 0.65000 to 0.68000, saving model to /content/drive/My Drive/Colab Notebooks/NLP Projects/sentiment/Model.h5\n",
            "13/13 [==============================] - 1s 79ms/step - loss: 0.4470 - accuracy: 0.8385 - val_loss: 0.8815 - val_accuracy: 0.6800\n",
            "Epoch 9/10\n",
            "13/13 [==============================] - ETA: 0s - loss: 0.3435 - accuracy: 0.8648\n",
            "Epoch 9: val_accuracy did not improve from 0.68000\n",
            "13/13 [==============================] - 1s 66ms/step - loss: 0.3435 - accuracy: 0.8648 - val_loss: 0.9485 - val_accuracy: 0.6700\n",
            "Epoch 10/10\n",
            "13/13 [==============================] - ETA: 0s - loss: 0.2778 - accuracy: 0.8836\n",
            "Epoch 10: val_accuracy did not improve from 0.68000\n",
            "13/13 [==============================] - 1s 65ms/step - loss: 0.2778 - accuracy: 0.8836 - val_loss: 1.0058 - val_accuracy: 0.6750\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plt.title('GRU:Loss and accuracy')\n",
        "plt.plot(history.history['accuracy'])\n",
        "plt.plot(history.history['loss'])\n",
        "plt.legend(['accuracy', 'loss'])\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 293
        },
        "id": "yhIwLfznYwyO",
        "outputId": "f98f4a70-0f83-4bd9-bb89-77bc61e03ff0"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plt.title('GRU:val_loss and val_accuracy')\n",
        "plt.plot(history.history['val_loss'])\n",
        "plt.plot(history.history['val_accuracy'])\n",
        "plt.legend(['val_loss', 'val_accuracy'])\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 293
        },
        "id": "wcdCvjmeXJ7n",
        "outputId": "4787a058-fada-47d2-de4d-b1c0eb57919d"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX8AAAEUCAYAAADDdzb+AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdeXxM9/748dfMZCb7viEICSEksSaxxBYUodWqlnu7qeqmSqu06OJ729vS3i6/6sqt6qK9KLWW2IVSa+1RIpZERHZZZJ85vz8iw5RIZJss7+fjkYfJ52zv88l4n3M+53M+R6UoioIQQohGRW3uAIQQQtQ+Sf5CCNEISfIXQohGSJK/EEI0QpL8hRCiEZLkL4QQjZAkf1FpM2bMIDw8/K6W+fXXX2nXrh2XLl2qoajqvnbt2vHZZ5/V+nbDw8OZMWNGrW9X1E2S/KtJeno6H3/8Mffeey9dunQhMDCQgQMH8uqrr7J3716TeUsTYOlPhw4d6NOnDzNmzCApKemWdYeHh/PUU0/ddruXLl2iXbt2LFiwoEb2SwjRMFmYO4CG4Pjx4zz77LNkZ2cTERHB2LFj0el0XLp0iW3btvHEE0+wYMEC+vXrZ7Lciy++SIsWLSgsLOTIkSOsXLmSQ4cOsW7dOiwtLc20N0KIxkCSfxVlZWXxwgsvoFarWbVqFb6+vibTX375ZTZt2oStre0ty4aFhdG5c2cAHnroIZydnfnvf//L1q1biYiIqJX4hWhI8vLysLa2NncY9YI0+1TRkiVLSEpKYtasWbck/lL33HMP3bt3L3ddpfPEx8dXKaaioiJiY2NJTk42lj377LP079+f243mMW7cOJO2+19//ZVx48bRu3dvAgICuOeee5g/fz4Gg6FKcd3Jxo0bGTVqFEFBQYSGhjJ16lQuX75sMk9qaiqvv/46/fr1IyAggF69ejFhwgRiYmKM85w8eZKnn36aHj16EBgYyIABA5g+fTr5+fl33P7Bgwd56aWXGDBgAAEBAYSFhfHGG29w9epVk/k+++wz2rVrx7lz55gxYwbdu3enW7duzJw5k7y8PJN5CwsLee+99+jRowddunThueee48qVK+XWRVFRESEhIUyfPv2Wafn5+XTt2tXYdl9YWMi8efN48MEHCQ4OJigoiNGjR7Nly5Zyt1MRFa0XgOTkZN566y369u1LQEAA4eHhvPHGG+Tk5Bjnyc7O5v3332fgwIEEBATQp08fXnnlFWNzZ1n3hPbt20e7du3Yt2+fseyxxx5j6NChnDp1iscee4zOnTvzr3/9q1rjvnDhAu3atWPRokW3LPfXX3/Rrl07fv7558pVrpnJmX8Vbdu2DSsrKwYPHlzldSUkJADg4OBQpfUkJSURERHBAw88wNy5cwGIiIhgx44dHD58mK5duxrnTUtLY//+/YwfP95Y9tNPP+Hr60u/fv3Q6XTs3buXjz/+mOzsbKZNm1al2G5n9erVvPrqq3Ts2JGpU6eSkZHBDz/8wKFDh1i5ciUuLi4ATJ48mTNnzvDII4/QvHlzMjIy2L9/PxcuXKBt27akp6czfvx4nJycmDBhAo6OjiQmJrJt2zZyc3OxsrIqM4bIyEiys7N5+OGHcXV15fTp0/zyyy/ExMSwZMkSVCqVyfxTp06lRYsWTJ06lejoaH755RdcXFxMEvbrr7/OmjVrGDFiBF27dmXfvn0888wz5daHVqtl8ODBbNiwgYKCApMmwKioKK5du2a8MszJyWHp0qVEREQwatQoCgsLWbt2LS+88MJtmxrvVkXrJSUlhYceeoiMjAwefvhh2rZtS3JyMps3b+bq1avY2dmRm5vLo48+SkxMDA888AABAQFcvXqVqKgoLl68iKen513Hl52dzVNPPcU999zDiBEjsLe3r9a4W7VqRZcuXVizZg1PPvmkybbXrFmDVqutv1fpiqiS4OBgZeTIkbeUZ2dnK2lpacafnJwc47QVK1Yofn5+ys6dO5W0tDQlMTFRiYyMVHr06KEEBAQoiYmJJusaMGCAMn78+NtuPz4+XvHz81Pmz59/S9lrr71mEk9QUJDyzjvvmCy/ePFixc/PT4mOjjaW5ebm3rKdN954Q+ncubNSUFBgLHvttdeUAQMGlFU1t1W67/Hx8YqiKEphYaHSq1cvZdiwYUpeXp5xvr179yp+fn7K3LlzFUVRlMzMTMXPz0/55ptvylz35s2bFT8/P+XYsWN3FZOi3H6f16xZo/j5+SkHDhwwls2bN0/x8/NTZsyYYTLvCy+8oISEhBh/P3XqlOLn56fMnj3bZL7p06crfn5+yrx58+4Yz+7duxU/Pz9l06ZNJuWTJ09WQkNDlaKiIkVRFKW4uNjkb6IoilJQUKAMHz5ceeKJJ0zKBwwYYPKdqIiK1strr72mtG/fXjly5Mgt8xsMBkVRbtTd+vXry5zn79+PUqXfh7179xrLHn30UcXPz0/5/vvvazTun3/+WfHz81NiYmKM0/R6vdK3b1/lhRdeuGW5+kKafaooJycHGxubW8rffPNNevbsafx55513bplnwoQJ9OzZk379+jF58mRsbGz46quvaNKkSZViat68OadPnzae9QPY2dnRr18/IiMjTZpvNmzYgI+PD/7+/say0jZTvV5PZmYm6enpBAcHk5uby7lz56oU29+dOHGC1NRU/vGPf5icmYeGhtKxY0d27NgBgJWVFVqtlv3799/20h0wnvXt2LGDoqKiu4qjdJ8VRSEnJ4f09HS6dOkClDQl/d3DDz9s8nv37t25evWqsYkjKioKgEcffdRkvscee6xC8YSGhuLm5sb69euNZbm5uURFRXHPPfdgYVFy0a7RaNDpdEBJE1BpDN27d79t3HerIvViMBjYvHkzffv2pVOnTreso/Qse+PGjbRt25Zhw4aVOc/dsrCwYMyYMTUad0REBDqdjtWrVxun7du3jytXrjBy5MhKxV0XSLNPFdna2nLt2rVbyp9//nkeeughAJ577rnbLvvGG2/g6+tLdnY2K1eu5MCBA3dsmqiqiIgINm7cyIEDBwgNDSUpKYlDhw4xceJEk/kOHjzIJ598wtGjR29JotnZ2dUaU2m7fuvWrW+Z5uvry8aNGwHQ6XRMmzaNDz74gN69exMUFES/fv0YOXIkTZs2BSAkJIQhQ4bw+eefs2jRIkJCQggPD2fEiBG3PUDfLDExkQ8++MDYrHKz2+1zs2bNTH4vbarLzMzEzs6OhIQEVCoVLVu2NJnvdvt5OxqNhiFDhvDrr7+Sm5uLjY0N27ZtIy8vj+HDh5vM+8svv/Ddd98RGxtrck+nsgn1ZhWpl/T0dHJycmjbtu0d1xUXF8fAgQOrHNPNPDw8btszrjrjdnR0JDw8nHXr1jF16lRUKhVr1qzBycmpys1q5iTJv4p8fX05deoURUVFaLVaY7mfnx9+fn5AyX/k2wkMDDT29hk0aBCPPvooU6dOJTIy0iRZ6XQ6CgoKbruO0huZFeka2r9/f2xtbVm/fj2hoaFs2LABg8Fg0mYZHx/Pk08+SatWrZg5cybNmjXD0tKSkydP8uGHH9boTd/yjBs3joEDB7J161b27NnDl19+yfz58/n6668JDQ1FpVIxb948jh07xvbt29mzZw9vvvkm8+fPZ9myZbi6ut52vXq9nvHjx5Oens6zzz6Lr68v1tbWGAwGJkyYcNub5Gr17S+abzdvZUVERPDTTz+xY8cOIiIiWL9+PR4eHgQHBxvnWbNmDW+88QYDBgzg6aefxsXFBQsLC1asWMG6deuqtP3K1EtVlXXAKut7d7vvfU3Eff/99xMZGcmBAwfo1KkTmzZtYsSIEcarrvpIkn8VDRgwgMOHD7Nx40ZGjBhR6fVoNBpeeeUVHnnkERYvXmxyY9DLy4vTp0/fdrnSZhgvL69yt2FlZUV4eDibNm3irbfeYsOGDbRv396kl9LWrVspLCzk66+/NllnTT2RW3oGff78ecLCwkymnTt37pb9atGiBePGjWPcuHHGy+7S5F8qKCiIoKAgpkyZQlRUFM888wzLli3j+eefv20MZ86c4dy5c8ydO5cHHnjAWH7hwoVK75eXlxeKohAXF0ebNm2M5efPn6/wOrp160bTpk1Zv349ffv2ZdeuXYwdO9bkwBMZGUmLFi346quvTBLnihUrKh17qYrWi4uLC3Z2dia9rm6nZcuW5c5TegX196ut0s4Q5ogboE+fPri6urJ69WpSU1PJycmp100+IF09q2zs2LG4u7szd+5cYmNjq7Su7t2706VLF77//nuTM/2+ffuSkpJCZGSkyfzFxcX873//w9rampCQEGP57bp6lho+fDjp6en8+uuvHDly5JYmhNKrlJvPjgoLC1m8eHGV9q0sAQEBuLm5sXTpUpN9PnjwICdOnKB///5ASf/tv3fXbNKkCa6urmRlZQElTS5/P6vr2LEjgHGe2ylNpn9f9ttvv63cTlHyNwNuqbe7qUeVSsWwYcPYuXMnq1atorCwsEJ/r/j4+Grp6lnRelGr1QwePJidO3dy9OjRW9ZTuvyQIUOIiYlhw4YNZc5T2kx24MAB4zS9Xs+yZcvMFjeU3Fu499572bhxIytWrMDb29uk11x9JGf+VeTo6MiXX37JM888w/3338/w4cMJDAxEp9ORmJjI5s2byc3NNbZLl2f8+PG8+OKLLF++nEceeQSAMWPGsGLFCqZOncquXbvo2LEjOTk5bNy4kRMnTjBz5kyT7qG36+pZKiwsDEdHR+bMmQNwSze1sLAwtFotzz33HGPGjKGwsJDVq1eX2cxRVVqtlunTp/Paa6/xz3/+k/vuu4/09HR+/PFHPD09efrpp4GSs7YnnniCoUOH0qZNG3Q6HVFRUcTGxvLaa68BsHLlSn7++WcGDRpEy5Ytyc/P59dff0Wj0TB06NAyY/Dx8aFVq1a8//77XLlyBUdHR3bt2lWhPvll8ff3Z8SIEfzvf/8jOzvb2NXzbq8mIiIi+Pbbb/n444/x8vIyNhOWKr2Se/755wkPDycpKYmff/6Z1q1bc+rUqUrHD3dXL1OnTmX37t089thjjBkzhjZt2pCamsrmzZv5/PPPad68OU899RSbNm3ilVdeYffu3XTs2JHs7Gx27tzJ5MmTCQkJoW3btnTu3JmPP/6YzMxMHB0dWb9+PcXFxWaLu9T999/Pd999x++//86LL75YuUqtQyT5V4OgoCB+++03Fi1axI4dO4iMjESv1+Ph4UHXrl2ZNWsWPXv2rNC6Bg0ahLe3NwsXLmTMmDFYWFhgZWXFTz/9xJdffsmWLVtYvXo1Wq0Wf39/5s2bx5AhQyoca2kf8uXLl9OpUyeTLzeU3JD84osv+OSTT/jPf/6Ds7Mz999/PyEhISbPAlSn+++/H2tra+bPn8+HH36ItbU1/fr1Y9q0acY+/k2aNOHee+/ljz/+YO3atahUKlq1asW7777L6NGjgZIbvsePHycyMpKUlBTs7Ozo0KEDb7755m17c9xcJ1999RXvvvsuCxcuRKPR0KdPH7755ht69+5d6f167733cHZ2Zu3atWzbto3Q0NC77nsfGBiIt7c3Fy9e5J///Oct0x944AHS0tL43//+x549e/D29mbmzJnExcVVOfnfTb14eHjwyy+/8Omnn/Lbb7+RlZWFh4cHYWFhODs7A2BjY8PixYv57LPP2Lx5M6tWrcLFxYXg4GC8vb2N6/rwww956623WLBgAQ4ODowePZrQ0NBb+tnXVtyl/P39adeuHadPn+a+++6rTJXWKSqlJu7aCCFEA/Tggw+i1WpZsmSJuUOpMmnzF0KICjh16hQnTpwwuYlcn0mzj6gW2dnZ5Y6f4+LiUma3V1G70tPT0ev1ZU7XaDTGJrfG7syZM5w8eZLvvvsOV1fXBtHkA5L8RTV59913Wbly5R3n2bp16y33GIR5jB49+o7dJ728vNi2bVstRlR3bdy4kS+++AJvb28+/vjjBjNqqLT5i2px9uzZ23YtvVm3bt3kPQV1xKFDh8p8cBBKHp7q1q1bLUYkapskfyGEaITkhq8QQjRC9abNPyPjGgbD3V+kuLrakZaWU/6MjYTUhympjxukLkzV9/pQq1U4O9/6BsFS9Sb5GwxKpZJ/6bLiBqkPU1IfN0hdmGrI9SHNPkII0QhJ8hdCiEZIkr8QQjRCkvyFEKIRqtAN34sXL7Jw4UKOHj1KTEwMPj4+FX5LUHZ2NvPmzWPjxo2kp6fj4eHByJEjmTJlSpUCF0IIUXkVSv4xMTFERUXRqVMnDAZDhV+Dlpuby6OPPopKpWL69Ol4eHgQHx9fpXHS75Y8wybMIS/vGjk5V9HrKz4OvbklJ6vN+prOuqYu14dGY4GdnRPW1mV35SxPhZJ/eHg4gwYNAmDGjBmcOHGiQitfsGAB2dnZrF27FlvbkiBvft1eTTt5IZ1li/bzypjOONjU33dtivolL+8a2dkZODm5o9XqquVF6rXBwkJNcXHdTHbmUFfrQ1EUiooKuXo1BaDSB4AKtflX9i1Oy5cvZ/To0cbEX9tc7C1JSLnGmt8r/t5UIaoqJ+cqTk7u6HSW9Sbxi/pDpVKh01ni5OROTs7VSq+nxm74Xrp0iZSUFJydnXnuuecIDAyke/fuvPrqq2RmZtbUZk00dbVlWM9W7Dh8mYTUa7WyTSH0+mK0WrnSFDVLq9VVqVmxxp7wTU1NBeCDDz4gPDyc+fPnk5CQwEcffURaWhoLFy68q/W5utpVKo5/3NOOHYfiWb37ArMn9KjUOhoad3d7c4dQp1R3fSQnq9Fq6+d7CywspAPgzep6fajV6kp/f2ss+ZfeKPH29ubDDz80Xv7a29szZcoUjh07RlBQUIXXl5aWU6lHrd3d7RnesxXLtp9l+/4LBLR2vet1NCTu7vakpGSbO4w6oybqw2Aw1Mm24vLU1TZuc6kP9WEwGMr8/qrVqjueNNfYYc3R0RGAnj17mrR7lr7IPCYmpqY2fYuB3Zrj4WTN0q1n0dfRu/dCNFRhYd35+ecfKzz/6NH38vHH79dgRAJqMPm3aNECna7sds87vUiiumkt1Dw0wJeE1GvsOppYa9sVQoi6qsaSv06no3fv3uzZs8ekr/3u3bsBCAgIqKlN31ZXP3f8Wjixctc5cvPrT99rIYSoCRVK/nl5eURGRhIZGUlCQgI5OTkmvwPMmjWLDh06mCw3adIkYmNjmTp1Krt27WLp0qX861//Iiws7K7a+6uDSqVi7MA2ZOcW8dveC7W6bSHqmw0b1tG3bwipqSkm5Xl5eQwaFMbixd8RF3eB2bNnMWrUcMLDe/PII6P58cdFFBdX/8nVzp07GD/+EcLDe3Hvvffw4Ydzyc3NNU4vLi7myy/n8eCDIxgwoCf33TeEadMmk5GRXqHpjVGFbvimpaXdMhxD6e9z5sxh1KhRGAwG9Hq9yTwBAQF88803fPTRR0ycOBE7OzsiIiKYNm1aNYV/d1o1caBXQBM2H4inf2cv3J0axouYhahuffv25z//mcPWrZsYM+YRY/muXVEUFBQwaNBQLl++RIsWLRk8eAg2NracO3eWhQsXkJWVxQsvVN/wLb//HsXrr0+nf/+BPP30RBITLzN//ufEx8fx6adfArB48XesXLmc559/kdatfcjKyuTgwQPG5uXypjdGFUr+zZs35/Tp03ecZ+7cucydO/eW8h49evDLL79ULroa8GA/Xw7+lczyHbE8f3/tNj2Jxmv38UR+P2a++01hQU3pHdi0wvPb2trRq1dvtmzZaJL8t2zZSGBgJ5o0aUKTJk3o2rU7UPLUaVBQZwwGhW+/nc/EiZOr7QG3b79dQPv2HXjnnRv5xdHRidmzZ/Lnnwfp2rU70dEnCAkJZdSoh4zz9OsXbvxc3vTGqG53Yq0BzvaWDA1tyYG/kom5VPmn44Ro6AYPHsqpU9EkJFwCICsrk/37/2Dw4KFASaeNhQvnM2bM/YSH96J//x7Mm/cROTk5pKenVUsMubm5xMScITx8sEl5//7haDQajh07AoCfX3v++GMPCxfOJzr6xC2tEOVNb4zqzWscq9OwUG92Hr3Mkq1nef3xbqjlEXxRw3oH3t2Zd13Qs2cYdnZ2bN4cybhxE9ixYxuKojBgQMk4X1999Rlr1qzkyScn0L69P3Z29hw4sI8FC76ksLCwWmLIyclGURRcXU2fz9FoNDg6OpGVVTJawOOPj0etVhMZ+RuLFv0XR0dHRo58kKeeehaNRlPu9Mao0Z35A1jqNDzYz5fziVnsi04ydzhC1Ek6nY6+fQewZcsmoKTJJzg4FCcnJwC2b9/CyJGjeOyxJwkO7oG/f8dqH9bCzs4elUp1y5WEXq8nM/MqDg6OxljHj3+GZctWs2zZakaPHsuPPy5i7dqVFZreGDXK5A/QM6AJ3k3sWb4jloIiuQQU4nYGDx7ChQvn2Lt3D0eO/Gls8oGSZh+tVmv8XVEUNm+OrNbt29jY0LatH9u2bTEpj4rajl6vJyio8y3LNGvmxZNPPo27uwcXLtw6qGN50xuLRtnsA6BWqRgb3ob3fz7Mpv1x3Nu7tblDEqLO6do1GFdXV+bOfQedTkefPv2N04KDQ1mz5le8vVvh4uLCmjWryM6u/qFDxo9/hpkzpzF79iyGDRvBlSuX+frrz+nWLcR4w3nmzFfw82tP27btsLGxYd++P0hOTqJ795AKTW+MGm3yB2jX0plufu6s3xtHn07NcLKzNHdIQtQpGo2GAQMGs3z5EgYOHIyNjY1x2tSpr/Kf/8xh3ryP0Gp13HPPUAYMGMjbb79ZrTGEhfXj3//+gO+++y8zZ76Cra0dgwYNZeLEF43zBAZ2Zvv2zSxd+hN6vZ7mzVvw+uv/R1hYvwpNb4xUSj151VVVBna708BdyRm5vP7fffQMaML4CP+qhFgvyMBupmqiPq5cuUiTJt7Vus7aUB8GMqtN9aE+7vRdM9vAbvWFh7MNg7o3Z/exROKSJCkKIRqHRp/8Ae7t1Qpbay1LtsbIO3+FqCGKolBcXFzmj/S9r12Nus2/lI2VlpFhrflp8xmOxKTSxc/d3CEJ0eAcPnyIyZOfK3N6585d+fzzBbUYUeMmyf+6/l2ase3PSyzbfpZAX1csNHJRJER1at/en2+++aHM6TffTBY1T5L/dRq1mjHhbfh/vxxj258J3BPcwtwhCdGg2NjY0r59h/JnFLVCTm9vEujjSsfWLqzdfZ6cvCJzhyOEEDVGkv9NVCoVY8LbkFtQzJrfG++Tf0KIhk+S/980d7ejX6dmbD+cQGLaNXOHI4QQNUKS/22M7OOD1kLNL9tjzR2KEELUCEn+t+Foq2NEr1YcOZtK9IXG+5o3IUTDJcm/DIO7N8fN0YolW89WalgJIYSoyyT5l0FroWF0f18upeTw+3HzvX5PiPouLKw7P//8o7nDEH8jyf8Ogtt70MbLkV93niOvoNjc4QghRLWR5H8HKpWKMQPbkHWtkA37Lpo7HCFEHdBQxiGS5F8O32aO9Ojgycb98aRl5ps7HCFqxYYN6+jbN4TU1BST8ry8PAYNCmPx4u+Ii7vA7NmzGDVqOOHhvXnkkdH8+OMiiourdpW8bNnPTJjwOEOG9GPEiEFMnfoi587d2vPuxIljvPzyC9xzTz8GD+7L008/wYEDe43TCwsLWbDgSx56aCQDBvTkgQciePfd/zNOnzTpGV599SWTdSYmXiYsrDvbt994c9jo0ffy8cfvs2TJYh566D7Cw3uRmppS4f2/UxxRUdsJC+tOfHycyTIFBfkMHdqfH374trLVWK4KDe9w8eJFFi5cyNGjR4mJicHHx4d169bd1YY2b97MpEmTaNu27V0va24P9vPl0JkUlkfF8ux9Hc0djhA1rm/f/vznP3PYunUTY8Y8YizftSuKgoICBg0ayuXLl2jRoiWDBw/BxsaWc+fOsnDhArKysnjhhSmV3nZSUhKjRj1EkyZNycvLY+3alTz//Hh++mkFbm5uABw7doQpU56nQ4cAXnvtDezt7fnrr2iSkq4Y1/PGG69y6NABHnvsSTp2DOTq1QyiorZXKqaoqO00bdqMSZNeQqvVYWdnR0JCxfb/TnH07t0HV1c3fvttDc89N8m4zPbtW8nLyyMi4t5KxVsRFUr+MTExREVF0alTJwwGw10Pe5yXl8d7771n/MPVN66OVgwJacm6PRcY1K05vl6O5g5J1DNFZ3ZTdHqn2bavbdcXrV/vCs9va2tHr1692bJlo0ny37JlI4GBnWjSpAlNmjQxvkZRURSCgjpjMCh8++18Jk6cjEqlqlSsL774svGzXq8nJKQHDzwQwZYtkYwd+ygAX301Dy+vFsyb9zUajQaAkJAexuUOHNjLnj2/M3v2v03eO3zz57tRVFTExx9/bjL4XNeu3cvd//LisLCwICLiXtavX8vTTz9v3Jd161YTGtoLN7eaG2G4Qsk/PDycQYMGATBjxgxOnDhxVxv58ssvad68OV5eXne9bF0R0aMlu45eZsm2GGY92q3SX2wh6ovBg4fy+uuvkpBwCS+v5mRlZbJ//x9MmTINKHmB++LF37Fp0waSk5MoKroxHlZ6ehqurpU72Ttx4jgLF37NmTN/kZmZaSwvbRrJz8/n5MkTPPvsC8Zk+XcHDx7AysqKQYOGVCqGv+vatdsto45WZP8rEsd99z3A4sXf8ccfuwkL60t8fBxHjvzJnDkfVkvsZalQ8lerK39rIDY2lh9//JFly5bx7bc1135V06x0Fozq68OiDX9x4K9kQvw9zR2SqEe0fr3v6sy7LujZMww7Ozs2b45k3LgJ7NixDUVRGDCg5ETwq68+Y82alTz55ATat/fHzs6eAwf2sWDBlxQWFlZqm1euXGHq1En4+bXjlVdm4ubmhk6n4403XjOuMzs7C4PBcMez4qysTFxd3artJM3Z2fWWsorsf0XiaNq0GcHBofz222rCwvqybt1qXFxc6dkzrFpiL0uND+n89ttvM3r0aPz8/Gp6UzWud2BTth66xC/bY+nS1g2txe3POoRoCHQ6HX37DmDLlk2MGzeBLVs2EhwcipOTEwDbt29h5MhRPPbYk8Zljhw5XKVt7tu3h7y8XN5770McHETvWygAACAASURBVByM5VlZWcbPdnb2qNXqW25G38zBwZG0tFQURSkz8ep0liZn63/fzs1ut4qK7H9F4oCSs//Zs2eRnJxEZOQ6hg4djoVFzabnGu3t89tvv3HmzBkmT55ck5upNWp1yaifaVn5bDoQb+5whKhxgwcP4cKFc+zdu4cjR/40abcuKChAq9Uaf1cUhc2bI6u0vYKCAlQqlUni+/33KHJzbwyyaG1tTceOgURG/lZml8vu3UPIz89n27bNZW7L09OTixcvYDDceEn7/v1/3FWs5e1/ReIACAvrh6OjE2+//SZpaWmMGDGywnFUVo0dWnJycpg7dy5Tp041OYJX1p3eQl8ed3f7Km//5nXtPH6F9XvjGDmgLc72VtW27tpSnfXREFR3fSQnq7GwqJ+9qP8ed0hIKK6ubrz//jvodDoGDAg3zhMa2oM1a37Fx6c1Li6urF69kpycbAA0GtM60GhUFaqTkJAQAObM+RcPPDCauLiL/PDDIpydna8fFErW8cILk5k06Vleemkio0c/jL29A6dPn8LJyYl7772fnj170qtXGHPmvE1iYgIdOwaQlZXFtm1bePfd9wEYNOge1q5dxSeffMCAAQM5deokGzasM8Z/c32o1bfGX5H9r0gcJdvRMWLEffzwwyI6d+6Cj0/rivy5UKvVlf7+1ljy//rrr3FycmLw4MHGS6mioiIMBgNZWVlYWVmh0+kqvL60tJxKjbHj7m5PSkr2XS93JyN7t+LgqSQWrjrOE0PbV+u6a1pN1Ed9VhP1YTAYKC42lD9jHWNhob5N3CoGDBjE8uVLGDhwMDqdlXGel16azn/+M4dPPvkQrVbHPfcMpV+/cN5++030etM60OuVCtVJq1a+vP76//HttwuYNu0lfH19mT37XT788L3rL4AvWUdAQCc++2w+//3vV7zzzmzUag2tW/vw9NPPG+d55533WbTov6xcuYJvvpmPi4srwcGhxunduoXw4osv88svS9iwYR3dugUza9ZsnntuPHp9yTyl8xoMt8Zf0f0vL45Sffr054cfFjF8+MgKf38MBkOZ31+1WnXHk2aVcpf9Nkt7+5TXV3/ixIls3bq1zOkzZ85k3LhxFd5uXUr+AD9vOcPWQ5f415MhNPeo/FVJbZPkb6om6uPKlYs0aeJdreusDbdP/o1XbdfHokX/ZenSn1i1KhIrq4q1KNzpu1Ze8q+xM/+XXnqJJ554wqRswYIFnD9/njlz5uDtXf/+c9zsvt6t+ePEFZZsi+GVMZ2l66cQolLi4i4QF3eRpUt/5r77RlU48VdVhZJ/Xl4eUVFRACQkJJCTk0NkZMmNjcDAQLy8vJg1axarVq0iOjoa4La9e1auXElSUhKhoaHVFb/Z2Flrua93a/63NYZjsWl0alM/H2ATorYoinLHMXFUKlWZ/fYbsg8+eI/o6BMEB4cybtyEWttuhZJ/WloaU6aYPq5d+vucOXMYNWoUBoOhQQx2dDcGdPVi2+EElm0/S8fWLlho6udNPiFqw+HDh5g8+bkyp3fu3JXPP19QixHVDeba57tu8zeXutbmX+pwTAqfrTjOI4P9GNiteY1tp7pIm78pafO/oabbuHNzrxEXV/bouDY2NrRs2arGtn+36sM9kDrZ5t9YdG7jhr+3M6t/P0+Pjp7YWmnLX0iIRsjGxpb27TuYOwxxnbRTVJFKVfLg17W8ItbuvmDucIQQokIk+VeDlp72hAWVDP2QlJ5r7nBEHVBPWlNFPVbV75gk/2ryQF8fLDRqftlx60snROOi0VhQVFS5gc2EqKiiokI0msq33EvyryZOdpZE9PTmzzMp/HUxw9zhCDOys3Pi6tUUCgsL5ApAVDtFUSgsLODq1RTs7JwqvR654VuNhgS3IOpIAku2xfDWuGDU8uBXo2RtbQtAZmYqen3VXmlYm9RqtckgZ41dXa4PjcYCe3tn43etMiT5VyOdVsPofr4sWBvNnuNXCAtqau6QhJlYW9tW6T+mOUg3YFMNvT6k2aeahXbwxKeZAyt2xlJQ2LgeehNC1B+S/KuZSqVibHhbMnMK2bCv7AdahBDCnCT514A2zR0J8fcgcl8c6Vn55g5HCCFuIcm/hozu54tBgRVR58wdihBC3EKSfw1xc7LmnuAW/HHyCucTb/9eUCGEMBdJ/jVoeE9vHGy0LNkaI/29hRB1iiT/GmRtacH9fX2IuZTJodMp5g5HCCGMJPnXsD5BTfFyt+V/W2PIvCaP/Ash6gZJ/jVMo1YzYXgHruUV8cWvxymq4+ODCyEaB0n+tcC7iT3jh/tzNiGTHzeelvZ/IYTZyfAOtSTE35OElGus3XMBL3dbhoS0NHdIQohGTM78a9HIPq3p5ufOsu1nORabZu5whBCNmCT/WqRWqZgwogPN3e2Yv+YEl1OvmTskIUQjJcm/llnqNEx+MAitRs28FcfIySsyd0hCiEZIkr8ZuDpaMWlUEOlZ+Xy16gTFeukBJISoXZL8zaRNc0ceH9KeUxczWLr1rLnDEUI0MhXq7XPx4kUWLlzI0aNHiYmJwcfHh3Xr1t1xmZycHBYtWsTOnTs5f/48FhYWdOzYkalTp9KxY8dqCb6+CwtqSkJqDhv3x+Plbkv/Ll7mDkkI0UhU6Mw/JiaGqKgovL298fX1rdCKL1++zNKlS+nVqxeffPIJc+bMwWAwMHbsWE6ePFmloBuSh/q3IdDHlZ82n5F3/wohao1KqcATRwaDAbW65DgxY8YMTpw4Ue6Zf25uLiqVCmtra2NZQUEBAwcOpE+fPsyZM+euAk1Ly8FguPuHo+rDq9hy84t598eDZF0r5M1xwXg4WZe/UCXVh/qoTVIfN0hdmKrv9aFWq3B1tSt7esVWcve3BmxsbEwSP4ClpSW+vr4kJyff9foaMhsrCyaPDgLgs+XHyCuoPy/9FkLUT7V6wzc3N5dTp07h4+NTm5utFzydbXj+/gAS03JZsOZkpa5yhBCiomp1eIf/9//+H3l5eTz66KN3veydLl/K4+5uX+lla1M/d3tyCvR8vfI4Gw7EM25EzdwYry/1UVukPm6QujDVkOuj1pL/2rVr+f7773nrrbfw9va+6+Ubcpv/zYL93Pirixcrtp/F2VZLr4Cm1br++lYfNU3q4wapC1P1vT6qpc2/qnbv3s3MmTN56qmneOSRR2pjk/WWSqXin4Pa0r6lE99tOE1sQqa5QxJCNEA1nvyPHTvGpEmTGDZsGNOnT6/pzTUIFho1Ex8IxNlex2e/Hic9K9/cIQkhGpgaTf6xsbE8/fTTdO3alffeew+VSlWTm2tQ7Ky1TH4wiMIiPZ+tOE5Bkd7cIQkhGpAKJf+8vDwiIyOJjIwkISGBnJwck98BZs2aRYcOHYzLpKWl8dRTT6HVapkwYQInT57kyJEjHDlyhOjo6JrZmwbGy92OZ+7rSFxSNt/+dkpeAiOEqDYVuuGblpbGlClTTMpKf58zZw6jRo3CYDCg1984Oz179iyJiYkAjBs3zmRZLy8vtm3bVpW4G43ObdwYPcCXX7bH4uVuy329W5s7JCFEA1ChJ3zrgsbS2+d2FEXhm3Wn+OPkFSbeH0D39h6VXldDqI/qJPVxg9SFqfpeH3Wit4+oGpVKxbhh7fBt5sA3v0UTl1R/v5BCiLpBkn89obXQMGlUILZWWuatOEbmtUJzhySEqMck+dcjjnaWTH4wiJzcIr749ThFxfISGCFE5Ujyr2e8m9jz1IgOnE3I5IeNf0kPICFEpUjyr4eC23twX+9W7D5+hU0H4s0djhCiHpLkX0/dF9aabu3cWbb9LMdi08wdjhCinpHkX0+pVSomDO9AC3c75q85weXUa+YOSQhRj0jyr8csdRpefDAIrUbNvOXHyMkrMndIQoh6QpJ/PefqaMWkUUGkZ+fz1aoTFOulB5AQonyS/BuANs0deWJoe05dzGDJ1hhzhyOEqAdq9U1eoub0DmxKQso1IvfH4eVux4AuXuYOSQhRh8mZfwMyur8vQb6u/Lz5DH9dzDB3OEKIOkySfwOiVqt45t6OeDhb88XK4yRfzTN3SEKIOkqSfwNjY2XB5NFBAMxbfoy8gmIzRySEqIsk+TdAns42TLw/gCtpuSxYc7JSQ2ELIRo2Sf4NlH8rFx4Z3JajsWmsiIo1dzhCiDpGevs0YAO6NudSyjU27IvDy92WXgFNzR2SEKKOkOTfwP1jUFsS067x3Ya/8HS2wd3d3twhCSHqAGn2aeAsNGomPhCIi70Vn/16nPikbBkGWgghZ/6NgZ21lhdHB/HuDweZ+ME2dFo1Hk42eLpY4+FsjaezDZ7O1ng42+Bkp0OlUpk7ZCFEDZPk30h4udny1rhg4lJzORefQVJ6Lgkp1zgSk4r+pt5ANx8YPJ1trh8c5MAgREMjyb8RaeJiQ2A7T1JS3IxlBoNCWlY+SRm5JGfkkZSeR3LG7Q8MlloNHs5/v1qwxtPFBkdbOTAIUZ9UKPlfvHiRhQsXcvToUWJiYvDx8WHdunUV2sCqVav4+uuvSUhIoGXLlrzwwgtERERUKWhRfdRqFe5O1rg7WUNr02l6g4H0rAKSMnKvHxTySMrI5dIdDgylVwme1w8KHs7WcmAQog6qUPKPiYkhKiqKTp06YTAYKnzDMDIyktdee41nnnmG3r17s2XLFqZOnYqtrS39+vWrUuCi5mnUauOBIeA2B4a0rAKS03NJun5QSM7IIz7lGofvcGAoPSAE+bjiaGdZy3skhCilUiqQyQ0GA2p1ScegGTNmcOLEiQqd+Q8bNgw/Pz8+/fRTY9n48ePJyspi+fLldxVoWlpOpZ5UdXe3JyUl+66Xa6hqoz7KOjAkpeeSmpmP3qCgtVDTN6gZw3q0xMXBqkbjuRP5ftwgdWGqvteHWq3C1dWuzOkVOvMvTfx3Iz4+nnPnzvHyyy+blI8YMYKZM2eSnp6Oi4vLXa9X1H0atRoPJ2s8nKwJ+Ns0vcFAYmoumw/Gs+NIAjuOJNA7sAkRPbzxcLYxS7xC1FWKotRYk2mN3fA9d+4cAL6+viblbdq0MU6X5N/4aNRqmnvY8WSEP/f2bkXkvjh2Hk1k17FEenTwJKJnK7zcbM0dphBmk3w1j6MxqRw5m8q5xCxeGh1Eu5bO1b6dGkv+mZmZADg4OJiUOzo6mkwXjZebozWP3tOOEb1asXF/HDsOX2bvySS6tnNnRM9WeDeRp5FFw2cwKMRezuTI2VSOnk3jcuo1AJq62hDe1YtWTR3KWUPl1JuunndquyqPDGlgqq7Vh7u7PW1bu/HY8I6s3XWOtb+f49DpFLr7ezJmkB/tW9XsFWJdqw9zkrowVVP1kZtfxOHTKeyPvsLBU0lkXStEo1bR0ceViN6tCenQhKY1fAVcY8m/9Aw/KysLd3d3Y3npGX/p9IqSG77Vo67Xx5DuzekT4MnWPxPYfCCe6Z/twt/bmXt7taJdS6dqb/+s6/VRm6QuTFV3faRczbt+dp/K6bir6A0KtlYWBPq60rmNGwGtXbCx0pbMrBiqvO1queFbGT4+PkBJ2/7N7f6xsbEm04X4OxsrLff2asXg7s3ZcfgyG/fH8cH/DtPGy5ERvVoR6OMizw2IOs9gUDh3OcuY8BNuas4ZHNyCTr6utGnuiKYSHWqqQ40l/xYtWuDj48P69esZPHiwsXzdunUEBgbKzV5RLiudBUNDWxLe1YtdxxLZsO8i/++Xo3h72jOiVyu6+LmhloOAqEPyCoo5eT6dI2dTORabRk5eEWqVCr8WjowNb0Ontm541pFebRVK/nl5eURFRQGQkJBATk4OkZGRAAQGBuLl5cWsWbNYtWoV0dHRxuUmT57Myy+/TMuWLenVqxdbt25l9+7dzJ8/vwZ2RTRUOq2Ggd2a069zM/44cYXf/rjIFyuP4+Vmy/Be3oS090StloOAMI/Um5pz/rq5OcfHlU5t3Aj0uak5pw6pUPJPS0tjypQpJmWlv8+ZM4dRo0ZhMBjQ6/Um8wwbNoz8/Hy+/vprFi5cSMuWLfnoo4/k6V5RKRYaNX06NaNXYBMOnEpm3R8XWbAmmtW7zhPRw5ueAU2w0Mgo5aJmGQwK5xKzOHq2pDtmQkpJc46niw2Dujencxs3szbnVFSFnvCtC+SGb/VoSPVhUBQOn0ll3Z4LXEzKxtXBkmE9vOkT1BSthaZC62hI9VFVUhembq6PvIJioi/caM7Jzr3RnNOpjRud2rjRxKVuNOeUMtsNXyFqmlqlols7d7r6uXH8XDpr95xn8aYzrN19gaGhLenf2QtLXcUOAkL8XXJ6LtsOXeLI2VROx2VQrFewsSzpndOpjSuBPq7Y1sHmnIqS5C/qPZVKRZCvK4E+LvwVd5V1ey6wdNtZfvvjIoODWzCwa3NsrOSrLsqXlpnP/lNJ7ItOIi45BwBPZ2vCu95ozmkoTYvyP0I0GCqVCn9vZ/y9nTl7KZN1f1xg5c5zRO6LY2C35gzu3hx7G525wxR1THZuIQdPp7Dv5BXOXCp5DsmnmQNPjuhIm6Z2NHVtmMONSJt/I9PY6uPilWzW/XGBQ6dTsNRqGNDFiyEhLYzDSTe2+riTxlQX+YXFHIlJZW90EifPp6M3KDR1taFHB09CO3ji4WxT7+tD2vxFo+bdxJ4XHggkISWH3/ZeZOOBOLYcukTfTk0ZFuotwxk0IsV6AyfOp7MvOonDMSkUFhlwtrdkcHALenTwpIWHXaN6eFDO/BuZxl4fSRm5bNh7kd3Hr6Ao0MzdFndHK5q42tDUxZamrjY0cbWp1zfyKqshfjcMikJM/FX2RSdx4K9kruUXY2tlQXB7D0I7eNK2hVOZDwrW9/qQM38hbuLpbMO4Yf7c26s1u45dJiWrgIuJWRyLTTN5+5iDrY6mLjbXDwo2NHEtOTC4OljJA2V1nKIoxCfnsDe65MZtRnYBOq2aLm3dCe3gSUBrlwZz07YqJPmLRsnV0Yr7+/gYz+70BgOpV/NJTMslMf0aiWm5XEnP5eD1s8VSWgs1ns43DgpNXW1o6mqLp4s1Vjr572ROyRm57ItOYm90EolpuSWjZLZ24aH+vnRu6yZ/n7+R2hCCkpfMeLrY4OliQ2fcTKZl5xYaDwaJade4kpZLXFI2h04nc3OjqYuDZclVgottycHh+oHByU5eYF9TMnMK2P9XMvuikzh3OQsAv+aODBrSju7t3KV31x1I8heiHPY2OuxtdPi1cDIpLyo2kJyRe/1qIZcrade4kp7L7hOJ5BfeGOrEUqcxaUJq6lpycPB0tq7wk8jihtz8Yv48k8K+6CtEX8xAUaCFhx0P9fclxN8TV0fzvRO6PpHkL0QlaS3UeLnb4eVuelNNURSu5hRyJe0aieklB4cradeIib/K3pNJxvlUKnB3tMbL3ZbObd3o5udeJwcAqwuKivUci01jb3QSR8+mUaw34OZoxfCe3oT6e97yNxDlk+QvRDVTqVQ421vibG+J/9/eQlZQqC9pPkovaT5KTMvl3OUsDsek8uPG0wS0diXY34PObdywtmzc/z0NBoVTcRnsO5nEoTPJ5BXocbDR0q9zM3p08MSnmYM0p1VB4/52CVHLLHUavJvYm7yfWFEUzidms/9USXfEI2dT0VqoCfJ1JcTfkyBfVyy1jaN5qLQu9kZf4cCpZDKvFWKl09DNz53Qjp74ezvX+dEy6wtJ/kKYmUqlwqeZAz7NHHg4vA1nL2Vy4FQyB04nG59M7tSm5EAQ6OPS4O4TZGQXEH0hnb8uZhB9MYOM7AIsNCqCfN3o0aHk4KdrJAe/2iTJX4g6pGSYYCf8Wjjxj0FtOR1/lQOnkjh4OoX9p5KxttTQpa07If4edGhVP/ur5+QV8dfFDE5d/7mSnguAnbWW9i2dCPRxpVs7uf9R0yT5C1FHqdU3Bqr752A//rqYwf5Tyfx5JoU9J65ga2VBVz93Qvw9ae/tVGebQ/ILizkTn3n9zD6d+KQcFEqawNq1cKJvp2Z0aOVMcw87eS1nLZLkL0Q9YKFRE+DjSoCPK48PbceJ8+kcOJXE/r+S2XUsEXsbLd3beRDi70Hb5k5mfQq5qNjAucuZnLrejHP+chZ6g4KFRoVvM0dG9mlNB28XWjW1r5dXLg2FJH8h6hkLjZrObdzo3MaNwiI9x8+lsf9UMruPJ7L9cAKOdjqC23kQ0sET31roEWMwKFxMyjY248TEX6Ww2IBKBa2a2DMkpCX+3s60ae7YaG5c1weS/IWox3RaDd3aedCtnQcFhXqOxqayLzqJHUcus+XQJVwdLAlu70mwvwetmthXy4FAURQup+WWNONcSOd03FVyC0qGwPBys6VPp2Z08HamXUsnabevwyT5C9FAWOo0hPh7EuLvSV5BMYdjSm4Sbz4YT+T+ODycrAn29yDE35Pm7rZ3dSBIzczj1IUMTsWVnN1n5hQC4OZoRbd27sZ7E6XvSRB1nyR/IRoga0sLegU0pVdAU67lF/Hn6RT2n0piw944fvvjIk1dbQhuX3IgaOZ265uqsq4V8ldcBtEXMjh1MZ2Uq/kAONhoae/tTIdWLvh7O+PuZF3buyaqiYzn38hIfZhqbPWRlVvIodMpHDiVxOm4qyhAc3dbgv098fdx5cCJK5y6mM6llGsAWFtqaNfCGf9WJWf2Xm53d8VQn9X370Z54/lL8m9kpD5MNeb6yMgu4ODpZA6cSuZsQsm7a7UWato2d7zejOOCdxO7OtuFtKbV9+9GtbzM5cKFC7zzzjv8+eefWFpaMnz4cKZNm4a19Z0v+XJzc/nyyy+JjIwkJSUFT09P7rvvPp555hl0OhlqVQhzcra3ZHD3Fgzu3oK0zHwMGjVOVpoG9wSxuL1yk39WVhaPP/44zZo149NPPyU9PZ05c+aQnp7OJ598csdl/+///o8tW7bw8ssv07ZtW44dO8a8efPIyspi1qxZ1bYTQoiqcXW0qvdnuuLulJv8lyxZQlZWFqtWrcLFpWSEQo1Gw7Rp05g4cSJt27a97XLFxcVERkYyYcIEHnvsMQB69OjB5cuXWbdunSR/IYQwo3Ib83bu3EmPHj2MiR9gyJAh6HQ6du7cWeZyiqKg1+uxt7c3KXdwcKCe3GYQQogGq9zkHxsbS5s2bUzKdDodLVu25Ny5c2Uup9VqGTlyJD/++CNHjx7l2rVr7N27l2XLlvHII49UPXIhhBCVVqE2fwcHh1vKHRwcyMzMvOOyb7/9NrNnz+bhhx82lo0bN45JkybddaB3umtdHnd3+/JnakSkPkxJfdwgdWGqIddHjT7k9dFHHxEVFcW///1vWrVqxZEjR/jiiy9wc3Pj6aefvqt1SVfP6iH1YUrq4wapC1P1vT6q3NXTwcGBrKysW8qzsrLw8fEpc7kzZ87w7bff8uWXXzJw4EAAgoODKS4uZt68efzjH//Azk7euymEEOZQbpu/r68vsbGxJmWFhYXExcXdMfmfPXsWAH9/f5PyDh06UFhYSFJS0u0WE0IIUQvKTf59+/Zl7969ZGRkGMs2b95MYWEh/fr1K3M5Ly8vAE6ePGlSfuLECVQqFc2aNatszEIIIaqo3GafsWPHsnjxYiZOnMjEiRNJS0tj7ty5REREmPQCmjVrFqtWrSI6OhqAgIAAgoKCmD17NmlpaXh7e3Ps2DEWLFjAgw8+WO7TwUIIIWpOhdr8v//+e/7973/z4osvGod3mD59usl8BoMBvV5v/F2j0fD111/z6aefsmDBAlJTU2natCnjx4/n2Wefrf49EUIIUWEysFsjI/VhSurjBqkLU/W9Psrr7dM4h+sTQohGTpK/EEI0QpL8hRCiEZLkL4QQjZAkfyGEaIQk+QshRCMkyV8IIRohSf5CCNEISfIXQohGSJK/EEI0QpL8hRCiEZLkL4QQjVCNvsZRCCEqSikuwHD1CkpRPiqdNSqtFWitSj5rtOYOr8GR5C+EqDWKoqDkZWG4mnjTz2UMVxNRctLKXlCtQaW1Bp0VKu31A8P1A4TJZ50VaG/32brkQKK1Ao0WlUpVezsNKIoB9MVg0KPoi65/LkbRF4O+6KbPxWAouvFZrcaiZWdUFrpqj0mSvxCi2ikGPUp2Coariegzrif5zJJ/Kbh2Y0YLHWqnpmiatEXt1A+1U1NUOhuUonwoykMpzP/b5zwoykcpzEPJz0bJSkYpKp0nv2LBqTXGA4HxoGA8kJT+bkW6jSUF2ddKErGhJBkr1xM1+mKU6//e+Fx0m3n115O7vvy4ymA9/FUsvDpUevmySPIXQlSaUpSP4eoV49m78SczqSQJXqeydkTt1BStbyhqp6bGH5WtMypV9dx6VBQDFBVcPxjkwfUDh1J40wHj+kHixueSf5X8HJTs1JvKCyhEAbUFaCxKmp3UmpKrBk1JGerr5RZaVGrrknK1xfV5NKDWXl+2rPK/rfPmbZWuX2uF2s6lWurn7yT5CyHuqKSpJhNDxt8S/NVElGvpN2ZUqVE5eKBxaopFy04lCd65GWrHJqgsbWs8TpVKXdL8o7MGnKu0LkUx4O7uQGpqTvUEVwdJ8hdCAKDoi9HffAafcSPJU5R3Y0atVUlTTbP218/gm5X86+BRcgbbAKhU6lq/L1DbGsZfSoh6TCnKRykuvOkmYHk3BG8qN2l7LkIpbWP+W5vz7dqnb15W0ReRXZRv0jatsnUuaapp2wu1840kr7JxavCJsTGQ5C+EmRgyr1CwfznF5w9WzwpL24yv/4tGi+p6m/LNbcklXSdvbmMu+Wzr5ES+zvVGe7zOunriEnWSJH8hapkhP5vCQ6spit4OFlp0nSJQ2blcT9alidsC1fUbg6YJ/XblJTcOq3o27lLPX1gu7o4kfyFqiVJcSOGJTRQe/g2KC9C274eu2/2obRzNHZpohCqU/C9cuMA777zDn3/+iaWlJcOHD2fatGlYW5d/WZidnc28efPYuHEj6enpeHh4MHLkSKZMmVLl4IWoDxTFjy2isgAADTFJREFUQHHMHgoO/IpyLR0L7y7oQh9C49TM3KGJRqzc5J+VlcXjjz9Os2bN+PTTT0lPT2fOnDmkp6fzySef3HHZ3NxcHn30UVQqFdOnT8fDw4P4+HiuXLlSbTsgRF1WfOkEBfuWYkiLR+3eGqvwZ7Fo2s7cYQlRfvJfsmQJWVlZrFq1CheXkocNNBoN06ZNY+LEibRt27bMZRcsWEB2djZr167F1rakn29oaGg1hS5E3aVPi6dg31L0l06gsnfDKvw5LHxDqu2BJiGqqtxv4s6dO+nRo4cx8QMMGTIEnU7Hzp0777js8uXLGT16tDHxC9HQGa5lkLdjIbkr3kKfch7LHmOxfXgO2jY9JPGLOqXcM//Y2FgefPBBkzKdTkfLli05d+5cmctdunSJlJQUnJ2dee6559i9ezeWlpaEh4fz+uuv4+goN7lEw6EU5lF4dD2FxzaCYkAbNATLLvfWypOtQlRGhdr8HRwcbil3cHAgMzOzzOVSU1MB+OCDDwgPD2f+/PkkJCTw0UcfkZaWxsKFC6sQthB1g2IopuhUFIWHVqHkZ2Ph2wPL4AdRO7ibOzQh7qjGunoaDAYAvL29+fDDD419kO3t7ZkyZQrHjh0jKCiowutzdbWrdCzu7vaVXrYhkvowVZn6UBSF3DP7Sd+2mKL0y1i17IjrwMexbNamBiKsPfLdMNWQ66Pc5O/g4EBWVtYt5VlZWfj4+JS5XGmzTs+ePU0ePunZsycAMTExd5X809JyMBiUCs9fyl0eXDEh9WGqMvWhT46lYO9S9FfOoHZqivWQKWhadiZLpYJ6XLfy3TBV3+tDrVbd8aS53OTv6+tLbGysSVlhYSFxcXGMGjWqzOVatGiBTlf2CwgKCgrK27QQdYohK7lkOIZz+1FZO2AZ9gTa9n1LhlAQop4pN/n37duXr776ioyMDJydS4ZJ3bx5M4WFhfTr16/M5XQ6Hb1792bPnj0oimI8+9+9ezcAAQEB1RG/EDVOyc+h4M/VFEVvA7UGXdeR6IKGytg3ol4rt+/Z2LFjsbe3Z+LEiezatYtVq1bxzjvvEBERQZs2N9o3Z82aRYcOpm+bmTRpErGxsUydOpVdu3axdOlS/vWvfxEWFnZXTT5CmINSXEjBkfXkLJlO0cktaP16YzvmfSy7PyCJX9R7FWrz//777/n3v//Niy++aBzeYfr06SbzGQwG9HrTV5UFBATwzTff8NFHHzFx4kTs7OyIiIhg2rRp1bsXQlQjRTFQfHYvBQdWoOSkoWkRhGXoGDQuXuYOTYhqo1IU5e7vopqB3PCtHlIfpv5eH8UJ0SXDMaReRO3qjWWPMTXy/tS6SL4bpup7fVT5hq8QjYE+PaFkOIb4Y6jsXLEa8AwW8lSuaMAk+YtGrTg7g/ydP1B0ehdordCFPIwuYBAqi7J7qgnREEjyF42OUlyIIfUixXFHiT+5GUVfjLbjICy7jkRlVfmHCYWoTyT5iwZNMRgwXE1An3wOQ/J59CnnMaRfAqWkc4Ktf0/o9ABqBw8zRypE7ZLkLxoMRVFQslPRp5xHn3IOQ/I59KkXofj6A4U6azTuPug6DUPt4YPGwwdP7xb1+qaeEJUlyV/UW4a8LAz/v727DWmq7+MA/t2OW8lst+atxLzJfEAzsWl3oSJo6guhuoRE8EWxkDJQesBABYOgtOzhhYnSKEhJQyxiaAW96IEQIgyMzCRMZmrKzd0up53Vcm47536xOVv2dF237T93fp83zv/Z9OtBv//539k5Jtezeef7UQimtxDn3EXOBUEeHg3FxmxwETHgImIh+0ckvYBLiBuVP1kRRLsNzj/HIJhG4XQv34gWk3urDPKwKHDrU8FFxoCLjIU87F+QcfTrTcj30F8H8Tui4IBgnnKtz78fdS3hzEwB7rekyELCwUXGgtuUC3lELLh/RtM7bgn5i6j8A4goCoDTAQgOiE7HV7ftgODA3LwKTss8IA9yPTPmFAAXBJk8COCCXONfnIX192cWIfLv4fQ8ox+F8Oe4Ky8A2aoQyCNjoNzwb3CRMZBHxEIevPT6EoSQvyagy190OvBp+Bnss37wgp4geBex0wFRcBe00w5RcC6OO+2A+/PF+yzc3+76Gu5x0f0YCA7XY37C+itZ3ROBZ0Lw3P5qouAUrjNaLoy7Jw9wii9uL4x/cR+ZHMLsf9wvzL4FbJ9c35dTgovYAMWmPNc6fWQsZGsifDoZESIVAV3+jokX+O/9FtYxfm5J2X5VqgvblcHfKNVfKFvPfTio1cH4MGNZnHS+nkgWJp0l4w6vCUt02gH7HMS/OxnJ5JCvjYIiZqvryJuIGMjDouj0yIT4SECXvyJmKyLKmzFtWnoxGp+TyfximUUVsQZWHx3auLgM5fSaFETBAXlIOGRBq3ySgxCyVECXPwAo1mrAOQP3Umz+TCaTA+7TJMhAL8gS4k/ooGdCCJEgKn9CCJEgKn9CCJEgKn9CCJEgKn9CCJEgKn9CCJGgFXOop1z+94+F/38eG4hof3ij/bGI9oW3lbw/fpZ9xVzAnRBCyPKhZR9CCJEgKn9CCJEgKn9CCJEgKn9CCJEgKn9CCJEgKn9CCJEgKn9CCJEgKn9CCJEgKn9CCJGggCz/sbEx7N+/H2lpacjIyEBdXR0+f/7MOhYT9+7dQ0VFBXJycpCamoo//vgDnZ2dEASBdTTmPn36hOzsbCQmJmJwcJB1HGa6u7tRVFSEzZs3Iz09HaWlpTCbzaxjMfHgwQMUFxcjLS0NWVlZOHz4MMbGxljH+i1WzLl9fhXP89DpdNBoNGhqaoLZbEZDQwPMZjMaGxtZx/O5trY2aDQaVFdXIzw8HH19fTh9+jTevXuHmpoa1vGYamlpgdP5g4vMS4Ber8eVK1dw8OBB1NTUwGKxoK+vD3a7nXU0n3v69CkOHTqEwsJCVFZWgud5tLS0oLS0FHfu3EFISAjriMtLDDCXL18WtVqtOD097Rm7ffu2mJCQIL5584ZhMja+3A8Lzpw5I6akpIg2m41BIv8wPDwspqamil1dXWJCQoL48uVL1pF8zmg0ips2bRIfPXrEOopfqK2tFXNzc0VBEDxjAwMDYkJCgvj48WOGyX6PgFv26e3tRUZGBtauXesZKygogFKpRG9vL8NkbHy5HxYkJSXBZrNhdnaWQSL/cOrUKezZswcbNmxgHYUZg8EAjUaD3Nxc1lH8gsPhgEqlgky2eDbMNWvWMEz0ewVc+RuNRsTHx3uNKZVKrF+/HqOjo4xS+Zf+/n6EhoYiPDycdRQmuru7MT4+jvLyctZRmBoYGEBiYiIuXbqErKwsJCcno7i4GM+ePWMdjYndu3djdHQUHR0d4Hkek5OTOHfuHOLi4pCZmck63rILuPLneR5qtXrJuFqtxocPHxgk8i+Dg4MwGAzYt28fOI5jHcfnLBYLLly4gKqqKqhUKtZxmDKZTHjy5AkMBgOOHz8OvV6PkJAQHDhwAJOTk6zj+VxGRgaam5vR2NiIbdu2IT8/H1NTU2hra4NSqWQdb9kFXPmT7zOZTDhy5AhSUlJQVlbGOg4TFy9eRHR0NAoLC1lHYU4URVitVjQ3N2PHjh3Izs72TABXr15lHc/nnj9/jpqaGhQXF+PatWtoamqCTCZDeXk55ubmWMdbdgF3tI9arQbP80vGeZ5HbGwsg0T+wWKxoKysDKtXr4Zer4dCoWAdyedGRkbQ1dWF1tZWz++I1Wr1fPz48WPgHdHxA2q1GqGhoUhKSvKMBQcHQ6vVYmRkhGEyNurr65Geno7a2lrPWGpqKrZv346enh6UlJQwTLf8Aq784+LiYDQavcbm5+cxMTGBoqIiRqnYstlsKC8vx/T0NLq6uhAWFsY6EhPj4+NwOBzQ6XRLtul0OmzcuBE9PT0MkrERHx+PiYmJb26z2Ww+TsOe0WhEXl6e19i6desQFhb23f20kgVc+S/86zozM+Mpufv372N+fh45OTmM0/mew+HA0aNHMTw8jI6ODkRFRbGOxMyWLVvQ3t7uNfb69Ws0NDTg5MmTSE5OZpSMjdzcXBgMBgwNDXl+dqvVihcvXqCgoIBxOt/TaDQYGhryGpuamsLMzExA/t0E3DV8eZ7Hrl27EBUVhYqKCkxPT+Ps2bPIzMyU5Ju8Tpw4gRs3bqCqqgpbt2712hYfHy+pZY5v6evrg06nw61bt5CSksI6jk8JgoCSkhKYzWZUVlZCpVKhtbUVr169Qnd3N6Kjo1lH9Knr16+jrq4Oe/fuRX5+PmZnZ6HX62E2m3H37t2A+4854MofAN6+fYv6+nr09/dj1apV2LlzJ6qqqhAcHMw6ms/l5eVhamrqm9va29uRnp7u40T+RcrlDwBmsxnnz5/Hw4cPYbPZoNVqUV1dLcl9IYoibt68ic7OTkxMTEClUkGr1eLYsWOIi4tjHW/ZBWT5E0II+TE61JMQQiSIyp8QQiSIyp8QQiSIyp8QQiSIyp8QQiSIyp8QQiSIyp8QQiSIyp8QQiSIyp8QQiTofwamp1olVC3SAAAAAElFTkSuQmCC\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plt.title('GRU:accuracy and val_accuracy')\n",
        "plt.plot(history.history['accuracy'])\n",
        "plt.plot(history.history['val_accuracy'])\n",
        "plt.legend(['accuracy', 'val_accuracy'])\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 293
        },
        "id": "K5twqclIY9-b",
        "outputId": "bd193623-ddc2-4cb2-b2c6-4edf5995a900"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plt.title('CRU:loss and val_loss')\n",
        "plt.plot(history.history['loss'])\n",
        "plt.plot(history.history['val_loss'])\n",
        "plt.legend(['loss', 'val_loss'])\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 293
        },
        "id": "YEEv63EDXKw9",
        "outputId": "c3bb0d0b-e71f-4adb-c9c6-5b971776c51c"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.metrics import classification_report, confusion_matrix\n",
        "# load the Saved model from directory\n",
        "# model = load_model(path+\"Model.h5\")\n",
        "# predictions = model.predict(test_padded)\n",
        "# y_pred = np.argmax(predictions, axis=1)\n",
        "\n",
        "# cm = confusion_matrix(testing_label_seq, y_pred) \n",
        "\n",
        "# # Transform to df for easier plotting\n",
        "# cm_df = pd.DataFrame(cm,\n",
        "                     \n",
        "#                      index = ['negative' , 'neutral' , 'positive'], \n",
        "#                      columns = ['negative' , 'neutral' , 'positive'])\n",
        "\n",
        "# plt.figure(figsize=(8,6))\n",
        "# sns.heatmap(cm_df, annot=True,cmap=\"YlGnBu\", fmt='g')\n",
        "# plt.title('GRU \\nAccuracy: {0:.2f}'.format(accuracy_score(testing_label_seq, y_pred)*100))\n",
        "# plt.ylabel('True label')\n",
        "# plt.xlabel('Predicted label')\n",
        "# plt.xticks(rotation = 45)\n",
        "# plt.yticks(rotation = 45)\n",
        "# plt.show()"
      ],
      "metadata": {
        "id": "j50GaANNRZh7"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print('GRU')\n",
        "print('Accuracy: {0:.2f}'.format(np.max(history.history['accuracy'])*100))\n",
        "print('loss: {0:.2f}'.format(np.min(history.history['loss'])))\n",
        "print('val_accuracy: {0:.2f}'.format(np.max(history.history['val_accuracy'])*100))\n",
        "print('val_loss: {0:.2f}'.format(np.min(history.history['val_loss'])))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "bBBxbyFk4POA",
        "outputId": "ee8b820f-073a-4978-d196-9a64807e6abd"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "GRU\n",
            "Accuracy: 88.36\n",
            "loss: 0.28\n",
            "val_accuracy: 68.00\n",
            "val_loss: 0.83\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "for i in history.history['accuracy']:\n",
        "  print(f'{i:.4f} ')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZcJshrDHqwaR",
        "outputId": "e661fb30-d7a4-4da4-9cb2-015c478af00f"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "0.5332 \n",
            "0.5544 \n",
            "0.4869 \n",
            "0.5557 \n",
            "0.5594 \n",
            "0.7121 \n",
            "0.8073 \n",
            "0.8385 \n",
            "0.8648 \n",
            "0.8836 \n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "for i in history.history['loss']:\n",
        "  print(f'{i:.4f} ')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "K9nBgKWQq3l3",
        "outputId": "ad564620-19e4-4da6-ef86-158848fa9a03"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1.7213 \n",
            "1.4237 \n",
            "1.0905 \n",
            "0.9849 \n",
            "0.8842 \n",
            "0.7234 \n",
            "0.5442 \n",
            "0.4470 \n",
            "0.3435 \n",
            "0.2778 \n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "for i in history.history['val_accuracy']:\n",
        "  print(f'{i:.4f} ')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9Bxu_gfAq5co",
        "outputId": "36b878a3-c47c-4da9-e43a-31cc7f4281de"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "0.5000 \n",
            "0.5000 \n",
            "0.5050 \n",
            "0.5000 \n",
            "0.5400 \n",
            "0.6450 \n",
            "0.6500 \n",
            "0.6800 \n",
            "0.6700 \n",
            "0.6750 \n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "for i in history.history['val_loss']:\n",
        "  print(f'{i:.4f} ')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "2iqxmdQKq7sW",
        "outputId": "b66c826e-ba70-49d0-b158-9495d68a5c12"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1.6145 \n",
            "1.2768 \n",
            "1.0460 \n",
            "1.0187 \n",
            "0.9265 \n",
            "0.8968 \n",
            "0.8319 \n",
            "0.8815 \n",
            "0.9485 \n",
            "1.0058 \n"
          ]
        }
      ]
    }
  ]
}