News_Classification_Iqbal_Maulana.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "News_Classification_Iqbal_Maulana.ipynb",
      "provenance": [],
      "toc_visible": true,
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/IqbalLx/Machine-Learning-Deep-Learning-Project/blob/master/News_Classification_Iqbal_Maulana.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LmrQTLfdntwg"
      },
      "source": [
        "# News Classification\n",
        "---"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "K2AZJzddoBj2"
      },
      "source": [
        "##Prepare dataset from Kaggle"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Q4n5ukNAoyb0"
      },
      "source": [
        "from google.colab import files\n",
        "files.upload() #Upload kaggle API"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "y_7j8D3ipGd0",
        "outputId": "36c4eb7d-c48e-4318-cc42-12f1c1252cd0",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 68
        }
      },
      "source": [
        "!pip install -q kaggle                                                           # Installing required kaggle library\n",
        "!mkdir ~/.kaggle                                                                 # Create new Kaggle directory to store our API key in root\n",
        "!cp kaggle.json ~/.kaggle                                                        # Copy the API to root folder\n",
        "!chmod 600 ~/.kaggle/kaggle.json                                                 # Give access\n",
        "!kaggle datasets download -d amananandrai/ag-news-classification-dataset          # Download dataset"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Downloading ag-news-classification-dataset.zip to /content\n",
            " 79% 9.00M/11.4M [00:00<00:00, 15.9MB/s]\n",
            "100% 11.4M/11.4M [00:00<00:00, 16.7MB/s]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "08If26IfpWl1",
        "outputId": "e251c34e-f270-4adb-ce9a-21b2bec0902c",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 68
        }
      },
      "source": [
        "!unzip ag-news-classification-dataset.zip"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Archive:  ag-news-classification-dataset.zip\n",
            "  inflating: test.csv                \n",
            "  inflating: train.csv               \n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ck3wqgRapchj"
      },
      "source": [
        "## Pre-process Dataset"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fOrBLUsZpa1U"
      },
      "source": [
        "import pandas as pd\n",
        "import numpy as np\n",
        "from sklearn.model_selection import train_test_split"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Ffi--PFwpp4L",
        "outputId": "42ea52cf-8a2a-4fc6-af13-63505a670cbe",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        }
      },
      "source": [
        "train_df = pd.read_csv(\"train.csv\")\n",
        "test_df = pd.read_csv(\"test.csv\")\n",
        "\n",
        "news = pd.concat([train_df, test_df], axis=0)\n",
        "news.head()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Class Index</th>\n",
              "      <th>Title</th>\n",
              "      <th>Description</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>3</td>\n",
              "      <td>Wall St. Bears Claw Back Into the Black (Reuters)</td>\n",
              "      <td>Reuters - Short-sellers, Wall Street's dwindli...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>3</td>\n",
              "      <td>Carlyle Looks Toward Commercial Aerospace (Reu...</td>\n",
              "      <td>Reuters - Private investment firm Carlyle Grou...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>3</td>\n",
              "      <td>Oil and Economy Cloud Stocks' Outlook (Reuters)</td>\n",
              "      <td>Reuters - Soaring crude prices plus worries\\ab...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>3</td>\n",
              "      <td>Iraq Halts Oil Exports from Main Southern Pipe...</td>\n",
              "      <td>Reuters - Authorities have halted oil export\\f...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>3</td>\n",
              "      <td>Oil prices soar to all-time record, posing new...</td>\n",
              "      <td>AFP - Tearaway world oil prices, toppling reco...</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "   Class Index  ...                                        Description\n",
              "0            3  ...  Reuters - Short-sellers, Wall Street's dwindli...\n",
              "1            3  ...  Reuters - Private investment firm Carlyle Grou...\n",
              "2            3  ...  Reuters - Soaring crude prices plus worries\\ab...\n",
              "3            3  ...  Reuters - Authorities have halted oil export\\f...\n",
              "4            3  ...  AFP - Tearaway world oil prices, toppling reco...\n",
              "\n",
              "[5 rows x 3 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_VmwFVXepuAC",
        "outputId": "09024782-9e49-47f9-b977-9aa919e0d1d6",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        }
      },
      "source": [
        "news['Description'] = news['Title'] + ' ' + news['Description']\n",
        "news.drop('Title', axis=1, inplace=True)\n",
        "news.head()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Class Index</th>\n",
              "      <th>Description</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>3</td>\n",
              "      <td>Wall St. Bears Claw Back Into the Black (Reute...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>3</td>\n",
              "      <td>Carlyle Looks Toward Commercial Aerospace (Reu...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>3</td>\n",
              "      <td>Oil and Economy Cloud Stocks' Outlook (Reuters...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>3</td>\n",
              "      <td>Iraq Halts Oil Exports from Main Southern Pipe...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>3</td>\n",
              "      <td>Oil prices soar to all-time record, posing new...</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "   Class Index                                        Description\n",
              "0            3  Wall St. Bears Claw Back Into the Black (Reute...\n",
              "1            3  Carlyle Looks Toward Commercial Aerospace (Reu...\n",
              "2            3  Oil and Economy Cloud Stocks' Outlook (Reuters...\n",
              "3            3  Iraq Halts Oil Exports from Main Southern Pipe...\n",
              "4            3  Oil prices soar to all-time record, posing new..."
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "CB0KBez5BIPf"
      },
      "source": [
        "### EDA"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8cAxI8YqBKOy"
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "\n",
        "plt.style.use('seaborn')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "itgzOCRKJl8w",
        "outputId": "c801ae42-3376-46d3-db16-12c06e569816",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 170
        }
      },
      "source": [
        "news.info()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "Int64Index: 127600 entries, 0 to 7599\n",
            "Data columns (total 2 columns):\n",
            " #   Column       Non-Null Count   Dtype \n",
            "---  ------       --------------   ----- \n",
            " 0   Class Index  127600 non-null  int64 \n",
            " 1   Description  127600 non-null  object\n",
            "dtypes: int64(1), object(1)\n",
            "memory usage: 2.9+ MB\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jWmq2QpDLuy6",
        "outputId": "ceac5c63-baf7-421f-d02b-7e8ead30c5d5",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "news['Class Index'].unique()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([3, 4, 2, 1])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Yfa7mkabBMpg"
      },
      "source": [
        "news['length'] = news['Description'].apply(len)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bz1Bgv3YBY-Z",
        "outputId": "c5537071-90f2-4174-fb2c-d5b0643b16ea",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 378
        }
      },
      "source": [
        "plt.hist(news[news['Class Index']==1]['length'], bins=30, color='red', alpha=0.3, label='World')\n",
        "plt.hist(news[news['Class Index']==2]['length'], bins=30, color='green', alpha=0.3,label='Sports')\n",
        "plt.hist(news[news['Class Index']==3]['length'], bins=30, color='blue', alpha=0.3,label='Business')\n",
        "plt.hist(news[news['Class Index']==4]['length'], bins=30, color='yellow', alpha=0.3,label='Science')\n",
        "\n",
        "plt.xlabel(\"Desc Length\")\n",
        "plt.ylabel(\"Frequency\")\n",
        "\n",
        "plt.legend()\n",
        "plt.plot()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 11
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfgAAAFYCAYAAAC/NO6RAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3deXhTZf7//2fSJHShhbY2KAyCgIJCWQqCgFUZLZvzUVSK0B/oOCgiRdGpbB0UXEEQLlxwQHFBFEGro+hHWj7OAINaqhBFUBkHEESWNoVCoVvSNr8/+BopLZC0TVtOX4/rmmuS+yx5592D79z3Oec+Jo/H40FEREQMxVzfAYiIiEjtU4EXERExIBV4ERERA1KBFxERMSAVeBEREQNSgRcRETEgS30HUJuczuN+bxMZGUpeXmEAomlclMeaUw5rh/JYc8ph7aiLPMbEhJ9xWaPvwVssQfUdgiEojzWnHNYO5bHmlMPaUd95bPQFXkRExIhU4EVERAxIBV5ERMSAVOBFREQMSAVeRETEgFTgRUREDEgFXkRExIBU4EVEpFG6447b2b//V+/70aMTycz83Pt++vSHycrKPOs+HI7NzJgxpVL7jBlTyMrKqr1gq8FQM9mJiMj5yZaxplb35xo05JzrxMX14ttvHbRq9QeOHj1KUVER3377DX37Xg3ADz9s59FHn6jVuOqSCryIiDRKPXr04osv/s2NN97Ed999y6BBQ/nuu28B2LPnZ1q2bMmXX37OqlVvExQURMeOl/Pggw/z6qtLOHBgPwcPHuAvfxnn3d/bby/js88yuPDCiygoKKivr+WlIXoREWmUevSI8xb0rVu/oVev3pSVlVFSUsy33zro3LkrL7+8iIULX+Lvf3+VAwf243BsBqC01M1LLy3FbD5ZRo8fP84//pHG4sWv88gjj7N79656+16/UYEXEZFGKSKiGSEhITidOfzww3Y6d+7CFVd05vvvt/Pdd98SHX0Bf/jDxYSGhgLQo0dPfvppBwCXX965wr7279/HJZe0o0mTJoSGhtGx4+V1/n1OpyF6qRMZGf49dGHQoLIARSIi8ru4uF5kZWViMplo0iSYrl27s23bVn744XtuvTWRr7/+/UK50lI3TZo0AcBqtVbYj8fjwWQyn/K+vG6+wFkErAdfUFDAxIkTGTNmDCNHjmTjxo3s2LGDkSNHMnLkSGbOnOldd+nSpQwfPpzExEQ2bNgAnBzuGDduHKNGjWLs2LEcPXo0UKGKiEgj1aNHLz766AO6dIkFoGvX7nz55edccMEFtG9/Kb/++guFhSfPp3/zjYOOHa+ocj+tWv2BvXt/xu12U1Bwgv/858c6+w5nErAe/D/+8Q8uueQSUlJSyM7O5s477yQmJobU1FS6du1KSkoKGzZsoF27dnz66aesXLmSEydOkJSUxNVXX82yZcvo3bs3d999N6tWreKVV15h8uTJgQpXREQaoe7d4/jb3yZz551/ASAyMor8/GPccMMgQkJCSE6eRErK/ZhMZrp27U63bt3ZvLny7W8REc0YMuRP3HvvXbRs2YpOnTpXWqeuBazAR0ZG8p///AeA/Px8mjdvzv79++natSsAAwYMIDMzE6fTSXx8PDabjaioKFq1asXOnTvJzMzk6aef9q47fvz4QIUqIiL1zJfb2gKhadOmbNhQsWC/884H3tfXXvtHrr32jxWWjx17r/d1XFwv4uJ6AfDnP9/Nn/98t3dZTEw4TufxQITtk4AV+BtvvJEPPviAhIQE8vPz+fvf/87jjz/uXR4dHY3T6aR58+ZERUV526OionA6neTm5nrbo6OjycnJOednRkaGYrH4d64XTv4RpObOlsdmzfzdVw2DOU/pWKwdymPNKYe1oz7zGLAC/9FHH9GyZUteffVVduzYQXJyMuHhv39Rj8dT5XZVtZ9p3dPl5RX6HWd9/8IyinPl8dgx/354OZ2N7yI7HYu1Q3msOeWwdtRFHs/2AyJgF9k5HA6uvvrkbECdOnWipKSEvLw87/Ls7Gzsdjt2u53c3Nwq251OZ4U2ERER8U3ACnybNm3YunUrAPv37ycsLIz27duzefPJSQLWrl1LfHw8V111FevXr8flcpGdnU1OTg4dOnSgf//+pKenV1hXREREfBOwIfrbb7+d1NRURo8eTWlpKbNmzSImJoZHH32U8vJyunXrRr9+/QAYMWIEo0ePxmQyMWvWLMxmM2PGjGHy5MkkJSURERHBvHnzAhWqiIiI4Zg8vp7gPg9U51yHzjXVjnPlURPdnJuOxdqhPNacclg76vscvGayExGRRuv9998lI+NTbDYbJSXFjBuXzJVX9qnWvgoKTvD999vp3fuqWo6yelTgRUSk3mXsqd3HxQ5qe+776g8ePMDHH3/I0qVvYrFY2LfvF5555slqF/j//GcHX321SQVeRESkPp04cQKXqwS3243FYqF164t58cWXmThxHJdf3pkdO36gpKSExx+fzYUXXsRLLz3Htm1bKS0t47bbRjB48I1MnDiOdu3aA+BwbKGwsIDWrS/moota8vrrSwgKshIZGcXMmU9isdRtyVWBFxGRRunSSy/j8ss7k5h4E3379ueqq/pz7bUDgJNTz77wwhLS0lby7rsruOaaAezevYu///01ioqKuPPOkVxzzXUAtGvXnmHDhvPppx+ze/cubr75VqZOfYhp06bRpk1HNmz4F8eOHSU6+oI6/X4q8CIi0mg98sjj7NnzM199lcmKFW/y4YdpAFx5ZW8AunTpyqZNX7Jjxw907x4HQEhICG3btmPfvn0AXH55l0r7HTDgBmbOnMkf/ziQG24YVOfFHfQ8eBERaaQ8Hg8lJSW0bXsJI0Yk8fLLb+B05pCdfYjy8nLvOiaTCZPJxKn3nJWWujGbTQBYrZX7yoMH38ibb75Js2bNmTr1Ifbu3VMXX6kCFXgREWmUPvnkI+bOfco7HXpBwQnKy8tp3jySrVu/BWD79m20bduOTp068803WwAoLCxk//5f+cMfLq6wP5PJRFnZyVt833hjKRaLhZtvvpXrrx/Inj276/CbnaQhehERaZSGDv0f9u7dw7hxdxISEkppaSkPPjiZFSveJDv7EH/96/2cOHGcp56aS0yMnY4dO5GcfA+lpaWMHz+RkJCQCvvr2LETixe/QEyMnRYtLuSuu+4iODiM8PBwRo4cXeffTxPdaEKHWqGJbmpOx2LtUB5rrrHncOLEcfz1r1No165DjfajiW7kvHG2+1SbHQnhWH5RhTZf7kMVEZHAUIEXERE5xYsvvlzfIdQKXWQnIiJiQCrwIiIiBqQCLyIiYkAq8CIiIgaki+xERKRROnjwAHfcMZKOHTthMplwuVxMmDCJbt26+7T94cO5vPrqEqZM+VuAI60eFXgREal3/s6VcS6+zqVx8cVtvFfNf/utg2XLlrJgwYs+bRsdfUGDLe6gAi8iIgLAkSNHuOCCGJ56ahbXXXc9/fvH88UXG1m//p9MnTqDxx9/hMOHc3G5XIwdey9t2rRlxoypvPrqcm6/fRg333wrX3yxEZfLxXPPvURZWSizZz/OgQP7KS0t5e67x9Oz55WsWfMJH3zwLhaLlQ4dLiMlZWqVbTWlAi8iIo3WL7/sZeLEcbhcLnJzncyf/wIrVrxZab1du3Zy7NhRFi16hePHj5OZ+UWF5WVlZVx8cVuSku5g5szpbN78Nd98U0Z09AVMn/4oR48eZdKk8SxbtpKVK99i7tyFtGhxIf/7v6spKSmusq1Jk+AafTcVeBERabROHaLfu3cPjzwylQ4dLqu0Xps2bSksLOCJJx7hmmsGcMMNA8nOPlRhnW7degAQE9OCgoIT7Nq1g6ysr/juu5MPrikpKcHtdnPDDYNITZ3MoEFDuOGGQTRpElxlW03pKnoRERFOFvEmTZrgdOZ420pLSwEIDg5myZI3uOmm28jM/II5c56otH1Q0O/XEXg8HqxWK3fc8RdefPFlXnzxZVau/AdWq5UxY+7iqafmUV5ezgMP3MexY0erbKspFXgREREgP/8Yhw8fpk2bthw+nAvg7X3/5z87+L//S6dbt+48/PB09uz5+Zz769atG59/vgGAvLwjLFmyiPLycpYsWcQFF1zAyJGj6dIllkOHDlXZVlMaohcRkUbrt3PwAC6Xi4cemkxMjJ3HHpvB+vX/4tJLTw7XX3RRS5YsWcRHH32A2WwmKWnMOfc9ZMgQ1q/fyPjxf6GsrIy//GUcZrOZ0NAw7r33Lpo2bUrLlq249NLL+OqrTZXaakqPi23kj0X0x1mfJhdx9qfJ6XGx56ZjsXYojzWnHNYOPS5WDOvUHwSOnIvOum6cvWegwxERaVR0Dl5ERMSAAtaDf++991i9erX3/fbt23nnnXeYNWsWAB07duSxxx4DYOnSpaSnp2MymZg4cSLXXnstx48fJyUlhePHjxMaGsr8+fNp3rx5oMIVERExlIAV+MTERBITEwH46quvWLNmDU899RSpqal07dqVlJQUNmzYQLt27fj0009ZuXIlJ06cICkpiauvvpply5bRu3dv7r77blatWsUrr7zC5MmTAxWuiIiIodTJEP2iRYu455572L9/P127dgVgwIABZGZmkpWVRXx8PDabjaioKFq1asXOnTvJzMwkISGhwroiIiLim4BfZPfdd99x0UUXERQUREREhLc9Ojoap9NJ8+bNiYqK8rZHRUXhdDrJzc31tkdHR5OTk1Np36eLjAzFYvH/gQVnuwpRftfsSMjZl0eceXlYSJOzb9sstML7mBjf4zISHYu1Q3msOeWwdtRnHgNe4NPS0rjlllsqtZ/p7ryq2n29ky8vr9C/4NDtIP44/Ta4U1V1m9ypCopKzr7vYxX/dk6nbpOT6lEea64x5fD9998lI+NTbDYbJSXFjBuXzJVX9qm03nPPzScxcSQtW7byed+Gv00uKyuLGTNmYDKZOHr096n3srOzsdvt2O12fv755yrbnU4n4eHh3jYRETEmm+3M82xUh8s15JzrHDx4gI8//pClS9/EYrGwb98vPPPMk1UW+EmTUmo1vroQ0HPw2dnZhIWFYbPZsFqttGvXjs2bNwOwdu1a4uPjueqqq1i/fj0ul4vs7GxycnLo0KED/fv3Jz09vcK6IiIiteXEiRO4XCcfAAPQuvXFvPjiy/z00w7uvfcu7rvvLyxa9BwAEyeOY/funRQWFjBjxhQmTbqPiRPHsXPnfwG4/fZhrFjxJsnJ93DPPXdSWFiA2+1m1qy/cd99f2HSpPtwOnMoKytj9uzHuf/+e7nvvrFs2fJ1wL5fQHvwTqezwvn11NRUHn30UcrLy+nWrRv9+vUDYMSIEYwePRqTycSsWbMwm82MGTOGyZMnk5SUREREBPPmzQtkqCIi0shceullXH55ZxITb6Jv3/5cdVV/rr12AAsXPsvkyal06HApTzzxKIcOHfRu8+6779CnTz/+53+G8fPPu3nuuWdZuPClKh8XW15eTHR0NLNmPcVnn2Xw+ef/JiQkpMpHyAZCQAt8ly5dWLp0qfd9hw4dWLFiRaX1xowZw5gxFef1DQsL46WXXgpkeCIi0sg98sjj7NnzM199lcmKFW/y4Ydp/PLLHjp0uNS7/FTbtn3H0aN5ZGR8CkBJSbF32emPi929+z/ExnYH4IYbBgHw7LOz2br1m0qPkLVarbX+3TRVrYiINEoejweXy0XbtpfQtu0l3Hbb7fx//9/wCteLnc5qtfDQQ5Pp0qVrpWWnPy42KCiI8vKKF4lbLCcfIZuQMLj2vsgZaKpaERFplD755CPmzn3Ke6dWQcEJysvL6dGjJ99/vx2A2bMfr/Bo2Cuu6MK//70egJ9/3s3KlW+dcf+xsbE4HCfPsX/xxUbefPM1rriiS6VHyAaKevAiItIoDR36P+zdu4dx4+4kJCSU0tJSHnxwMi1aXMizz84GoHPnWNq2vcS7zfDht/PUU7OYMOFuysvLefDBh8+y/6H8618bmDhxHEFBFmbMmEVkZBQOx9cVHiEbKHpcbCO637Om/H1c7KkcG/17mpweFyvVpTzWnHJYO+r7PngN0YuIiBiQhuilTlzR7ouzLm/VLLvCe5ut3KeJKkREpGrqwYuIiBiQCryIiIgBaYheGiSHw8z+/f49GbAxXpgnInIm6sGLiIgYkHrw0iAF7dqJ5YctPq1bGtfz3CuJiDQy6sGLiIgYkAq8iIiIAWmIXqrl9JnpwkKaUFBUcsb1r2i3O9AhiYjIKdSDFxERMSAVeBEREQNSgRcRETEgFXgREREDUoEXERExIBV4ERERA1KBFxERMSAVeBEREQNSgRcRETEgFXgREREDUoEXERExIBV4ERERAwrow2ZWr17N0qVLsVgsPPDAA3Ts2JEpU6ZQVlZGTEwM8+bNw2azsXr1apYtW4bZbGbEiBEkJibidruZNm0aBw4cICgoiNmzZ9O6detAhisiImIYAevB5+XlsWjRIlasWMHixYv55z//yfPPP09SUhIrVqygTZs2pKWlUVhYyKJFi3jjjTdYvnw5y5Yt4+jRo3zyySdERETwzjvvMH78eObPnx+oUEVERAwnYAU+MzOTvn370rRpU+x2O0888QRZWVlcf/31AAwYMIDMzEy2bt1KbGws4eHhBAcHExcXh8PhIDMzk4SEBAD69euHw+EIVKgiIiKGE7Ah+l9//ZXi4mLGjx9Pfn4+999/P0VFRdhsNgCio6NxOp3k5uYSFRXl3S4qKqpSu9lsxmQy4XK5vNtXJTIyFIslyO9YY2LC/d6mMWp2JMT7OiykSaXlVbX9JthmPeu+w8JO2zbYWrntjIGFAhAT49vqDZmOxdqhPNacclg76jOPAT0Hf/ToUV588UUOHDjAHXfcgcfj8S479fWp/G0/VV5eod8xxsSE43Qe93u7xuhYfpH3dUFRSYVlYSFNKrWdqtjlPuu+CwoqbhtU7K7Udialx07+3Z3OMp/Wb6h0LNYO5bHmlMPaURd5PNsPiIAN0UdHR9OjRw8sFgsXX3wxYWFhhIWFUVxcDEB2djZ2ux273U5ubq53u5ycHG+70+kEwO124/F4ztp7FxERkd8FrMBfffXVbNq0ifLycvLy8igsLKRfv35kZGQAsHbtWuLj4+nWrRvbtm0jPz+fgoICHA4HvXr1on///qSnpwOwbt06+vTpE6hQRUREDCdgQ/QtWrRg0KBBjBgxAoAZM2YQGxvL1KlTWbVqFS1btmTYsGFYrVZSUlIYO3YsJpOJ5ORkwsPDGTp0KF9++SWjRo3CZrMxZ86cQIUqdcR88NAZlwXlnnGRiIhUQ0DPwY8cOZKRI0dWaHv99dcrrTd48GAGDx5coe23e99FRETEf5rJTkRExIBU4EVERAxIBV5ERMSAVOBFREQMSAVeRETEgAJ6Fb0Y1xXtvqjwPthmPedsdSIiUnfUgxcRETEgFXgRERED0hC9NAg73ZVnufuxZFeV6/Zs0j7Q4YiInPfUgxcRETEgFXgREREDUoEXERExIBV4ERERA1KBFxERMSBdRS8N1uVdvq2yvaW14hX3Za2y/9+rQQGOSETk/KEevIiIiAGpwIuIiBiQCryIiIgBqcCLiIgYkAq8iIiIAanAi4iIGJAKvIiIiAGpwIuIiBiQCryIiIgBqcCLiIgYkAq8iIiIAQVsLvqsrCwmTZrEpZdeCsBll13G3XffzZQpUygrKyMmJoZ58+Zhs9lYvXo1y5Ytw2w2M2LECBITE3G73UybNo0DBw4QFBTE7Nmzad26daDCFRERMZSAPmymd+/ePP/8897306dPJykpiSFDhrBgwQLS0tIYNmwYixYtIi0tDavVyvDhw0lISGDdunVEREQwf/58Pv/8c+bPn8/ChQsDGa6IiIhh1OkQfVZWFtdffz0AAwYMIDMzk61btxIbG0t4eDjBwcHExcXhcDjIzMwkISEBgH79+uFwOOoyVBERkfNaQHvwO3fuZPz48Rw7doyJEydSVFSEzWYDIDo6GqfTSW5uLlFRUd5toqKiKrWbzWZMJhMul8u7vYiIiJxZwAp827ZtmThxIkOGDGHfvn3ccccdlJWVeZd7PJ4qt/O3/VSRkaFYLEF+xxoTE+73No1RsyMhv79xWistD7ZVbqvAWjsDRsHBp31OWBPAGH9HI3yHhkB5rDnlsHbUZx59KvAejweTyeTXjlu0aMHQoUMBuPjii7ngggvYtm0bxcXFBAcHk52djd1ux263k5ub690uJyeH7t27Y7fbcTqddOrUCbfbjcfjOWfvPS+v0K8Y4WTync7jfm/XGB3LL/K+Lna5KywLtlkrtZ3O7C6vlTiKqfg5ZQUlAOf931HHYu1QHmtOOawddZHHs/2A8KnADxgwgJtvvpnhw4f7fCX76tWrcTqdjB07FqfTyeHDh7n11lvJyMjg5ptvZu3atcTHx9OtWzdmzJhBfn4+QUFBOBwOUlNTOXHiBOnp6cTHx7Nu3Tr69Onj27cVv2TsWVPfIYiISAD4VODfe+89MjIySE1NxWKxcOuttzJo0KCz9qj/+Mc/8vDDD/PPf/4Tt9vNrFmzuPzyy5k6dSqrVq2iZcuWDBs2DKvVSkpKCmPHjsVkMpGcnEx4eDhDhw7lyy+/ZNSoUdhsNubMmVNrX1pERMToTB5fTm6fYu/evUyfPp1du3YxcuRIJkyYQJMmTQIVn1+qMxTS2IeiqtuDL96/2/vafPAQNqsZVy0NwZ9LB+uFFd6Xte8AQJcug+rk8wOlsR+LtUV5rDnlsHbU9xC9z1c9ff3110yfPp177rmHuLg4VqxYQUREBJMmTaqVIEVERKT2+DREn5CQQKtWrRgxYgSPP/44VuvJq5jbt2/PZ599FtAARURExH8+FfilS5fi8Xho27YtAD/88ANXXHEFACtWrAhYcCIiIlI9Pg3Rf/DBByxZssT7/uWXX+bZZ58F8Pv2OREREQk8nwp8VlYWs2fP9r5fuHAhW7ZsCVhQIiIiUjM+FXi3243L5fK+LygooLS0NGBBiYiISM34dA5+5MiRDB06lC5dulBeXs62bduYOHFioGMTERGRavKpwCcmJtK/f3+2bduGyWRi+vTpXHTRRYGOTURERKrJpwJfUlLCDz/8wIkTJ/B4PHzxxRcADB8+PKDBiYiISPX4VODHjh2L2WymVatWFdpV4EVERBomnwp8aWkpK1euDHQsIiIiUkt8uoq+Q4cO5OXlBToWERERqSU+9eAPHTrEwIEDad++PUFBQd72t99+O2CBiYiISPX5VODHjRsX6DhERESkFvk0RN+7d28KCwv56aef6N27NxdeeCFXXnlloGMTERGRavKpwM+bN4+0tDQ++OADAD7++GOefPLJgAYmIiIi1efTEP3XX3/Nu+++y5gxYwBITk5m5MiRAQ1M6larZv49W2DX/sgAReK/oF07AbDtLz/nuq5BQwIdjohIg+BTD75JkybA70+OKysro6ysLHBRiYiISI341IOPi4tj+vTp5OTk8Prrr7N27Vp69+4d6NhERESkmnwq8A899BDp6ekEBwdz6NAh7rrrLgYOHBjo2ERERKSafCrw+/bto3PnznTu3LlCW+vWrQMWmIiIiFSfTwX+zjvv9J5/d7lcHDlyhEsvvZQPP/wwoMGJiIhI9fhU4P/1r39VeP/f//6XtLS0gAQkIiIiNefTVfSnu/TSS/n+++9rOxYRERGpJT714J977rkK7w8dOkR+fn5AAhIREZGa86kHHxQUVOF/HTt25JVXXgl0bCIiIlJNPvXgJ0yYUGV7efnJmcPM5qp/JxQXF/OnP/2JCRMm0LdvX6ZMmUJZWRkxMTHMmzcPm83G6tWrWbZsGWazmREjRpCYmIjb7WbatGkcOHCAoKAgZs+erSv2RURE/OBTge/atWuVM9d5PB5MJhM//vhjldv9/e9/p1mzZgA8//zzJCUlMWTIEBYsWEBaWhrDhg1j0aJFpKWlYbVaGT58OAkJCaxbt46IiAjmz5/P559/zvz581m4cGENvqaIiEjj4lOBT05OpkOHDvTv3x+TycS6devYs2fPGXv2ALt27WLnzp1cd911AGRlZfHYY48BMGDAAF577TUuueQSYmNjCQ8PB07OmOdwOMjMzGTYsGEA9OvXj9TU1Jp8RxERkUbHp3PwmzZtIiEhgdDQUEJCQhg6dChZWVln3eaZZ55h2rRp3vdFRUXYbDYAoqOjcTqd5ObmEhUV5V0nKiqqUrvZbMZkMuFyufz+ciIiIo2VTz34o0ePsmHDBnr16gXA5s2bOXLkyBnX//DDD+nevfsZz5t7PJ5aaT9dZGQoFkuQT+ueKiYm3O9tjKLZkRAAwkJtfm0XbLP+/sZ68neizVqtuy79FhxsrbK9WbOQc2/cwP/WjflYrE3KY80ph7WjPvPoU4F/4oknmDNnDg899BAAl112GTNnzjzj+uvXr2ffvn2sX7+eQ4cOYbPZCA0Npbi4mODgYLKzs7Hb7djtdnJzc73b5eTk0L17d+x2O06nk06dOuF2u/F4PN7e/9nk5RX68nUqiIkJx+k87vd2RnEsvwiApib/RkiKXW7va7O7HJvVjMt97se11oZi3FW2H6PonNu6GvDfurEfi7VFeaw55bB21EUez/YDwueL7FasWOG9qO5cTr0g7oUXXqBVq1Z88803ZGRkcPPNN7N27Vri4+Pp1q0bM2bMID8/n6CgIBwOB6mpqZw4cYL09HTi4+NZt24dffr08SVMERER+X98GlPdsWMHt956K0OGDAHgpZdeYuvWrX590P3338+HH35IUlISR48eZdiwYQQHB5OSksLYsWO56667SE5OJjw8nKFDh1JeXs6oUaN4++23SUlJ8f+biYiINGI+9eAff/xxnn76aZ566ikAhgwZwvTp01m5cuU5t73//vu9r19//fVKywcPHszgwYMrtP1277uIiIhUj089eIvFQqdOnbzvL7nkEiwWn34biIiISD3wucDv27fPe/59w4YNPl/ZLiIiInXPp2741KlTmTBhAj///DM9e/akVatWzJ07N9CxiYiISDX5VOAjIyP5+OOPOXLkCDabjaZNmwY6LqlDjo0XUdwusr7DEBGRWuTTEP3DD0TfxVQAABjeSURBVD8MnJxpTsVdRESk4fOpB9+2bVumTJlCjx49sFp/n0Vs+PDhAQtMREREqu+sBX7Hjh3e2eSCgoLYsGEDkZG/D+WqwIuIiDRMZy3wTz/9NG+++ab3nvQ77riDxYsX10lgIiIiUn1nPQevW+FERETOT2ct8KfPO6+CLyIicn7w6/mevjxoRkREROrfWc/Bf/PNN1x33XXe94cPH+a6667zPlVu/fr1AQ5PREREquOsBT49Pb2u4hAREZFadNYC36pVq7qKQ0RERGqRX+fgRURE5PygAi8iImJAKvAiIiIGpAIvIiJiQD49bEYaH/PBQ/UdgoiI1IB68CIiIgakAi8iImJAKvAiIiIGpAIvIiJiQCrwIiIiBqQCLyIiYkAq8CIiIgYUsPvgi4qKmDZtGocPH6akpIQJEybQqVMnpkyZQllZGTExMcybNw+bzcbq1atZtmwZZrOZESNGkJiYiNvtZtq0aRw4cICgoCBmz55N69atAxWuiIiIoQSsB79u3Tq6dOnCW2+9xcKFC5kzZw7PP/88SUlJrFixgjZt2pCWlkZhYSGLFi3ijTfeYPny5SxbtoyjR4/yySefEBERwTvvvMP48eOZP39+oEIVERExnIAV+KFDh3LPPfcAcPDgQVq0aEFWVhbXX389AAMGDCAzM5OtW7cSGxtLeHg4wcHBxMXF4XA4yMzMJCEhAYB+/frhcDgCFaqIiIjhBHyq2pEjR3Lo0CEWL17MXXfdhc1mAyA6Ohqn00lubi5RUVHe9aOioiq1m81mTCYTLpfLu72IiIicWcAL/MqVK/nxxx+ZPHkyHo/H237q61P5236qyMhQLJYgv2OMiQn3exujaHYkhLCQJgTbrBUXWP0f3LFVY5vqCA62VtnerFnIuTdu4H/rxnws1iblseaUw9pRn3kMWIHfvn070dHRXHTRRVx++eWUlZURFhZGcXExwcHBZGdnY7fbsdvt5ObmerfLycmhe/fu2O12nE4nnTp1wu124/F4ztl7z8sr9DvOmJhwnM7jfm9nFMfyiygoKqHY5a7QbnaX+7Ufm9WMy89tqmu7e1+V7bsLcqpsHxLW0/va1YD/1o39WKwtymPNKYe1oy7yeLYfEAHrcm3evJnXXnsNgNzcXAoLC+nXrx8ZGRkArF27lvj4eLp168a2bdvIz8+noKAAh8NBr1696N+/P+np6cDJC/b69OkTqFBFREQMJ2A9+JEjR/K3v/2NpKQkiouLefTRR+nSpQtTp05l1apVtGzZkmHDhmG1WklJSWHs2LGYTCaSk5MJDw9n6NChfPnll4waNQqbzcacOXMCFaqIiIjhBKzABwcHV3lr2+uvv16pbfDgwQwePLhC22/3vouIiIj/NJOdiIiIAanAi4iIGJAKvIiIiAGpwIuIiBiQCryIiIgBBXwmO5G6suW7y6te0OQi78tSfp/pcNCgskCHJCJSb9SDFxERMSAVeBEREQNSgRcRETEgFXgRERED0kV2YhiXd/m2yvaW1kPe12Wtsr2vbbZyXK4hAY9LRKQ+qAcvIiJiQCrwIiIiBqQCLyIiYkAq8CIiIgakAi8iImJAuoregGy2NT6v26rZForbRQYwGhERqQ/qwYuIiBiQCryIiIgBqcCLiIgYkAq8iIiIAanAi4iIGJAKvIiIiAGpwIuIiBiQCryIiIgBqcCLiIgYUEBnsps7dy5btmyhtLSUe++9l9jYWKZMmUJZWRkxMTHMmzcPm83G6tWrWbZsGWazmREjRpCYmIjb7WbatGkcOHCAoKAgZs+eTevWrQMZroiIiGEErMBv2rSJ//73v6xatYq8vDxuueUW+vbtS1JSEkOGDGHBggWkpaUxbNgwFi1aRFpaGlarleHDh5OQkMC6deuIiIhg/vz5fP7558yfP5+FCxcGKlwRERFDCdgQ/ZVXXslzzz0HQEREBEVFRWRlZXH99dcDMGDAADIzM9m6dSuxsbGEh4cTHBxMXFwcDoeDzMxMEhISAOjXrx8OhyNQoYqIiBhOwAp8UFAQoaGhAKSlpXHNNddQVFSEzWYDIDo6GqfTSW5uLlFRUd7toqKiKrWbzWZMJhMulytQ4YqIiBhKwJ8m99lnn5GWlsZrr73GwIEDve0ej6fK9f1tP1VkZCgWS5DfMcbEhPu9TcMW4vOaYcdtBNuslRdY/f/tZ6vGNnUhOPiU7xfWxPuyWTOAhvW3N96xWD+Ux5pTDmtHfeYxoAV+48aNLF68mKVLlxIeHk5oaCjFxcUEBweTnZ2N3W7HbreTm5vr3SYnJ4fu3btjt9txOp106tQJt9uNx+Px9v7PJC+v0O8YY2LCcTqP+71dQ2azFfm8bkGhi2KXu1K72V3u32dazbj83KauFPP79ysrKPG+PnasHJer4fztjXgs1gflseaUw9pRF3k82w+IgHW5jh8/zty5c1myZAnNmzcHTp5Lz8jIAGDt2rXEx8fTrVs3tm3bRn5+PgUFBTgcDnr16kX//v1JT08HYN26dfTp0ydQoYqIiBhOwHrwn376KXl5eTz44IPetjlz5jBjxgxWrVpFy5YtGTZsGFarlZSUFMaOHYvJZCI5OZnw8HCGDh3Kl19+yahRo7DZbMyZMydQoYqIiBiOyePLye3zRHWGQs73oShbxppKbZZWW6pctzSuZ6U2R/YWdv0QWandfPCQf3E04CH6DtYLva/L2nfwvo6LK8flGlIfIVXpfD8WGwrlseaUw9ph2CF6ERERqT8q8CIiIgakAi8iImJAKvAiIiIGpAIvIiJiQCrwIiIiBqQCLyIiYkAq8CIiIgakAi8iImJAAX+anNSczVZ5trrfnGnWOhERadxU4BsRi6Pyj4Ggkp2YD7auh2hERCSQNEQvIiJiQCrwIiIiBqQCLyIiYkAq8CIiIgaki+ykUQnatdP72kIe7D/zuq5BDedZ8SIi/lIPXkRExIBU4EVERAxIBV5ERMSAVOBFREQMSAVeRETEgFTgRUREDEgFXkRExIB0H7zBbCnZee6VRETE8NSDFxERMSAVeBEREQMKaIH/6aefuOGGG3jrrbcAOHjwIGPGjCEpKYlJkybhcrkAWL16NbfddhuJiYm89957ALjdblJSUhg1ahSjR49m3759gQxVRETEUAJ2Dr6wsJAnnniCvn37etuef/55kpKSGDJkCAsWLCAtLY1hw4axaNEi0tLSsFqtDB8+nISEBNatW0dERATz58/n888/Z/78+SxcuDBQ4UojZWm15cwLbZWbXC7NTy8i54eA9eBtNhuvvPIKdrvd25aVlcX1118PwIABA8jMzGTr1q3ExsYSHh5OcHAwcXFxOBwOMjMzSUhIAKBfv344HI5AhSoiImI4ASvwFouF4ODgCm1FRUXYbCe7RdHR0TidTnJzc4mKivKuExUVVandbDZjMpm8Q/oiIiJydvV2m5zH46mV9lNFRoZisQT5HUtMTLjf29StkDMvCqs4jhxcjT+pzVo7v/Nqaz+1LTjYWmV7WFgVY/CnalZV3gN7rDT8Y/H8oDzWnHJYO+ozj3Va4ENDQykuLiY4OJjs7Gzsdjt2u53c3FzvOjk5OXTv3h273Y7T6aRTp0643W48Ho+3938meXmFfscUExOO03nc7+3qks1WdMZlloKKoxrFJaV+79/lLvd7m9PZrOZa2U8gbHdXfYFm8ZHK7T2bdPC+Lj1WOe8uV+COlfPhWDwfKI81pxzWjrrI49l+QNRpge/Xrx8ZGRncfPPNrF27lvj4eLp168aMGTPIz88nKCgIh8NBamoqJ06cID09nfj4eNatW0efPn3qMtTz2s6fW9d3CCIiUs8CVuC3b9/OM888w/79+7FYLGRkZPDss88ybdo0Vq1aRcuWLRk2bBhWq5WUlBTGjh2LyWQiOTmZ8PBwhg4dypdffsmoUaOw2WzMmTMnUKGKiIgYjsnjy8nt80R1hkLOh6Eom23NGZdZHBVv89pSsrNeevANeYj+TDpcco4h+rielZYH8ja58+FYPB8ojzWnHNaO+h6ib5hXRYmIiEiNqMCLiIgYkAq8iIiIAanAi4iIGJAKvIiIiAHV20x24juH48y/w4J2RVZ4v9Ote+B9VeXdBtbf81l22u/fuLjz6y4BEWnc1IMXERExIBV4ERERA9IQvcgZBO3aWeG9hTwsVP38+KomxQE9P15E6o968CIiIgakHrxIAGVk+P744mbN4KqrAhiMiDQq6sGLiIgYkAq8iIiIAanAi4iIGJDOwYv4aMtpkwqd6vRJcX7TqtWnPu8/LKwJNluRrrwXkVqhHryIiIgBqcCLiIgYkIboRU6x033I53U7WC8MYCQiIjWjAn8e2HVs5xmXmf0oSFK7Tv0xUH7s3Ou3b9bhnOs4HGb27/f93vlBg8p8XldEGhcVeJFaYD547h9aQbkn/7+s/bkLvYhITanAi9SR33r8Z+rtBxdZKXa5oenzPvX2AWy2cl11LyJV0kV2IiIiBqQefANky1hT4b35uM6zG8kZh/OtZszuckDD+SJSc+rBi4iIGJB68CIN2OnPpD/d1l0Ai/zaZ8/2eRR2mVH9oETkvKAevIiIiAE16B78008/zdatWzGZTKSmptK1a9f6DkmkTgR6wp3Q7U/6tF5pXE8cDv/6AXFxurJfpCFosAX+q6++Yu/evaxatYpdu3aRmprKqlWr6jsskUbF4thC0FkesnOq3y4I9HeyHtCEPSKB0GALfGZmJjfccAMA7du359ixY5w4cYKmTZvWc2Qi57ezPRWvJk69XuBinj/ruqffHbB9u3+fZaRRApttzblXOoVRvrcEXoMt8Lm5uXTu3Nn7PioqCqfTqQIvchp/hvP9Faj59s918eC5nO3iwp+DrRQXu8+5j9q+BTEurrxW93cm/v4gqJ4RAf+EhvjDpvZzG4LNVlShpS5/oDXYAn86j8dzznViYsKrte/qbhcwoyv+4/pLHfxjExE5VeD/u9gQ/7tW+zE1a1bru/RZg72K3m63k5ub632fk5NDTExMPUYkIiJy/miwBb5///5kZGQA8P3332O32zU8LyIi4qMGO0QfFxdH586dGTlyJCaTiZkzZ9Z3SCIiIucNk8eXk9siIiJyXmmwQ/QiIiJSfSrwIiIiBtRgz8HXBU2F65+5c+eyZcsWSktLuffee4mNjWXKlCmUlZURExPDvHnzsNlsrF69mmXLlmE2mxkxYgSJiYn1HXqDUlxczJ/+9CcmTJhA3759lcNqWL16NUuXLsVisfDAAw/QsWNH5dEPBQUFTJ06lWPHjuF2u0lOTiYmJoZZs2YB0LFjRx577DEAli5dSnp6OiaTiYkTJ3LttdfWY+QNw08//cSECRP485//zOjRozl48KDPx5/b7WbatGkcOHCAoKAgZs+eTevWrQMTqKeRysrK8owbN87j8Xg8O3fu9IwYMaKeI2rYMjMzPXfffbfH4/F4jhw54rn22ms906ZN83z66acej8fjmT9/vuftt9/2FBQUeAYOHOjJz8/3FBUVeW688UZPXl5efYbe4CxYsMBz6623et5//33lsBqOHDniGThwoOf48eOe7Oxsz4wZM5RHPy1fvtzz7LPPejwej+fQoUOeQYMGeUaPHu3ZunWrx+PxeP7617961q9f7/nll188t9xyi6ekpMRz+PBhz6BBgzylpaX1GXq9Kygo8IwePdozY8YMz/Llyz0ej8ev4++DDz7wzJo1y+PxeDwbN270TJo0KWCxNtoh+jNNhStVu/LKK3nuuecAiIiIoKioiKysLK6//noABgwYQGZmJlu3biU2Npbw8HCCg4OJi4vD4XDUZ+gNyq5du9i5cyfXXXcdgHJYDZmZmfTt25emTZtit9t54oknlEc/RUZGcvToUQDy8/Np3rw5+/fv945i/pbDrKws4uPjsdlsREVF0apVK3burNkshOc7m83GK6+8gt1u97b5c/xlZmaSkJAAQL9+/QJ6TDbaAp+bm0tk5O9zcv82Fa5ULSgoiNDQUADS0tK45pprKCoqwmazARAdHY3T6SQ3N5eoqCjvdsprRc888wzTpk3zvlcO/ffrr79SXFzM+PHjSUpKIjMzU3n004033siBAwdISEhg9OjRTJkyhYiICO9y5fDMLBYLwcHBFdr8Of5ObTebzZhMJlwuV2BiDchez0Me3S3ok88++4y0tDRee+01Bg4c6G0/U/6U1999+OGHdO/e/Yzn25RD3x09epQXX3yRAwcOcMcdd1TIkfJ4bh999BEtW7bk1VdfZceOHSQnJxMe/vvUtMph9fmbu0DmtNEWeE2F67+NGzeyePFili5dSnh4OKGhoRQXFxMcHEx2djZ2u73KvHbv3r0eo2441q9fz759+1i/fj2HDh3CZrMph9UQHR1Njx49sFgsXHzxxYSFhREUFKQ8+sHhcHD11VcD0KlTJ0pKSigtLfUuPzWHP//8c6V2qciff8d2ux2n00mnTp1wu914PB5v77+2Ndohek2F65/jx48zd+5clixZQvPmzYGT549+y+HatWuJj4+nW7dubNu2jfz8fAoKCnA4HPTq1as+Q28wFi5cyPvvv8+7775LYmIiEyZMUA6r4eqrr2bTpk2Ul5eTl5dHYWGh8uinNm3asHXrVgD2799PWFgY7du3Z/PmzcDvObzqqqtYv349LpeL7OxscnJy6NChdp/CZwT+HH/9+/cnPT0dgHXr1tGnT5+AxdWoZ7J79tln2bx5s3cq3E6dOtV3SA3WqlWreOGFF7jkkku8bXPmzGHGjBmUlJTQsmVLZs+ejdVqJT09nVdffRWTycTo0aO56aab6jHyhumFF16gVatWXH311UydOlU59NPKlStJS0sD4L777iM2NlZ59ENBQQGpqakcPnyY0tJSJk2aRExMDI8++ijl5eV069aN6dOnA7B8+XI+/vhjTCYTDz74IH379q3n6OvX9u3beeaZZ9i/fz8Wi4UWLVrw7LPPMm3aNJ+Ov7KyMmbMmMGePXuw2WzMmTOHiy66KCCxNuoCLyIiYlSNdoheRETEyFTgRUREDEgFXkRExIBU4EVERAxIBV5ERMSAGu1ENyJG8+uvvzJ48GB69OgBgNvtplevXiQnJxMSEhKQz0tKSuLf//53re/7N9nZ2ezevZu+ffvywgsvUFpaykMPPRSwzxMxEvXgRQwkKiqK5cuXs3z5cpYtW0ZBQQEpKSn1HVa1ZWVlsWnTpvoOQ+S8pB68iEE1adKE1NRUBg0axM6dO+nQoQMLFizA4XBQXFzMlVdeyZQpU8jJyeHhhx8GTj6r/vbbb2f48OHs2bOHRx55hPLycpo0acLs2bNp0aKFT5/96aef8tZbb+HxeIiKiuLJJ58kMjKSnj17Mn78eDZu3IjT6WThwoV07NiRDRs2MH/+fJo1a0Z8fDxvvfUWb7/9NgsXLsTj8XhnT8zOzuaBBx5g9+7d9O7dm0cffTRg+RM536kHL2JgVquVLl268NNPP7FmzRqys7N56623SEtL45dffmHdunWsWbOGdu3asXz5ct566y2Ki4sBmDlzJmPHjuXtt9/mtttuY82aNT595sGDB1m8eDFvvPEG77zzDr1792bJkiUAnDhxgssuu4w333yTG2+8kffeew+Px8PMmTOZO3cuy5cv5/jx4wC0bt2aW265hZtuuom77roLgL1797JgwQLef/99/vGPf5CXlxeArIkYg3rwIgZ3/PhxzGYzmzZt4ttvv2XMmDHe9l9//ZX4+HhWrFjBtGnTuPbaa7n99tsB+O677+jduzdw8vGivvrmm29wOp2MHTsWAJfLxR/+8Afv8quuugqAli1bsnfvXu988r9NFT1o0CA++uijKvfds2dPLBYLFouFyMhIjh8/XuGxzyLyOxV4EQMrKirixx9/pHPnzjgcDkaMGOEtvKf63//9X77++mvS09NZtmwZK1euBKC8vNzvz7TZbHTt2tXbaz9dUFCQ97XH48Hj8WAymapcfrZtf9teRKqmIXoRg3K73Tz55JP079+f1q1b07NnT/7v//7P+1jQF198kT179vDxxx+zbds2+vXrx8yZMzl48CClpaXExcWxceNG4OQ59QULFvj0ubGxsXz33Xc4nU4A1qxZw2effXbG9SMjIzGbzezevRs4+TSu35hMpgqPMRUR36kHL2IgR44cYcyYMZSVlZGfn0///v29F6INHDiQb7/9lpEjRxIUFMQVV1xB69atKSoqYubMmdhsNjweD/fccw8Wi4VHHnmERx55hBUrVmCxWHj66afP+Hm/iY2NZcqUKfztb3/j3nvvJSQkhODgYJ555pkzxmw2m0lNTSU5OZmWLVvSq1cvLJaT/2nq1asXDz30EFar9aw9exGpTE+TE5F699lnn9GxY0dat27N2rVrWbVqFa+++mp9hyVyXlMPXkTqXXl5Offffz9NmzalrKyMWbNm1XdIIuc99eBFREQMSBfZiYiIGJAKvIiIiAGpwIuIiBiQCryIiIgBqcCLiIgYkAq8iIiIAf3/GWvzjxpcnycAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 576x396 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-ausEEgiCOCI"
      },
      "source": [
        "Seems all desc lenght pretty equal"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LmN3yG5VDJdL",
        "outputId": "89da21ba-835d-47d6-c145-d5d885d4cf25",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 361
        }
      },
      "source": [
        "distribution = news.groupby('Class Index').agg('count').reset_index()\n",
        "plt.bar(distribution['Class Index'], distribution['Description'])\n",
        "\n",
        "plt.ylabel('Count')\n",
        "plt.xlabel('Class')\n",
        "\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAf0AAAFYCAYAAABZHSXVAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3de3BU9f3/8ddmN2uMboTEXSuVWhEEWiBIUWooYNBoTG2LlFso9hZbKNFKmxZiVMCxmBiNRZQpFwtkQgPRdMam1hKqjQxOtmlxWwpWCtqZNg2Y7CrhkgsJ4fz+6I/9EgPJRrJZN5/nY4aZ5LNnD+/3ebO8ck72YrMsyxIAABjwYiJdAAAA6B+EPgAAhiD0AQAwBKEPAIAhCH0AAAxB6AMAYAhHpAsIN7//REjbDR4cr6NHm8NcTf8aiD1JA7OvgdiTNDD7oqfoMRD7CrUnt9t13nXO9P8/h8Me6RL63EDsSRqYfQ3EnqSB2Rc9RY+B2NfF9kToAwBgCEIfAABDEPoAABiC0AcAwBCEPgAAhiD0AQAwBKEPAIAhCH0AAAxB6AMAYAhCHwAAQxD6AAAYgtAHAMAQA/5T9sLhuwV/jHQJA86m3Ol9vk/mFB7MKnowq+gQjjldCGf6AAAYgtAHAMAQhD4AAIYg9AEAMAShDwCAIQh9AAAMQegDAGAIQh8AAEMQ+gAAGILQBwDAEIQ+AACGIPQBADBE2D5wp6WlRbm5ufrggw906tQpLV68WKNGjdLSpUvV0dEht9utp556Sk6nUxUVFSouLlZMTIzmzJmj2bNnq729Xbm5uTp8+LDsdrvy8/M1dOhQHThwQCtXrpQkjRw5Uo899li4WgAAYEAJ25l+VVWVxowZo61bt2r16tUqKCjQmjVrNH/+fJWWluraa69VeXm5mpubtXbtWm3ZskUlJSUqLi5WY2OjXnnlFSUkJGjbtm1atGiRioqKJEmrVq1SXl6etm/frpMnT2rXrl3hagEAgAElbKGfkZGh733ve5KkI0eO6KqrrlJNTY1uu+02SVJqaqq8Xq/27t2rsWPHyuVyKS4uThMmTJDP55PX61VaWpokKSUlRT6fT21tbaqrq9O4ceM67QMAAPQsbJf3z5o3b57ef/99rVu3Tt/5znfkdDolSUlJSfL7/QoEAkpMTAxun5iY2GU9JiZGNptNgUBACQkJwW3P7gMAAPQs7KG/fft2vfPOO/rpT38qy7KC6+d+fa7erF9o23MNHhwvh8MeUq1utyuk7dD3envsmVXkMKvo0Ztjz5wipz8fU2EL/f379yspKUlXX321Ro8erY6ODl122WVqbW1VXFyc6uvr5fF45PF4FAgEgvdraGjQ+PHj5fF45Pf7NWrUKLW3t8uyLLndbjU2Nga3PbuP7hw92hxSvW63S37/iY/XLC5ab449s4osZhU9Qj32zCmywvGYutAPBmH7nf6ePXu0adMmSVIgEFBzc7NSUlJUWVkpSdq5c6emTJmi5ORk7du3T8ePH1dTU5N8Pp8mTpyoyZMna8eOHZL+96TASZMmKTY2VsOGDdOePXs67QMAAPQsbGf68+bN08MPP6z58+ertbVVy5cv15gxY7Rs2TKVlZVpyJAhmjFjhmJjY5WTk6OsrCzZbDZlZ2fL5XIpIyND1dXVyszMlNPpVEFBgSQpLy9Py5cv15kzZ5ScnKyUlJRwtQAAwIASttCPi4sLvszuXJs3b+6ylp6ervT09E5rZ1+b/1HDhw9XaWlp3xUKAIAheEc+AAAMQegDAGAIQh8AAEMQ+gAAGILQBwDAEIQ+AACGIPQBADAEoQ8AgCEIfQAADEHoAwBgCEIfAABDEPoAABiC0AcAwBCEPgAAhiD0AQAwBKEPAIAhCH0AAAxB6AMAYAhCHwAAQxD6AAAYgtAHAMAQhD4AAIYg9AEAMAShDwCAIQh9AAAMQegDAGAIQh8AAEMQ+gAAGILQBwDAEIQ+AACGIPQBADAEoQ8AgCEIfQAADEHoAwBgCEIfAABDEPoAABjCEc6dFxYW6q233tLp06e1cOFC/fGPf9Tbb7+tQYMGSZKysrJ06623qqKiQsXFxYqJidGcOXM0e/Zstbe3Kzc3V4cPH5bdbld+fr6GDh2qAwcOaOXKlZKkkSNH6rHHHgtnCwAADBhhC/0//elPOnTokMrKynT06FHdc889+uIXv6gf//jHSk1NDW7X3NystWvXqry8XLGxsZo1a5bS0tJUVVWlhIQEFRUV6c0331RRUZFWr16tVatWKS8vT+PGjVNOTo527dqladOmhasNAAAGjLBd3r/pppv07LPPSpISEhLU0tKijo6OLtvt3btXY8eOlcvlUlxcnCZMmCCfzyev16u0tDRJUkpKinw+n9ra2lRXV6dx48ZJklJTU+X1esPVAgAAA0rYQt9utys+Pl6SVF5erqlTp8put2vr1q365je/qR/96Ef68MMPFQgElJiYGLxfYmKi/H5/p/WYmBjZbDYFAgElJCQEt01KSpLf7w9XCwAADChh/Z2+JL322msqLy/Xpk2btH//fg0aNEijR4/Whg0b9Pzzz+vGG2/stL1lWefdz/nWL7TtuQYPjpfDYQ+pVrfbFdJ26Hu9PfbMKnKYVfTozbFnTpHTn4+psIb+7t27tW7dOr3wwgtyuVy65ZZbgrdNnz5dK1eu1J133qlAIBBcb2ho0Pjx4+XxeOT3+zVq1Ci1t7fLsiy53W41NjYGt62vr5fH4+m2hqNHm0Oq1e12ye8/0csO0Vd6c+yZVWQxq+gR6rFnTpEVjsfUhX4wCNvl/RMnTqiwsFDr168PPlv/gQceUG1trSSppqZGI0aMUHJysvbt26fjx4+rqalJPp9PEydO1OTJk7Vjxw5JUlVVlSZNmqTY2FgNGzZMe/bskSTt3LlTU6ZMCVcLAAAMKGE703/11Vd19OhRLVmyJLg2c+ZMLVmyRJdeeqni4+OVn5+vuLg45eTkKCsrSzabTdnZ2XK5XMrIyFB1dbUyMzPldDpVUFAgScrLy9Py5ct15swZJScnKyUlJVwtAAAwoIQt9OfOnau5c+d2Wb/nnnu6rKWnpys9Pb3T2tnX5n/U8OHDVVpa2neFAgBgCN6RDwAAQxD6AAAYgtAHAMAQhD4AAIYg9AEAMAShDwCAIQh9AAAMQegDAGAIQh8AAEMQ+gAAGILQBwDAEIQ+AACGIPQBADAEoQ8AgCEIfQAADEHoAwBgCEIfAABDEPoAABiC0AcAwBCEPgAAhiD0AQAwBKEPAIAhCH0AAAxB6AMAYAhCHwAAQxD6AAAYgtAHAMAQhD4AAIYg9AEAMAShDwCAIQh9AAAMQegDAGAIQh8AAEMQ+gAAGILQBwDAEIQ+AACGcIRz54WFhXrrrbd0+vRpLVy4UGPHjtXSpUvV0dEht9utp556Sk6nUxUVFSouLlZMTIzmzJmj2bNnq729Xbm5uTp8+LDsdrvy8/M1dOhQHThwQCtXrpQkjRw5Uo899lg4WwAAYMAI25n+n/70Jx06dEhlZWV64YUX9MQTT2jNmjWaP3++SktLde2116q8vFzNzc1au3attmzZopKSEhUXF6uxsVGvvPKKEhIStG3bNi1atEhFRUWSpFWrVikvL0/bt2/XyZMntWvXrnC1AADAgBK20L/pppv07LPPSpISEhLU0tKimpoa3XbbbZKk1NRUeb1e7d27V2PHjpXL5VJcXJwmTJggn88nr9ertLQ0SVJKSop8Pp/a2tpUV1encePGddoHAADoWdgu79vtdsXHx0uSysvLNXXqVL355ptyOp2SpKSkJPn9fgUCASUmJgbvl5iY2GU9JiZGNptNgUBACQkJwW3P7qM7gwfHy+Gwh1Sz2+3qVY/oO7099swqcphV9OjNsWdOkdOfj6mw/k5fkl577TWVl5dr06ZNuuOOO4LrlmWdd/verF9o23MdPdocUp1ut0t+/4mQtkXf682xZ1aRxayiR6jHnjlFVjgeUxf6wSCsz97fvXu31q1bp40bN8rlcik+Pl6tra2SpPr6enk8Hnk8HgUCgeB9Ghoagutnz+Lb29tlWZbcbrcaGxuD257dBwAA6FnYQv/EiRMqLCzU+vXrNWjQIEn/+918ZWWlJGnnzp2aMmWKkpOTtW/fPh0/flxNTU3y+XyaOHGiJk+erB07dkiSqqqqNGnSJMXGxmrYsGHas2dPp30AAICehe3y/quvvqqjR49qyZIlwbWCggI98sgjKisr05AhQzRjxgzFxsYqJydHWVlZstlsys7OlsvlUkZGhqqrq5WZmSmn06mCggJJUl5enpYvX64zZ84oOTlZKSkp4WoBAIABJWyhP3fuXM2dO7fL+ubNm7uspaenKz09vdPa2dfmf9Tw4cNVWlrad4UCAGAI3pEPAABDEPoAABiC0AcAwBCEPgAAhiD0AQAwBKEPAIAhCH0AAAxB6AMAYAhCHwAAQxD6AAAYIqTQf++997qs/e1vf+vzYgAAQPh0G/rHjx/Xf/7zH+Xl5am2tjb451//+peWLVvWXzUCAIA+0O0H7vz1r39VcXGx3nnnHX3rW98KrsfExOhLX/pS2IsDAAB9p9vQnzZtmqZNm6Zt27YpMzOzv2oCAABhENJH695+++0qLi7WsWPHZFlWcP3BBx8MW2EAAKBvhfREvoULF+rAgQOKiYmR3W4P/gEAANEjpDP9+Ph45efnh7sWAAAQRiGd6ScnJ5/3ZXsAACB6hHSmv3v3bm3ZskWDBw+Ww+GQZVmy2Wx64403wlweAADoKyGF/i9+8Ytw1wEAAMIspND3er3nXZ81a1afFgMAAMInpNB/6623gl+3tbXp73//uyZMmEDoAwAQRUIK/Y8+c7+lpUUPPfRQWAoCAADh8bE+Ze/SSy/Vf/7zn76uBQAAhFFIZ/rz58+XzWYLfl9fX6+RI0eGrSgAAND3Qgr9JUuWBL+22Wy6/PLLNWrUqLAVBQAA+l5Il/dvvvlmxcTE6O2339bbb7+t1tbWTmf+AADgky+k0H/22WdVWFiohoYG1dfX62c/+5nWr18f7toAAEAfCunyfk1NjbZv366YmP/9jHD69GktWLBACxcuDGtxAACg74R0pn/mzJlg4EuSw+Hg8j4AAFEmpDP9MWPGaNGiRUpJSZEkVVdXa8yYMWEtDAAA9K0eQ7+2tlZ5eXn6/e9/r71798pms2nixIm67777+qM+AADQR7q9vO/1epWZmammpiZ9+ctfVl5enmbOnKlt27Zp//79/VUjAADoA92G/vPPP69NmzbJ5XIF10aOHKl169Zp9erVYS8OAAD0nW5D37Is3XDDDV3WR4wYoVOnToWtKAAA0Pe6Df3m5uYL3tbY2Njjzg8ePKjbb79dW7dulSTl5ubqK1/5iu69917de++9euONNyRJFRUV+vrXv67Zs2frpZdekiS1t7crJydHmZmZWrBggWprayVJBw4c0Lx58zRv3jytWLEipCYBAEAPoT9ixAht27aty/rGjRuVnJzc7Y6bm5v1+OOP65Zbbum0/uMf/1glJSUqKSnRrbfequbmZq1du1ZbtmxRSUmJiouL1djYqFdeeUUJCQnatm2bFi1apKKiIknSqlWrlJeXp+3bt+vkyZPatWtXb3sGAMBI3T57f+nSpcrOztZvfvMbjRkzRmfOnJHP59Pll1/e4zvyOZ1Obdy4URs3bux2u71792rs2LHB5w1MmDBBPp9PXq9XM2bMkCSlpKQoLy9PbW1tqqur07hx4yRJqamp8nq9mjZtWsgNAwBgqm5D3+1268UXX5TX69WhQ4dkt9t111136aabbup5xw6HHI6uu9+6das2b96spKQkPfroowoEAkpMTAzenpiYKL/f32k9JiZGNptNgUBACQkJwW2TkpLk9/tDbhYAAJOF9OY8t9xyS5fL9B/H1772NQ0aNEijR4/Whg0b9Pzzz+vGG2/stI1lWee97/nWL7TtuQYPjpfDYQ+pPrfb1fNGCIveHntmFTnMKnr05tgzp8jpz8dUSKHfV879wWH69OlauXKl7rzzTgUCgeB6Q0ODxo8fL4/HI7/fr1GjRqm9vV2WZcntdnd6AmF9fb08Hk+3f+fRoxd+MuK53G6X/P4TvewIfaU3x55ZRRazih6hHnvmFFnheExd6AeDkN57v6888MADwWfh19TUaMSIEUpOTta+fft0/PhxNTU1yefzaeLEiZo8ebJ27NghSaqqqtKkSZMUGxurYcOGac+ePZKknTt3asqUKf3ZAgAAUStsZ/r79+/Xk08+qbq6OjkcDlVWVmrBggVasmSJLr30UsXHxys/P19xcXHKyclRVlaWbDabsrOz5XK5lJGRoerqamVmZsrpdKqgoECSlJeXp+XLl+vMmTNKTk4Ofh4AAADoXthCf8yYMSopKemyfuedd3ZZS09PV3p6eqc1u92u/Pz8LtsOHz5cpaWlfVcoAACG6NfL+wAAIHIIfQAADEHoAwBgCEIfAABDEPoAABiC0AcAwBCEPgAAhiD0AQAwBKEPAIAhCH0AAAxB6AMAYAhCHwAAQxD6AAAYgtAHAMAQhD4AAIYg9AEAMAShDwCAIQh9AAAMQegDAGAIQh8AAEMQ+gAAGILQBwDAEIQ+AACGIPQBADAEoQ8AgCEIfQAADEHoAwBgCEIfAABDEPoAABiC0AcAwBCEPgAAhiD0AQAwBKEPAIAhCH0AAAxB6AMAYAhCHwAAQ4Q19A8ePKjbb79dW7dulSQdOXJE9957r+bPn68HH3xQbW1tkqSKigp9/etf1+zZs/XSSy9Jktrb25WTk6PMzEwtWLBAtbW1kqQDBw5o3rx5mjdvnlasWBHO8gEAGFDCFvrNzc16/PHHdcsttwTX1qxZo/nz56u0tFTXXnutysvL1dzcrLVr12rLli0qKSlRcXGxGhsb9corryghIUHbtm3TokWLVFRUJElatWqV8vLytH37dp08eVK7du0KVwsAAAwoYQt9p9OpjRs3yuPxBNdqamp02223SZJSU1Pl9Xq1d+9ejR07Vi6XS3FxcZowYYJ8Pp+8Xq/S0tIkSSkpKfL5fGpra1NdXZ3GjRvXaR8AAKBnjrDt2OGQw9F59y0tLXI6nZKkpKQk+f1+BQIBJSYmBrdJTEzssh4TEyObzaZAIKCEhITgtmf3AQAAeha20O+JZVkXvX6hbc81eHC8HA57SDW53a6QtkPf6+2xZ1aRw6yiR2+OPXOKnP58TPVr6MfHx6u1tVVxcXGqr6+Xx+ORx+NRIBAIbtPQ0KDx48fL4/HI7/dr1KhRam9vl2VZcrvdamxsDG57dh/dOXq0OaTa3G6X/P4TH68xXLTeHHtmFVnMKnqEeuyZU2SF4zF1oR8M+vUleykpKaqsrJQk7dy5U1OmTFFycrL27dun48ePq6mpST6fTxMnTtTkyZO1Y8cOSVJVVZUmTZqk2NhYDRs2THv27Om0DwAA0LOwnenv379fTz75pOrq6uRwOFRZWamnn35aubm5Kisr05AhQzRjxgzFxsYqJydHWVlZstlsys7OlsvlUkZGhqqrq5WZmSmn06mCggJJUl5enpYvX64zZ84oOTlZKSkp4WoBAIABJWyhP2bMGJWUlHRZ37x5c5e19PR0paend1qz2+3Kz8/vsu3w4cNVWlrad4UCAGAI3pEPAABDEPoAABiC0AcAwBCEPgAAhiD0AQAwBKEPAIAhCH0AAAxB6AMAYAhCHwAAQxD6AAAYgtAHAMAQhD4AAIYg9AEAMAShDwCAIQh9AAAMQegDAGAIQh8AAEMQ+gAAGILQBwDAEIQ+AACGIPQBADAEoQ8AgCEIfQAADEHoAwBgCEIfAABDEPoAABiC0AcAwBCEPgAAhiD0AQAwBKEPAIAhCH0AAAxB6AMAYAhCHwAAQxD6AAAYgtAHAMAQhD4AAIZw9OdfVlNTowcffFAjRoyQJN1www267777tHTpUnV0dMjtduupp56S0+lURUWFiouLFRMTozlz5mj27Nlqb29Xbm6uDh8+LLvdrvz8fA0dOrQ/WwAAIGr1a+hL0s0336w1a9YEv3/ooYc0f/583XXXXXrmmWdUXl6uGTNmaO3atSovL1dsbKxmzZqltLQ0VVVVKSEhQUVFRXrzzTdVVFSk1atX93cLAABEpYhf3q+pqdFtt90mSUpNTZXX69XevXs1duxYuVwuxcXFacKECfL5fPJ6vUpLS5MkpaSkyOfzRbJ0AACiSr+f6b/77rtatGiRjh07pvvvv18tLS1yOp2SpKSkJPn9fgUCASUmJgbvk5iY2GU9JiZGNptNbW1twfufz+DB8XI47CHV5na7LqIzXIzeHntmFTnMKnr05tgzp8jpz8dUv4b+Zz/7Wd1///266667VFtbq29+85vq6OgI3m5Z1nnv19v1cx092hxSbW63S37/iZC2Rd/rzbFnVpHFrKJHqMeeOUVWOB5TF/rBoF8v71911VXKyMiQzWbTZz7zGV155ZU6duyYWltbJUn19fXyeDzyeDwKBALB+zU0NATX/X6/JKm9vV2WZXV7lg8AAP5Pv4Z+RUWFfvnLX0qS/H6/PvjgA82cOVOVlZWSpJ07d2rKlClKTk7Wvn37dPz4cTU1Ncnn82nixImaPHmyduzYIUmqqqrSpEmT+rN8AACiWr9e3p8+fbp+8pOf6PXXX1d7e7tWrlyp0aNHa9myZSorK9OQIUM0Y8YMxcbGKicnR1lZWbLZbMrOzpbL5VJGRoaqq6uVmZkpp9OpgoKC/iwfAICo1q+hf/nll2vdunVd1jdv3txlLT09Xenp6Z3Wzr42HwAA9F7EX7IHAAD6B6EPAIAhCH0AAAxB6AMAYAhCHwAAQxD6AAAYgtAHAMAQhD4AAIYg9AEAMAShDwCAIQh9AAAMQegDAGAIQh8AAEMQ+gAAGILQBwDAEIQ+AACGIPQBADAEoQ8AgCEIfQAADEHoAwBgCEIfAABDEPoAABiC0AcAwBCEPgAAhiD0AQAwBKEPAIAhCH0AAAxB6AMAYAhCHwAAQxD6AAAYgtAHAMAQhD4AAIYg9AEAMAShDwCAIQh9AAAM4Yh0AR/HE088ob1798pmsykvL0/jxo2LdEkAAHziRV3o//nPf9a///1vlZWV6b333lNeXp7KysoiXRYAAJ94UXd53+v16vbbb5ckXX/99Tp27JhOnjwZ4aoAAPjki7rQDwQCGjx4cPD7xMRE+f3+CFYEAEB0iLrL+x9lWVa3t7vdrpD3Feq2vy36Wsj7RHiEMivm9MnArKID//9Fj97k2kdF3Zm+x+NRIBAIft/Q0CC32x3BigAAiA5RF/qTJ09WZWWlJOntt9+Wx+PR5ZdfHuGqAAD45Iu6y/sTJkzQ5z//ec2bN082m00rVqyIdEkAAEQFm9XTL8UBAMCAEHWX9wEAwMdD6AMAYIio+53+xeruLXynT5+uT33qU7Lb7ZKkp59+WldddVWkSu2VgwcPavHixfr2t7+tBQsWdLqturpazzzzjOx2u6ZOnars7OwIVdk73fUUzbMqLCzUW2+9pdOnT2vhwoW64447grdF66y66ykaZ9XS0qLc3Fx98MEHOnXqlBYvXqzU1NTg7dE6p576isZZndXa2qq7775bixcv1syZM4Pr0Tor6cI9XdScLIPU1NRY3//+9y3Lsqx3333XmjNnTqfbU1NTrZMnT0aitIvS1NRkLViwwHrkkUeskpKSLrffdddd1uHDh62Ojg4rMzPTOnToUASq7J2eeorWWXm9Xuu+++6zLMuyPvzwQ2vatGmdbo/GWfXUUzTO6ne/+521YcMGy7Is67///a91xx13dLo9GudkWT33FY2zOuuZZ56xZs6caf3617/utB6ts7KsC/d0MXMy6kz/Qm/hG+0v+XM6ndq4caM2btzY5bba2lpdccUVuvrqqyVJ06ZNk9fr1fDhw/u7zF7prqdodtNNNwWvLiUkJKilpUUdHR2y2+1RO6vueopWGRkZwa+PHDnS6SwqWuckdd9XNHvvvff07rvv6tZbb+20Hs2zulBPF8uo0A8EAvr85z8f/P7sW/ieG/orVqxQXV2dvvCFLygnJ0c2my0SpfaKw+GQw3H+Ufr9fiUmJga/T0xMVG1tbX+V9rF119NZ0Tgru92u+Ph4SVJ5ebmmTp0aDMdonVV3PZ0VjbOSpHnz5un999/XunXrgmvROqdzna+vs6JxVk8++aQeffRRvfzyy53Wo3lWF+rprI87J6NC/6Osj7xa8Yc//KGmTJmiK664QtnZ2aqsrFR6enqEqkN3on1Wr732msrLy7Vp06ZIl9JnLtRTNM9q+/bteuedd/TTn/5UFRUVURGAobhQX9E4q5dfflnjx4/X0KFDI11Kn+mpp4uZk1Gh39Nb+M6YMSP49dSpU3Xw4MFP/D/4nny05/r6enk8nghW1DeieVa7d+/WunXr9MILL8jl+r/30I7mWV2oJyk6Z7V//34lJSXp6quv1ujRo9XR0aEPP/xQSUlJUT2n7vqSonNWb7zxhmpra/XGG2/o/fffl9Pp1Kc+9SmlpKRE7ay660m6uDkZ9ZK97t7C98SJE8rKylJbW5sk6S9/+YtGjBgRsVr7yjXXXKOTJ0/qv//9r06fPq2qqipNnjw50mVdlGie1YkTJ1RYWKj169dr0KBBnW6L1ll111O0zmrPnj3BKxaBQEDNzc3BT/eM1jlJ3fcVrbNavXq1fv3rX+vFF1/U7NmztXjx4mA4RuusuuvpYudk3DvyPf3009qzZ0/wLXz/8Y9/yOVyKS0tTcXFxXr55Zd1ySWX6HOf+5weffTRqLict3//fj355JOqq6uTw+HQVVddpenTp+uaa65RWlqa/vKXv+jpp5+WJN1xxx3KysqKcMU966mnaJ1VWVmZnnvuOV133XXBtUmTJmnkyJFRO6ueeorGWbW2turhhx/WkSNH1Nraqvvvv1+NjY3B/yuicU5Sz31F46zO9dxzz+nTn/60JEX9rM46X08XMyfjQh8AAFMZdXkfAACTEfoAABiC0AcAwBCEPgAAhiD0AQAwhFFvzgPg42toaFBhYaEOHjyoyy67TJL0wAMP6P3331d1dXXwZVEAPrkIfQA9sixL2dnZmjFjRjDc/3xp8TIAAAJkSURBVPnPf+q73/2ulixZEuHqAISK0AfQI6/XK5vNpm984xvBtZEjR+rVV1/V66+/Hlz7wx/+oBdeeEFOp1MdHR0qLCzUNddco+LiYlVUVOjSSy9VXFycnnrqKbW1teknP/mJpP+9aczcuXM1a9asfu8NMAmhD6BHhw4d0tixY7usX3HFFZ2+P378uH7+859ryJAhWr9+vX71q19p2bJlWrNmjSorK3XllVdq9+7damhokNfr1bBhw/TYY4/p1KlTeumll/qrHcBYhD6AHtntdnV0dPS43ZVXXqlly5bJsiz5/X7deOONkqRZs2bpvvvu05133qn09HRdd911cjgcKi0tVW5urqZNm6a5c+eGuw3AeDx7H0CPbrjhBv31r3/tsv7Pf/5TLS0tkqT29nYtWbJEjz/+uLZu3ap77703uN1DDz2ktWvXBj8KdNeuXbr++uv1u9/9Tl/96lfl9Xo7bQ8gPAh9AD26+eabddlll2nDhg3BtUOHDukHP/iB7Ha7JKmpqUkxMTH69Kc/rVOnTun1119XW1ubjh07pueee05XX3215s+fr2984xvat2+ffvvb32rfvn1KSUnRihUrdOTIEZ0+fTpSLQJG4PI+gJBs2LBB+fn5uvvuuzVo0CBdcsklWr16td59911J0qBBg3T33Xdr1qxZGjJkiLKysrR06VJVV1erqalJs2bNUkJCghwOh1atWqUPP/xQK1askNPplGVZ+t73vieHg/+SgHDiU/YAADAEl/cBADAEoQ8AgCEIfQAADEHoAwBgCEIfAABDEPoAABiC0AcAwBCEPgAAhvh/tnUchF4je24AAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 576x396 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZU4bvpIoEVGl"
      },
      "source": [
        "Really ?! This dataset is very very very balanced, I love it!"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "NNdRRqcF_LOR"
      },
      "source": [
        "### Cleaning"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "at2ihZyIEe0A"
      },
      "source": [
        "**Main objective:**\n",
        "- remove stopwords. Stopwords list available on nltk module\n",
        "- remove punctuation. Punc list abailable on string module"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9cOD1T0a_iix",
        "outputId": "a5cfdd30-58f3-4ef9-fd55-60f4e5da9b9d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 68
        }
      },
      "source": [
        "import string\n",
        "import nltk\n",
        "\n",
        "nltk.download('stopwords')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[nltk_data] Downloading package stopwords to /root/nltk_data...\n",
            "[nltk_data]   Unzipping corpora/stopwords.zip.\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "True"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 13
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fj6Az0uF_NOD"
      },
      "source": [
        "from nltk.corpus import stopwords\n",
        "\n",
        "def text_process(body):\n",
        "    STOPWORDS = stopwords.words('english')\n",
        "\n",
        "    #remove punctuation\n",
        "    nopunc = [char for char in body if char not in string.punctuation]\n",
        "    nopunc = ''.join(nopunc)\n",
        "    \n",
        "    # remove stopwords\n",
        "    nostopwords = ' '.join([word for word in nopunc.split() if word.lower() not in STOPWORDS])\n",
        "\n",
        "    return nostopwords"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Vrowe4dy_PIu"
      },
      "source": [
        "news['clean_desc'] = news['Description'].apply(text_process)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "e0FNYDyPqtUm"
      },
      "source": [
        "### Splitting"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4ebDi677ou7T"
      },
      "source": [
        "news['Class Index'] -= 1"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lNWnDf7OqZjr"
      },
      "source": [
        "desc = news['clean_desc'].values\n",
        "label = news['Class Index'].values\n",
        "\n",
        "train_news, test_news, train_label, test_label = train_test_split(desc, label, test_size=0.2, shuffle=True)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1XhX7RRrrmZV",
        "outputId": "507d952a-9b5b-459e-fa31-5f6f8e9b4611",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "print(\"Total train: \", len(train_news))\n",
        "print(\"Total test: \", len(test_news))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Total train:  102080\n",
            "Total test:  25520\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ymWNF5mirxBx"
      },
      "source": [
        "## More pre-process dataset"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tP04fALbrtkM"
      },
      "source": [
        "from tensorflow.keras.preprocessing.text import Tokenizer\n",
        "from tensorflow.keras.preprocessing.sequence import pad_sequences"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kSxqXFV3tRtQ"
      },
      "source": [
        "NUM_WORDS = 1000\n",
        "LENGTH = 250\n",
        "OOV_TOKEN = \"<OOV>\"\n",
        "EMBEDDING_DIMS = 16"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "eDy1VOwktM7j",
        "outputId": "bc354b7a-0c88-40f3-ab9f-4c61a27b35fb",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "tokenizer = Tokenizer(num_words=NUM_WORDS, oov_token=\"<OOV>\")\n",
        "tokenizer.fit_on_texts(train_news)\n",
        "\n",
        "print(\"total word index: \", len(tokenizer.word_index))\n",
        "\n",
        "train_seq = tokenizer.texts_to_sequences(train_news)\n",
        "test_seq = tokenizer.texts_to_sequences(test_news)\n",
        "\n",
        "train_padded = pad_sequences(train_seq, maxlen=LENGTH, padding='post', truncating='post')\n",
        "test_padded = pad_sequences(test_seq, maxlen=LENGTH, padding='post', truncating='post')\n",
        "\n",
        "print(\"lenght: \", len(train_padded[0]))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "total word index:  94303\n",
            "lenght:  250\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YQyH5Dl3urLk",
        "outputId": "c82cf13a-84ef-4830-eb78-cef6b737efba",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 309
        }
      },
      "source": [
        "print(train_news[1000])\n",
        "print(train_padded[1000])\n",
        "print(train_label[1000])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Dig hype companies dish Whenever mutualfund company makes announcement designed build public 39s excitement interest investors start looking answer one simple question quotWhat 39s big deal\n",
            "[  1   1 132   1   1   1  16 543   1 817 941 279   2   1 469 218 205 680\n",
            "   1  17   1   1   1   2 160  61   0   0   0   0   0   0   0   0   0   0\n",
            "   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0\n",
            "   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0\n",
            "   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0\n",
            "   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0\n",
            "   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0\n",
            "   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0\n",
            "   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0\n",
            "   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0\n",
            "   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0\n",
            "   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0\n",
            "   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0\n",
            "   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0]\n",
            "2\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Kxl8MSTIwDMD"
      },
      "source": [
        "## Modeling"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "85bmkB2ixQNy"
      },
      "source": [
        "### Build model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Y-SnBoTevkyg"
      },
      "source": [
        "import tensorflow as tf\n",
        "from tensorflow.keras.regularizers import l2"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "M2krNEWsy2NA"
      },
      "source": [
        "#stop if no improvement on val_acc after 20 epochs\n",
        "early_stop = tf.keras.callbacks.EarlyStopping(\n",
        "    monitor='val_acc', patience=50, min_delta=0.0001\n",
        "    )"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fKr9cQjigaRd"
      },
      "source": [
        "#start with big learning rate and reduce if val_acc plateaued\n",
        "reduce_on_plateau = tf.keras.callbacks.ReduceLROnPlateau(\n",
        "                    monitor='val_acc', factor=0.1, patience=20, verbose=0, mode='auto',\n",
        "                    min_delta=0.0001, cooldown=5, min_lr=1e-8\n",
        "                )"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7qIFFc5TwLCM",
        "outputId": "fccb4582-0530-4d13-bb4a-829a7b13e80a",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 527
        }
      },
      "source": [
        "model = tf.keras.models.Sequential([\n",
        "        tf.keras.layers.Embedding(input_dim=NUM_WORDS, output_dim=EMBEDDING_DIMS,\n",
        "                                  input_length=LENGTH),\n",
        "        tf.keras.layers.Bidirectional(tf.keras.layers.LSTM(128, return_sequences=True, kernel_regularizer=l2(1e-5))),\n",
        "        tf.keras.layers.Dropout(0.2),\n",
        "        tf.keras.layers.GlobalMaxPool1D(),\n",
        "        tf.keras.layers.Dense(32, activation='relu'),\n",
        "        tf.keras.layers.Dropout(0.25),\n",
        "        tf.keras.layers.Dense(16, activation='relu'),\n",
        "        tf.keras.layers.Dropout(0.25),\n",
        "        tf.keras.layers.Dense(8, activation='relu'),\n",
        "        tf.keras.layers.Dropout(0.25),\n",
        "        tf.keras.layers.Dense(4, activation='softmax')\n",
        "])\n",
        "\n",
        "model.compile(\n",
        "    loss='sparse_categorical_crossentropy',\n",
        "    optimizer=tf.keras.optimizers.Adam(lr=1e-2),\n",
        "    metrics=['acc']\n",
        ")\n",
        "\n",
        "model.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential_8\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "embedding_8 (Embedding)      (None, 250, 16)           16000     \n",
            "_________________________________________________________________\n",
            "bidirectional_17 (Bidirectio (None, 250, 256)          148480    \n",
            "_________________________________________________________________\n",
            "dropout_27 (Dropout)         (None, 250, 256)          0         \n",
            "_________________________________________________________________\n",
            "global_max_pooling1d_8 (Glob (None, 256)               0         \n",
            "_________________________________________________________________\n",
            "dense_31 (Dense)             (None, 32)                8224      \n",
            "_________________________________________________________________\n",
            "dropout_28 (Dropout)         (None, 32)                0         \n",
            "_________________________________________________________________\n",
            "dense_32 (Dense)             (None, 16)                528       \n",
            "_________________________________________________________________\n",
            "dropout_29 (Dropout)         (None, 16)                0         \n",
            "_________________________________________________________________\n",
            "dense_33 (Dense)             (None, 8)                 136       \n",
            "_________________________________________________________________\n",
            "dropout_30 (Dropout)         (None, 8)                 0         \n",
            "_________________________________________________________________\n",
            "dense_34 (Dense)             (None, 4)                 36        \n",
            "=================================================================\n",
            "Total params: 173,404\n",
            "Trainable params: 173,404\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1CVMNrRnxSsA"
      },
      "source": [
        "### Training"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jZH8DKCvxSa7",
        "outputId": "386744f7-9e87-44c2-a0fa-84515c4f8cc8",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "history = model.fit(train_padded, train_label,\n",
        "                    epochs=300,\n",
        "                    batch_size=2048,\n",
        "                    validation_data=(test_padded, test_label),\n",
        "                    callbacks=[reduce_on_plateau, early_stop])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/300\n",
            "50/50 [==============================] - 21s 421ms/step - loss: 1.3204 - acc: 0.3255 - val_loss: 1.0839 - val_acc: 0.4642\n",
            "Epoch 2/300\n",
            "50/50 [==============================] - 20s 391ms/step - loss: 1.0134 - acc: 0.4694 - val_loss: 0.9208 - val_acc: 0.5432\n",
            "Epoch 3/300\n",
            "50/50 [==============================] - 19s 383ms/step - loss: 0.8737 - acc: 0.5962 - val_loss: 0.7570 - val_acc: 0.6534\n",
            "Epoch 4/300\n",
            "50/50 [==============================] - 19s 376ms/step - loss: 0.7967 - acc: 0.6479 - val_loss: 0.6931 - val_acc: 0.7180\n",
            "Epoch 5/300\n",
            "50/50 [==============================] - 19s 374ms/step - loss: 0.7503 - acc: 0.7065 - val_loss: 0.6341 - val_acc: 0.7820\n",
            "Epoch 6/300\n",
            "50/50 [==============================] - 19s 383ms/step - loss: 0.7030 - acc: 0.7482 - val_loss: 0.5871 - val_acc: 0.7939\n",
            "Epoch 7/300\n",
            "50/50 [==============================] - 19s 386ms/step - loss: 0.6662 - acc: 0.7667 - val_loss: 0.5612 - val_acc: 0.8169\n",
            "Epoch 8/300\n",
            "50/50 [==============================] - 19s 386ms/step - loss: 0.6486 - acc: 0.7736 - val_loss: 0.5693 - val_acc: 0.7924\n",
            "Epoch 9/300\n",
            "50/50 [==============================] - 19s 387ms/step - loss: 0.6337 - acc: 0.7846 - val_loss: 0.5365 - val_acc: 0.8210\n",
            "Epoch 10/300\n",
            "50/50 [==============================] - 19s 384ms/step - loss: 0.6164 - acc: 0.7934 - val_loss: 0.5212 - val_acc: 0.8324\n",
            "Epoch 11/300\n",
            "50/50 [==============================] - 19s 382ms/step - loss: 0.6084 - acc: 0.7965 - val_loss: 0.5518 - val_acc: 0.7976\n",
            "Epoch 12/300\n",
            "50/50 [==============================] - 19s 380ms/step - loss: 0.6098 - acc: 0.7959 - val_loss: 0.5497 - val_acc: 0.8058\n",
            "Epoch 13/300\n",
            "50/50 [==============================] - 19s 384ms/step - loss: 0.5967 - acc: 0.8022 - val_loss: 0.5380 - val_acc: 0.8170\n",
            "Epoch 14/300\n",
            "50/50 [==============================] - 19s 388ms/step - loss: 0.5916 - acc: 0.8024 - val_loss: 0.5498 - val_acc: 0.8040\n",
            "Epoch 15/300\n",
            "50/50 [==============================] - 19s 388ms/step - loss: 0.5810 - acc: 0.8078 - val_loss: 0.5217 - val_acc: 0.8276\n",
            "Epoch 16/300\n",
            "50/50 [==============================] - 19s 388ms/step - loss: 0.5724 - acc: 0.8123 - val_loss: 0.5641 - val_acc: 0.8003\n",
            "Epoch 17/300\n",
            "50/50 [==============================] - 19s 385ms/step - loss: 0.5659 - acc: 0.8129 - val_loss: 0.5218 - val_acc: 0.8171\n",
            "Epoch 18/300\n",
            "50/50 [==============================] - 19s 381ms/step - loss: 0.5577 - acc: 0.8148 - val_loss: 0.5246 - val_acc: 0.8246\n",
            "Epoch 19/300\n",
            "50/50 [==============================] - 19s 381ms/step - loss: 0.5510 - acc: 0.8174 - val_loss: 0.4980 - val_acc: 0.8363\n",
            "Epoch 20/300\n",
            "50/50 [==============================] - 19s 385ms/step - loss: 0.5491 - acc: 0.8185 - val_loss: 0.5215 - val_acc: 0.8166\n",
            "Epoch 21/300\n",
            "50/50 [==============================] - 19s 389ms/step - loss: 0.5464 - acc: 0.8182 - val_loss: 0.4855 - val_acc: 0.8466\n",
            "Epoch 22/300\n",
            "50/50 [==============================] - 19s 389ms/step - loss: 0.5387 - acc: 0.8221 - val_loss: 0.4766 - val_acc: 0.8484\n",
            "Epoch 23/300\n",
            "50/50 [==============================] - 19s 388ms/step - loss: 0.5279 - acc: 0.8239 - val_loss: 0.5094 - val_acc: 0.8277\n",
            "Epoch 24/300\n",
            "50/50 [==============================] - 19s 387ms/step - loss: 0.5287 - acc: 0.8240 - val_loss: 0.5071 - val_acc: 0.8319\n",
            "Epoch 25/300\n",
            "50/50 [==============================] - 19s 381ms/step - loss: 0.5219 - acc: 0.8245 - val_loss: 0.4865 - val_acc: 0.8446\n",
            "Epoch 26/300\n",
            "50/50 [==============================] - 19s 380ms/step - loss: 0.5202 - acc: 0.8249 - val_loss: 0.5193 - val_acc: 0.8413\n",
            "Epoch 27/300\n",
            "50/50 [==============================] - 19s 385ms/step - loss: 0.5194 - acc: 0.8261 - val_loss: 0.4918 - val_acc: 0.8432\n",
            "Epoch 28/300\n",
            "50/50 [==============================] - 19s 389ms/step - loss: 0.5127 - acc: 0.8281 - val_loss: 0.5217 - val_acc: 0.8240\n",
            "Epoch 29/300\n",
            "50/50 [==============================] - 19s 388ms/step - loss: 0.5048 - acc: 0.8287 - val_loss: 0.5291 - val_acc: 0.8343\n",
            "Epoch 30/300\n",
            "50/50 [==============================] - 19s 389ms/step - loss: 0.5032 - acc: 0.8325 - val_loss: 0.5082 - val_acc: 0.8424\n",
            "Epoch 31/300\n",
            "50/50 [==============================] - 19s 387ms/step - loss: 0.4965 - acc: 0.8327 - val_loss: 0.5304 - val_acc: 0.8320\n",
            "Epoch 32/300\n",
            "50/50 [==============================] - 19s 382ms/step - loss: 0.4986 - acc: 0.8328 - val_loss: 0.5114 - val_acc: 0.8446\n",
            "Epoch 33/300\n",
            "50/50 [==============================] - 19s 382ms/step - loss: 0.4858 - acc: 0.8353 - val_loss: 0.5159 - val_acc: 0.8420\n",
            "Epoch 34/300\n",
            "50/50 [==============================] - 19s 383ms/step - loss: 0.4869 - acc: 0.8361 - val_loss: 0.5113 - val_acc: 0.8493\n",
            "Epoch 35/300\n",
            "50/50 [==============================] - 19s 388ms/step - loss: 0.4830 - acc: 0.8367 - val_loss: 0.5177 - val_acc: 0.8441\n",
            "Epoch 36/300\n",
            "50/50 [==============================] - 19s 387ms/step - loss: 0.4696 - acc: 0.8407 - val_loss: 0.5116 - val_acc: 0.8355\n",
            "Epoch 37/300\n",
            "50/50 [==============================] - 19s 390ms/step - loss: 0.4709 - acc: 0.8420 - val_loss: 0.5305 - val_acc: 0.8372\n",
            "Epoch 38/300\n",
            "50/50 [==============================] - 19s 386ms/step - loss: 0.4632 - acc: 0.8425 - val_loss: 0.5453 - val_acc: 0.8378\n",
            "Epoch 39/300\n",
            "50/50 [==============================] - 19s 380ms/step - loss: 0.4622 - acc: 0.8425 - val_loss: 0.5643 - val_acc: 0.8186\n",
            "Epoch 40/300\n",
            "50/50 [==============================] - 19s 381ms/step - loss: 0.4623 - acc: 0.8431 - val_loss: 0.5627 - val_acc: 0.8402\n",
            "Epoch 41/300\n",
            "50/50 [==============================] - 19s 384ms/step - loss: 0.4645 - acc: 0.8424 - val_loss: 0.5629 - val_acc: 0.8172\n",
            "Epoch 42/300\n",
            "50/50 [==============================] - 19s 389ms/step - loss: 0.4687 - acc: 0.8400 - val_loss: 0.5432 - val_acc: 0.8525\n",
            "Epoch 43/300\n",
            "50/50 [==============================] - 19s 388ms/step - loss: 0.4617 - acc: 0.8418 - val_loss: 0.5401 - val_acc: 0.8420\n",
            "Epoch 44/300\n",
            "50/50 [==============================] - 20s 391ms/step - loss: 0.4533 - acc: 0.8448 - val_loss: 0.6138 - val_acc: 0.8384\n",
            "Epoch 45/300\n",
            "50/50 [==============================] - 19s 385ms/step - loss: 0.4435 - acc: 0.8486 - val_loss: 0.5514 - val_acc: 0.8422\n",
            "Epoch 46/300\n",
            "50/50 [==============================] - 19s 381ms/step - loss: 0.4467 - acc: 0.8468 - val_loss: 0.5499 - val_acc: 0.8207\n",
            "Epoch 47/300\n",
            "50/50 [==============================] - 19s 379ms/step - loss: 0.4798 - acc: 0.8363 - val_loss: 0.5556 - val_acc: 0.8455\n",
            "Epoch 48/300\n",
            "50/50 [==============================] - 19s 383ms/step - loss: 0.4585 - acc: 0.8436 - val_loss: 0.5420 - val_acc: 0.8482\n",
            "Epoch 49/300\n",
            "50/50 [==============================] - 19s 389ms/step - loss: 0.4463 - acc: 0.8491 - val_loss: 0.5666 - val_acc: 0.8370\n",
            "Epoch 50/300\n",
            "50/50 [==============================] - 19s 390ms/step - loss: 0.4364 - acc: 0.8533 - val_loss: 0.5810 - val_acc: 0.8342\n",
            "Epoch 51/300\n",
            "50/50 [==============================] - 20s 391ms/step - loss: 0.4440 - acc: 0.8510 - val_loss: 0.6072 - val_acc: 0.8484\n",
            "Epoch 52/300\n",
            "50/50 [==============================] - 19s 384ms/step - loss: 0.4363 - acc: 0.8536 - val_loss: 0.5535 - val_acc: 0.8346\n",
            "Epoch 53/300\n",
            "50/50 [==============================] - 19s 381ms/step - loss: 0.4618 - acc: 0.8453 - val_loss: 0.5665 - val_acc: 0.8247\n",
            "Epoch 54/300\n",
            "50/50 [==============================] - 19s 380ms/step - loss: 0.4376 - acc: 0.8529 - val_loss: 0.6108 - val_acc: 0.8298\n",
            "Epoch 55/300\n",
            "50/50 [==============================] - 19s 384ms/step - loss: 0.4290 - acc: 0.8592 - val_loss: 0.5731 - val_acc: 0.8383\n",
            "Epoch 56/300\n",
            "50/50 [==============================] - 19s 388ms/step - loss: 0.4339 - acc: 0.8571 - val_loss: 0.6270 - val_acc: 0.8469\n",
            "Epoch 57/300\n",
            "50/50 [==============================] - 20s 391ms/step - loss: 0.4202 - acc: 0.8590 - val_loss: 0.6136 - val_acc: 0.8477\n",
            "Epoch 58/300\n",
            "50/50 [==============================] - 20s 390ms/step - loss: 0.4172 - acc: 0.8638 - val_loss: 0.6329 - val_acc: 0.8399\n",
            "Epoch 59/300\n",
            "50/50 [==============================] - 19s 385ms/step - loss: 0.4177 - acc: 0.8658 - val_loss: 0.6209 - val_acc: 0.8283\n",
            "Epoch 60/300\n",
            "50/50 [==============================] - 19s 380ms/step - loss: 0.4105 - acc: 0.8692 - val_loss: 0.6301 - val_acc: 0.8472\n",
            "Epoch 61/300\n",
            "50/50 [==============================] - 19s 381ms/step - loss: 0.4344 - acc: 0.8618 - val_loss: 0.6006 - val_acc: 0.8482\n",
            "Epoch 62/300\n",
            "50/50 [==============================] - 19s 384ms/step - loss: 0.4163 - acc: 0.8690 - val_loss: 0.6507 - val_acc: 0.8417\n",
            "Epoch 63/300\n",
            "50/50 [==============================] - 20s 391ms/step - loss: 0.3869 - acc: 0.8787 - val_loss: 0.6858 - val_acc: 0.8446\n",
            "Epoch 64/300\n",
            "50/50 [==============================] - 20s 391ms/step - loss: 0.3718 - acc: 0.8838 - val_loss: 0.7461 - val_acc: 0.8467\n",
            "Epoch 65/300\n",
            "50/50 [==============================] - 19s 389ms/step - loss: 0.3622 - acc: 0.8868 - val_loss: 0.7927 - val_acc: 0.8439\n",
            "Epoch 66/300\n",
            "50/50 [==============================] - 19s 385ms/step - loss: 0.3554 - acc: 0.8899 - val_loss: 0.8227 - val_acc: 0.8444\n",
            "Epoch 67/300\n",
            "50/50 [==============================] - 19s 380ms/step - loss: 0.3496 - acc: 0.8927 - val_loss: 0.8803 - val_acc: 0.8465\n",
            "Epoch 68/300\n",
            "50/50 [==============================] - 19s 381ms/step - loss: 0.3443 - acc: 0.8930 - val_loss: 0.9111 - val_acc: 0.8469\n",
            "Epoch 69/300\n",
            "50/50 [==============================] - 19s 384ms/step - loss: 0.3402 - acc: 0.8943 - val_loss: 0.9368 - val_acc: 0.8424\n",
            "Epoch 70/300\n",
            "50/50 [==============================] - 20s 390ms/step - loss: 0.3368 - acc: 0.8955 - val_loss: 0.9524 - val_acc: 0.8456\n",
            "Epoch 71/300\n",
            "50/50 [==============================] - 20s 390ms/step - loss: 0.3343 - acc: 0.8975 - val_loss: 1.0098 - val_acc: 0.8475\n",
            "Epoch 72/300\n",
            "50/50 [==============================] - 20s 390ms/step - loss: 0.3268 - acc: 0.9000 - val_loss: 1.0184 - val_acc: 0.8462\n",
            "Epoch 73/300\n",
            "50/50 [==============================] - 19s 385ms/step - loss: 0.3279 - acc: 0.8989 - val_loss: 1.0814 - val_acc: 0.8447\n",
            "Epoch 74/300\n",
            "50/50 [==============================] - 19s 380ms/step - loss: 0.3231 - acc: 0.9013 - val_loss: 1.1141 - val_acc: 0.8470\n",
            "Epoch 75/300\n",
            "50/50 [==============================] - 19s 381ms/step - loss: 0.3206 - acc: 0.9024 - val_loss: 1.1369 - val_acc: 0.8471\n",
            "Epoch 76/300\n",
            "50/50 [==============================] - 19s 384ms/step - loss: 0.3205 - acc: 0.9023 - val_loss: 1.1628 - val_acc: 0.8440\n",
            "Epoch 77/300\n",
            "50/50 [==============================] - 20s 391ms/step - loss: 0.3144 - acc: 0.9045 - val_loss: 1.2095 - val_acc: 0.8456\n",
            "Epoch 78/300\n",
            "50/50 [==============================] - 20s 390ms/step - loss: 0.3116 - acc: 0.9053 - val_loss: 1.2791 - val_acc: 0.8472\n",
            "Epoch 79/300\n",
            "50/50 [==============================] - 19s 389ms/step - loss: 0.3122 - acc: 0.9058 - val_loss: 1.2616 - val_acc: 0.8432\n",
            "Epoch 80/300\n",
            "50/50 [==============================] - 19s 385ms/step - loss: 0.3058 - acc: 0.9070 - val_loss: 1.3196 - val_acc: 0.8491\n",
            "Epoch 81/300\n",
            "50/50 [==============================] - 19s 380ms/step - loss: 0.3037 - acc: 0.9076 - val_loss: 1.3778 - val_acc: 0.8470\n",
            "Epoch 82/300\n",
            "50/50 [==============================] - 19s 380ms/step - loss: 0.3032 - acc: 0.9084 - val_loss: 1.3837 - val_acc: 0.8483\n",
            "Epoch 83/300\n",
            "50/50 [==============================] - 19s 384ms/step - loss: 0.3010 - acc: 0.9091 - val_loss: 1.4581 - val_acc: 0.8499\n",
            "Epoch 84/300\n",
            "50/50 [==============================] - 20s 391ms/step - loss: 0.2997 - acc: 0.9100 - val_loss: 1.4913 - val_acc: 0.8483\n",
            "Epoch 85/300\n",
            "50/50 [==============================] - 19s 389ms/step - loss: 0.2948 - acc: 0.9120 - val_loss: 1.5576 - val_acc: 0.8449\n",
            "Epoch 86/300\n",
            "50/50 [==============================] - 19s 389ms/step - loss: 0.2960 - acc: 0.9118 - val_loss: 1.5413 - val_acc: 0.8469\n",
            "Epoch 87/300\n",
            "50/50 [==============================] - 19s 384ms/step - loss: 0.2886 - acc: 0.9135 - val_loss: 1.5850 - val_acc: 0.8486\n",
            "Epoch 88/300\n",
            "50/50 [==============================] - 19s 380ms/step - loss: 0.2880 - acc: 0.9141 - val_loss: 1.6076 - val_acc: 0.8487\n",
            "Epoch 89/300\n",
            "50/50 [==============================] - 19s 380ms/step - loss: 0.2850 - acc: 0.9148 - val_loss: 1.6099 - val_acc: 0.8485\n",
            "Epoch 90/300\n",
            "50/50 [==============================] - 19s 385ms/step - loss: 0.2869 - acc: 0.9147 - val_loss: 1.6301 - val_acc: 0.8494\n",
            "Epoch 91/300\n",
            "50/50 [==============================] - 19s 390ms/step - loss: 0.2864 - acc: 0.9145 - val_loss: 1.6329 - val_acc: 0.8482\n",
            "Epoch 92/300\n",
            "50/50 [==============================] - 19s 389ms/step - loss: 0.2855 - acc: 0.9144 - val_loss: 1.6434 - val_acc: 0.8493\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6CgqvpzKxlv8"
      },
      "source": [
        "## Plot training result"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "uAUx5m0ox0EP"
      },
      "source": [
        "def plot(history):\n",
        "  graf = history.history\n",
        "  graf['epoch'] = history.epoch\n",
        "\n",
        "  fig, ax = plt.subplots(nrows=1, ncols=3, figsize=(20, 5))\n",
        "\n",
        "  ax[0].plot(graf['epoch'], graf['loss'], label='Loss')\n",
        "  ax[0].plot(graf['epoch'], graf['val_loss'], label='Val Loss', color='orange')\n",
        "\n",
        "  ax[0].set_xlabel('Epochs')\n",
        "  ax[0].set_ylabel('Score')\n",
        "  ax[0].set_title('Loss')\n",
        "  ax[0].legend()\n",
        "\n",
        "  ax[1].plot(graf['epoch'], graf['acc'], label='Accuracy')\n",
        "  ax[1].plot(graf['epoch'], graf['val_acc'], label='Val Accuracy', color='orange')\n",
        "\n",
        "  ax[1].set_xlabel('Epochs')\n",
        "  ax[1].set_ylabel('Score')\n",
        "  ax[1].set_title('Accuracy')\n",
        "  ax[1].legend()\n",
        "\n",
        "  ax[2].plot(graf['epoch'], graf['lr'], label='Learning Rate')\n",
        "\n",
        "  ax[2].set_xlabel('Epochs')\n",
        "  ax[2].set_ylabel('Value')\n",
        "  ax[2].set_title('Learning Rate')\n",
        "\n",
        "  return plt.show()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4_LxXuqd4mFa",
        "outputId": "08e2f78d-caac-4b12-90aa-2bbc51ae0371",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 286
        }
      },
      "source": [
        "plot(history)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1440x360 with 3 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hxXArgwc4nTE",
        "outputId": "be624940-904c-47c0-c096-9f56df457c28",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "print(\"Overfitting dikit :((\")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Overfitting dikit :((\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-NUbq3EkFsMt"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}