domain_classification_unbalanced_data_.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "domain_classification_unbalanced_data .ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "authorship_tag": "ABX9TyMZGc5W9yJLFa2Ue8aICsIi"
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "CsbZ6I8vhlcM"
      },
      "source": [
        "import pandas as pd\n",
        "import numpy as np\n",
        "import nltk\n",
        "import tensorflow as tf\n",
        "from tensorflow.keras.wrappers.scikit_learn import KerasClassifier\n",
        "from sklearn import model_selection, preprocessing, linear_model, naive_bayes, metrics, svm\n",
        "from sklearn.model_selection import RandomizedSearchCV\n",
        "# from sklearn.externals import joblib\n",
        "from tensorflow.keras.preprocessing.text import Tokenizer\n",
        "from sklearn.model_selection import train_test_split\n",
        "from nltk.tokenize import word_tokenize \n",
        "from sklearn.utils import shuffle\n",
        "from nltk.tokenize.treebank import TreebankWordDetokenizer as Detok\n",
        "from nltk.tokenize.treebank import TreebankWordDetokenizer\n",
        "from sklearn.naive_bayes import MultinomialNB\n",
        "from sklearn.pipeline import Pipeline\n",
        "from sklearn.feature_extraction.text import TfidfTransformer\n",
        "from sklearn.feature_extraction.text import CountVectorizer\n",
        "from sklearn.metrics import accuracy_score\n",
        "from sklearn.metrics import classification_report\n",
        "from sklearn.linear_model import SGDClassifier\n",
        "from tensorflow.keras.preprocessing.sequence import pad_sequences\n",
        "from tensorflow.keras.models import Sequential\n",
        "from tensorflow.keras.callbacks import ModelCheckpoint, EarlyStopping\n",
        "from tensorflow.keras.layers import *\n",
        "import matplotlib.pyplot as plt\n",
        "from tensorflow.keras.regularizers import l2\n",
        "from tensorflow.keras.wrappers.scikit_learn import KerasClassifier\n",
        "from tensorflow.keras.models import load_model\n",
        "import os.path\n",
        "from imblearn.over_sampling import ADASYN, SMOTE, RandomOverSampler\n",
        "from imblearn.pipeline import make_pipeline\n",
        "from sklearn.metrics import confusion_matrix\n",
        "from sklearn.utils.multiclass import unique_labels\n",
        "from sklearn.utils import class_weight\n",
        "from sklearn.neighbors import KNeighborsClassifier\n",
        "import tensorflow as tf\n",
        "\n",
        "from tensorflow.keras.optimizers import Adam"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "resources": {
            "http://localhost:8080/nbextensions/google.colab/files.js": {
              "data": "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",
              "ok": true,
              "headers": [
                [
                  "content-type",
                  "application/javascript"
                ]
              ],
              "status": 200,
              "status_text": ""
            }
          },
          "base_uri": "https://localhost:8080/",
          "height": 73
        },
        "id": "H3jlwMyOhGze",
        "outputId": "0f4392f6-96d4-4e3a-d27b-ff8492775100"
      },
      "source": [
        "from google.colab import files\n",
        "uploaded = files.upload()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              "\n",
              "     <input type=\"file\" id=\"files-d13dfeed-da4b-4815-a14a-384fffef0bda\" name=\"files[]\" multiple disabled\n",
              "        style=\"border:none\" />\n",
              "     <output id=\"result-d13dfeed-da4b-4815-a14a-384fffef0bda\">\n",
              "      Upload widget is only available when the cell has been executed in the\n",
              "      current browser session. Please rerun this cell to enable.\n",
              "      </output>\n",
              "      <script src=\"/nbextensions/google.colab/files.js\"></script> "
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Saving cleaned data-2.csv to cleaned data-2.csv\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 417
        },
        "id": "CXM7mE8jhZx8",
        "outputId": "5e8183f2-5ee0-428e-988c-48b6a95ba2f8"
      },
      "source": [
        "df = pd.read_csv(\"cleaned data-2.csv\", encoding='utf8')\n",
        "df"
      ],
      "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>Id</th>\n",
              "      <th>Text</th>\n",
              "      <th>Class</th>\n",
              "      <th>Filtered_sentence</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0</td>\n",
              "      <td>දින ඒකාබද්ධ විපක්ෂයේ මාධ්‍ය ලේකම් රේණුක පෙරේර...</td>\n",
              "      <td>Political</td>\n",
              "      <td>දින එකාබද්ධ විපක්ෂ මාධ්‍ ලේක රේනුක පෙරේර මහත න...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1</td>\n",
              "      <td>මා UNP හෝ ඒකාබද්ධ විපක්ෂය මගින් හෝ දැනට පවතින ...</td>\n",
              "      <td>Political</td>\n",
              "      <td>මා එක්සත් ජාතික පක්ෂ එකාබද්ධ විපක්ෂ පවති කිසි ...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>2</td>\n",
              "      <td>කිමද මේ තරම් කොහේදෝ ඇති විදපු ගලක් හා උරන මා...</td>\n",
              "      <td>Neutral</td>\n",
              "      <td>කිම තර කොහේද විද ගල උරන මා මිතුර පරය</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>3</td>\n",
              "      <td>එහෙම කොහොමද උනේ 😂😂 පරය අපිව අයින් කරල නේ 😂</td>\n",
              "      <td>Neutral</td>\n",
              "      <td>එහෙ කොහොම උනේ 😂😂 පරය අපි අය කරල නේ 😂</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>4</td>\n",
              "      <td>විමලයා ගොනෙක් වගෙ stage එකෙ මෑරෙන්න හදනවා පරය ...</td>\n",
              "      <td>Political</td>\n",
              "      <td>විමල ගොන වගෙ  එකෙ මෑරෙ හද පරය රටට වෑඩ කර යකො හොර</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1674</th>\n",
              "      <td>1705</td>\n",
              "      <td>පීටීඒ සහ අයිසීසීපීආර් පනත යටතේ වැරදි ලෙස ගොනු ...</td>\n",
              "      <td>Racist</td>\n",
              "      <td>පීටීඑ අයිසීසීපීඅර පනත යටත වැරදි ගො කර විමර්ෂ න...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1675</th>\n",
              "      <td>1706</td>\n",
              "      <td>අරාබි රටවල් මුස්ලිම් නොවන සඳහා සුපිරි වෙළඳසැල්...</td>\n",
              "      <td>Racist</td>\n",
              "      <td>අරාබි රට මුස්ලි නොව සුපිරි වෙලදසැල වල තිබ උරු ...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1676</th>\n",
              "      <td>1707</td>\n",
              "      <td>හැවිරිදි මුස්ලිම් පිරිමි ළමයෙක් ඉස්ලාම් භීතික...</td>\n",
              "      <td>Racist</td>\n",
              "      <td>හැවිරිදි මුස්ලි පිරි ලමය ඉස්ලා බීතික ඉලක්ක පත්...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1677</th>\n",
              "      <td>1708</td>\n",
              "      <td>ශ්‍රී ලංකාවේ මුස්ලිම් ප්‍රජාව වෙත එල්ල කරන දේ...</td>\n",
              "      <td>Racist</td>\n",
              "      <td>ලංක මුස්ලි ප්‍රජා එල් කරන දේෂපාල අගමික පීඩන සම...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1678</th>\n",
              "      <td>1709</td>\n",
              "      <td>රාජ්‍ය අංශයේ කාන්තා ඇඳුම් පැළඳුම් සංග්‍රහය පිළ...</td>\n",
              "      <td>Racist</td>\n",
              "      <td>රාජ්‍ අංෂ කාන්ත ඇද පැලද සංග්‍රහ දිනැ පබ්අඩ චක්...</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>1679 rows × 4 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "        Id  ...                                  Filtered_sentence\n",
              "0        0  ...  දින එකාබද්ධ විපක්ෂ මාධ්‍ ලේක රේනුක පෙරේර මහත න...\n",
              "1        1  ...  මා එක්සත් ජාතික පක්ෂ එකාබද්ධ විපක්ෂ පවති කිසි ...\n",
              "2        2  ...               කිම තර කොහේද විද ගල උරන මා මිතුර පරය\n",
              "3        3  ...               එහෙ කොහොම උනේ 😂😂 පරය අපි අය කරල නේ 😂\n",
              "4        4  ...   විමල ගොන වගෙ  එකෙ මෑරෙ හද පරය රටට වෑඩ කර යකො හොර\n",
              "...    ...  ...                                                ...\n",
              "1674  1705  ...  පීටීඑ අයිසීසීපීඅර පනත යටත වැරදි ගො කර විමර්ෂ න...\n",
              "1675  1706  ...  අරාබි රට මුස්ලි නොව සුපිරි වෙලදසැල වල තිබ උරු ...\n",
              "1676  1707  ...  හැවිරිදි මුස්ලි පිරි ලමය ඉස්ලා බීතික ඉලක්ක පත්...\n",
              "1677  1708  ...  ලංක මුස්ලි ප්‍රජා එල් කරන දේෂපාල අගමික පීඩන සම...\n",
              "1678  1709  ...  රාජ්‍ අංෂ කාන්ත ඇද පැලද සංග්‍රහ දිනැ පබ්අඩ චක්...\n",
              "\n",
              "[1679 rows x 4 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 59
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "6LfHRW8f_1vn",
        "outputId": "c4b83bbc-19c2-4387-9a5f-1e59a1b4b824"
      },
      "source": [
        "%matplotlib inline\n",
        "df['Class'].value_counts().plot(kind=\"bar\")\n",
        "\n",
        "def plot_history(history):\n",
        "    plt.title('Data-Set')\n",
        "    plt.plot(history.history['loss'], label='train')\n",
        "    plt.plot(history.history['val_loss'], label='test')\n",
        "    plt.legend()\n",
        "    plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LodxXD3ehaub"
      },
      "source": [
        "class_to_num = {\"Class\": {'Religion': 5,\n",
        "                'Racist': 4,\n",
        "                'Political': 3,\n",
        "                'sports':2,\n",
        "                'Sexism': 1,\n",
        "                'Neutral':0\n",
        "                }}"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 417
        },
        "id": "5IUcLjTVh2R1",
        "outputId": "a0a41b12-2cba-4767-ef9e-f0cb632ed684"
      },
      "source": [
        "df = df.replace(class_to_num)\n",
        "df"
      ],
      "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>Id</th>\n",
              "      <th>Text</th>\n",
              "      <th>Class</th>\n",
              "      <th>Filtered_sentence</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0</td>\n",
              "      <td>දින ඒකාබද්ධ විපක්ෂයේ මාධ්‍ය ලේකම් රේණුක පෙරේර...</td>\n",
              "      <td>3</td>\n",
              "      <td>දින එකාබද්ධ විපක්ෂ මාධ්‍ ලේක රේනුක පෙරේර මහත න...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1</td>\n",
              "      <td>මා UNP හෝ ඒකාබද්ධ විපක්ෂය මගින් හෝ දැනට පවතින ...</td>\n",
              "      <td>3</td>\n",
              "      <td>මා එක්සත් ජාතික පක්ෂ එකාබද්ධ විපක්ෂ පවති කිසි ...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>2</td>\n",
              "      <td>කිමද මේ තරම් කොහේදෝ ඇති විදපු ගලක් හා උරන මා...</td>\n",
              "      <td>0</td>\n",
              "      <td>කිම තර කොහේද විද ගල උරන මා මිතුර පරය</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>3</td>\n",
              "      <td>එහෙම කොහොමද උනේ 😂😂 පරය අපිව අයින් කරල නේ 😂</td>\n",
              "      <td>0</td>\n",
              "      <td>එහෙ කොහොම උනේ 😂😂 පරය අපි අය කරල නේ 😂</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>4</td>\n",
              "      <td>විමලයා ගොනෙක් වගෙ stage එකෙ මෑරෙන්න හදනවා පරය ...</td>\n",
              "      <td>3</td>\n",
              "      <td>විමල ගොන වගෙ  එකෙ මෑරෙ හද පරය රටට වෑඩ කර යකො හොර</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1674</th>\n",
              "      <td>1705</td>\n",
              "      <td>පීටීඒ සහ අයිසීසීපීආර් පනත යටතේ වැරදි ලෙස ගොනු ...</td>\n",
              "      <td>4</td>\n",
              "      <td>පීටීඑ අයිසීසීපීඅර පනත යටත වැරදි ගො කර විමර්ෂ න...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1675</th>\n",
              "      <td>1706</td>\n",
              "      <td>අරාබි රටවල් මුස්ලිම් නොවන සඳහා සුපිරි වෙළඳසැල්...</td>\n",
              "      <td>4</td>\n",
              "      <td>අරාබි රට මුස්ලි නොව සුපිරි වෙලදසැල වල තිබ උරු ...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1676</th>\n",
              "      <td>1707</td>\n",
              "      <td>හැවිරිදි මුස්ලිම් පිරිමි ළමයෙක් ඉස්ලාම් භීතික...</td>\n",
              "      <td>4</td>\n",
              "      <td>හැවිරිදි මුස්ලි පිරි ලමය ඉස්ලා බීතික ඉලක්ක පත්...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1677</th>\n",
              "      <td>1708</td>\n",
              "      <td>ශ්‍රී ලංකාවේ මුස්ලිම් ප්‍රජාව වෙත එල්ල කරන දේ...</td>\n",
              "      <td>4</td>\n",
              "      <td>ලංක මුස්ලි ප්‍රජා එල් කරන දේෂපාල අගමික පීඩන සම...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1678</th>\n",
              "      <td>1709</td>\n",
              "      <td>රාජ්‍ය අංශයේ කාන්තා ඇඳුම් පැළඳුම් සංග්‍රහය පිළ...</td>\n",
              "      <td>4</td>\n",
              "      <td>රාජ්‍ අංෂ කාන්ත ඇද පැලද සංග්‍රහ දිනැ පබ්අඩ චක්...</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>1679 rows × 4 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "        Id  ...                                  Filtered_sentence\n",
              "0        0  ...  දින එකාබද්ධ විපක්ෂ මාධ්‍ ලේක රේනුක පෙරේර මහත න...\n",
              "1        1  ...  මා එක්සත් ජාතික පක්ෂ එකාබද්ධ විපක්ෂ පවති කිසි ...\n",
              "2        2  ...               කිම තර කොහේද විද ගල උරන මා මිතුර පරය\n",
              "3        3  ...               එහෙ කොහොම උනේ 😂😂 පරය අපි අය කරල නේ 😂\n",
              "4        4  ...   විමල ගොන වගෙ  එකෙ මෑරෙ හද පරය රටට වෑඩ කර යකො හොර\n",
              "...    ...  ...                                                ...\n",
              "1674  1705  ...  පීටීඑ අයිසීසීපීඅර පනත යටත වැරදි ගො කර විමර්ෂ න...\n",
              "1675  1706  ...  අරාබි රට මුස්ලි නොව සුපිරි වෙලදසැල වල තිබ උරු ...\n",
              "1676  1707  ...  හැවිරිදි මුස්ලි පිරි ලමය ඉස්ලා බීතික ඉලක්ක පත්...\n",
              "1677  1708  ...  ලංක මුස්ලි ප්‍රජා එල් කරන දේෂපාල අගමික පීඩන සම...\n",
              "1678  1709  ...  රාජ්‍ අංෂ කාන්ත ඇද පැලද සංග්‍රහ දිනැ පබ්අඩ චක්...\n",
              "\n",
              "[1679 rows x 4 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 61
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Yeg03Jsx13kp"
      },
      "source": [
        "# split into train/test sets\n",
        "sentences = df['Filtered_sentence'].values.astype('U')\n",
        "y = df['Class']\n",
        "X_train, X_test, y_train, y_test = train_test_split(sentences, y, test_size=0.1, random_state=42)\n",
        "\n",
        "tags = ['0','1', '2','3','4','5']"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "N5vR_8bF3YgR"
      },
      "source": [
        "Multinomial Naive Bayes model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4KhGsWkEa7aW",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "8bfbfbc6-6a4b-48bf-da30-6e7eb4aa0ffc"
      },
      "source": [
        "def nb_train(X_train, X_test, y_train, y_test):\n",
        "    nb = Pipeline([('vect', CountVectorizer()),\n",
        "               ('tfidf', TfidfTransformer()),\n",
        "               ('clf', MultinomialNB())])\n",
        "    nb.fit(X_train, y_train)\n",
        "    y_pred = nb.predict(X_test)\n",
        "    print('Accuracy: %s' % accuracy_score(y_pred, y_test)) #test_size=0.1, random_state=42 - Multinomial Naive Bayes model\n",
        "    print(classification_report(y_test, y_pred,target_names=tags))\n",
        "    return nb\n",
        "\n",
        "nb_train(X_train, X_test, y_train, y_test)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Accuracy: 0.625\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.62      1.00      0.77       103\n",
            "           1       0.00      0.00      0.00        14\n",
            "           2       0.00      0.00      0.00         9\n",
            "           3       1.00      0.08      0.14        13\n",
            "           4       0.00      0.00      0.00        18\n",
            "           5       1.00      0.09      0.17        11\n",
            "\n",
            "    accuracy                           0.62       168\n",
            "   macro avg       0.44      0.19      0.18       168\n",
            "weighted avg       0.52      0.62      0.49       168\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/metrics/_classification.py:1272: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Pipeline(memory=None,\n",
              "         steps=[('vect',\n",
              "                 CountVectorizer(analyzer='word', binary=False,\n",
              "                                 decode_error='strict',\n",
              "                                 dtype=<class 'numpy.int64'>, encoding='utf-8',\n",
              "                                 input='content', lowercase=True, max_df=1.0,\n",
              "                                 max_features=None, min_df=1,\n",
              "                                 ngram_range=(1, 1), preprocessor=None,\n",
              "                                 stop_words=None, strip_accents=None,\n",
              "                                 token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\n",
              "                                 tokenizer=None, vocabulary=None)),\n",
              "                ('tfidf',\n",
              "                 TfidfTransformer(norm='l2', smooth_idf=True,\n",
              "                                  sublinear_tf=False, use_idf=True)),\n",
              "                ('clf',\n",
              "                 MultinomialNB(alpha=1.0, class_prior=None, fit_prior=True))],\n",
              "         verbose=False)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "lliczkIw244r",
        "outputId": "c0f2b92b-ffe6-4f93-91c0-67a67a8003ec"
      },
      "source": [
        "def nb_train(X_train, X_test, y_train, y_test):\n",
        "    nb = Pipeline([('vect', CountVectorizer()),\n",
        "               ('tfidf', TfidfTransformer()),\n",
        "               ('clf', MultinomialNB())])\n",
        "    nb.fit(X_train, y_train)\n",
        "    y_pred = nb.predict(X_test)\n",
        "    print('Accuracy: %s' % accuracy_score(y_pred, y_test)) #test_size=0.1, random_state=42 - Multinomial Naive Bayes model\n",
        "    print(classification_report(y_test, y_pred,target_names=tags))\n",
        "    return nb\n",
        "\n",
        "nb_train(X_train, X_test, y_train, y_test)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Accuracy: 0.625\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.62      1.00      0.77       103\n",
            "           1       0.00      0.00      0.00        14\n",
            "           2       0.00      0.00      0.00         9\n",
            "           3       1.00      0.08      0.14        13\n",
            "           4       0.00      0.00      0.00        18\n",
            "           5       1.00      0.09      0.17        11\n",
            "\n",
            "    accuracy                           0.62       168\n",
            "   macro avg       0.44      0.19      0.18       168\n",
            "weighted avg       0.52      0.62      0.49       168\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/metrics/_classification.py:1272: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Pipeline(memory=None,\n",
              "         steps=[('vect',\n",
              "                 CountVectorizer(analyzer='word', binary=False,\n",
              "                                 decode_error='strict',\n",
              "                                 dtype=<class 'numpy.int64'>, encoding='utf-8',\n",
              "                                 input='content', lowercase=True, max_df=1.0,\n",
              "                                 max_features=None, min_df=1,\n",
              "                                 ngram_range=(1, 1), preprocessor=None,\n",
              "                                 stop_words=None, strip_accents=None,\n",
              "                                 token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\n",
              "                                 tokenizer=None, vocabulary=None)),\n",
              "                ('tfidf',\n",
              "                 TfidfTransformer(norm='l2', smooth_idf=True,\n",
              "                                  sublinear_tf=False, use_idf=True)),\n",
              "                ('clf',\n",
              "                 MultinomialNB(alpha=1.0, class_prior=None, fit_prior=True))],\n",
              "         verbose=False)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 43
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "TuMS7Fwd3iyv",
        "outputId": "30976122-9579-4d02-d028-4463796407c0"
      },
      "source": [
        "#KNN\n",
        "def knn_train(X_train, X_test, y_train, y_test):\n",
        "    knn = Pipeline([('vect', CountVectorizer()),\n",
        "               ('tfidf', TfidfTransformer()),\n",
        "               ('clf', KNeighborsClassifier(n_neighbors=17, p=6,metric='euclidean'))])\n",
        "    knn.fit(X_train, y_train)\n",
        "    y_pred = knn.predict(X_test)\n",
        "    print('Accuracy: %s' % accuracy_score(y_pred, y_test)) #test_size=0.1, random_state=42 - Multinomial Naive Bayes model\n",
        "    print(classification_report(y_test, y_pred,target_names=tags))\n",
        "    return knn\n",
        "\n",
        "knn_train(X_train, X_test, y_train, y_test)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Accuracy: 0.6130952380952381\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.61      1.00      0.76       103\n",
            "           1       0.00      0.00      0.00        14\n",
            "           2       0.00      0.00      0.00         9\n",
            "           3       0.00      0.00      0.00        13\n",
            "           4       0.00      0.00      0.00        18\n",
            "           5       0.00      0.00      0.00        11\n",
            "\n",
            "    accuracy                           0.61       168\n",
            "   macro avg       0.10      0.17      0.13       168\n",
            "weighted avg       0.38      0.61      0.47       168\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/metrics/_classification.py:1272: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Pipeline(memory=None,\n",
              "         steps=[('vect',\n",
              "                 CountVectorizer(analyzer='word', binary=False,\n",
              "                                 decode_error='strict',\n",
              "                                 dtype=<class 'numpy.int64'>, encoding='utf-8',\n",
              "                                 input='content', lowercase=True, max_df=1.0,\n",
              "                                 max_features=None, min_df=1,\n",
              "                                 ngram_range=(1, 1), preprocessor=None,\n",
              "                                 stop_words=None, strip_accents=None,\n",
              "                                 token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\n",
              "                                 tokenizer=None, vocabulary=None)),\n",
              "                ('tfidf',\n",
              "                 TfidfTransformer(norm='l2', smooth_idf=True,\n",
              "                                  sublinear_tf=False, use_idf=True)),\n",
              "                ('clf',\n",
              "                 KNeighborsClassifier(algorithm='auto', leaf_size=30,\n",
              "                                      metric='euclidean', metric_params=None,\n",
              "                                      n_jobs=None, n_neighbors=17, p=6,\n",
              "                                      weights='uniform'))],\n",
              "         verbose=False)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 84
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "aPv1kBB-AOMe",
        "outputId": "62a3fb1b-c218-4539-9cdc-694d67a2b6db"
      },
      "source": [
        "#RF\n",
        "from sklearn import decomposition, ensemble\n",
        "def nb_train(X_train, X_test, y_train, y_test):\n",
        "    nb = Pipeline([('vect', CountVectorizer()),\n",
        "               ('tfidf', TfidfTransformer()),\n",
        "               ('clf', ensemble.RandomForestClassifier())])\n",
        "    nb.fit(X_train, y_train)\n",
        "    y_pred = nb.predict(X_test)\n",
        "    print('Accuracy: %s' % accuracy_score(y_pred, y_test)) #test_size=0.1, random_state=42 - Multinomial Naive Bayes model\n",
        "    print(classification_report(y_test, y_pred,target_names=tags))\n",
        "    return nb\n",
        "\n",
        "nb_train(X_train, X_test, y_train, y_test)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Accuracy: 0.6726190476190477\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.69      0.92      0.79       103\n",
            "           1       0.00      0.00      0.00        14\n",
            "           2       0.83      0.56      0.67         9\n",
            "           3       0.44      0.31      0.36        13\n",
            "           4       0.86      0.33      0.48        18\n",
            "           5       0.75      0.27      0.40        11\n",
            "\n",
            "    accuracy                           0.67       168\n",
            "   macro avg       0.60      0.40      0.45       168\n",
            "weighted avg       0.64      0.67      0.62       168\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Pipeline(memory=None,\n",
              "         steps=[('vect',\n",
              "                 CountVectorizer(analyzer='word', binary=False,\n",
              "                                 decode_error='strict',\n",
              "                                 dtype=<class 'numpy.int64'>, encoding='utf-8',\n",
              "                                 input='content', lowercase=True, max_df=1.0,\n",
              "                                 max_features=None, min_df=1,\n",
              "                                 ngram_range=(1, 1), preprocessor=None,\n",
              "                                 stop_words=None, strip_accents=None,\n",
              "                                 token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\n",
              "                                 tokenizer=None, vocabulary=Non...\n",
              "                 RandomForestClassifier(bootstrap=True, ccp_alpha=0.0,\n",
              "                                        class_weight=None, criterion='gini',\n",
              "                                        max_depth=None, max_features='auto',\n",
              "                                        max_leaf_nodes=None, max_samples=None,\n",
              "                                        min_impurity_decrease=0.0,\n",
              "                                        min_impurity_split=None,\n",
              "                                        min_samples_leaf=1, min_samples_split=2,\n",
              "                                        min_weight_fraction_leaf=0.0,\n",
              "                                        n_estimators=100, n_jobs=None,\n",
              "                                        oob_score=False, random_state=None,\n",
              "                                        verbose=0, warm_start=False))],\n",
              "         verbose=False)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 85
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "bKQy3XkTAVGJ",
        "outputId": "283f617f-b061-4f41-c99d-e27ae436edd8"
      },
      "source": [
        "#LR\n",
        "from sklearn import linear_model\n",
        "def nb_train(X_train, X_test, y_train, y_test):\n",
        "    nb = Pipeline([('vect', CountVectorizer()),\n",
        "               ('tfidf', TfidfTransformer()),\n",
        "               ('clf', linear_model.LogisticRegression())])\n",
        "    nb.fit(X_train, y_train)\n",
        "    y_pred = nb.predict(X_test)\n",
        "    print('Accuracy: %s' % accuracy_score(y_pred, y_test)) #test_size=0.1, random_state=42 - Multinomial Naive Bayes model\n",
        "    print(classification_report(y_test, y_pred,target_names=tags))\n",
        "    return nb\n",
        "\n",
        "nb_train(X_train, X_test, y_train, y_test)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Accuracy: 0.7023809523809523\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.69      0.99      0.82       103\n",
            "           1       0.00      0.00      0.00        14\n",
            "           2       0.71      0.56      0.63         9\n",
            "           3       0.75      0.23      0.35        13\n",
            "           4       0.71      0.28      0.40        18\n",
            "           5       1.00      0.27      0.43        11\n",
            "\n",
            "    accuracy                           0.70       168\n",
            "   macro avg       0.65      0.39      0.44       168\n",
            "weighted avg       0.66      0.70      0.63       168\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/metrics/_classification.py:1272: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Pipeline(memory=None,\n",
              "         steps=[('vect',\n",
              "                 CountVectorizer(analyzer='word', binary=False,\n",
              "                                 decode_error='strict',\n",
              "                                 dtype=<class 'numpy.int64'>, encoding='utf-8',\n",
              "                                 input='content', lowercase=True, max_df=1.0,\n",
              "                                 max_features=None, min_df=1,\n",
              "                                 ngram_range=(1, 1), preprocessor=None,\n",
              "                                 stop_words=None, strip_accents=None,\n",
              "                                 token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\n",
              "                                 tokenizer=None, vocabulary=None)),\n",
              "                ('tfidf',\n",
              "                 TfidfTransformer(norm='l2', smooth_idf=True,\n",
              "                                  sublinear_tf=False, use_idf=True)),\n",
              "                ('clf',\n",
              "                 LogisticRegression(C=1.0, class_weight=None, dual=False,\n",
              "                                    fit_intercept=True, intercept_scaling=1,\n",
              "                                    l1_ratio=None, max_iter=100,\n",
              "                                    multi_class='auto', n_jobs=None,\n",
              "                                    penalty='l2', random_state=None,\n",
              "                                    solver='lbfgs', tol=0.0001, verbose=0,\n",
              "                                    warm_start=False))],\n",
              "         verbose=False)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 86
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "z4d-3-GtyZiN",
        "outputId": "890ff228-a91b-4424-fde4-6c46b9560502"
      },
      "source": [
        "#Svm\n",
        "def svc_train(X_train, X_test, y_train, y_test):\n",
        "    sgd = Pipeline([('vect', CountVectorizer()),\n",
        "                ('tfidf', TfidfTransformer()),\n",
        "                ('clf', svm.SVC()),\n",
        "               ])\n",
        "    sgd.fit(X_train, y_train)\n",
        "    y_pred = sgd.predict(X_test)\n",
        "    print('Accuracy %s' % accuracy_score(y_pred, y_test)) #test_size=0.2, random_state=45-Multinomial Naive Bayes model\n",
        "    print(classification_report(y_test, y_pred,target_names=tags))\n",
        "    return sgd\n",
        "\n",
        "svc_train(X_train, X_test, y_train, y_test)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Accuracy 0.6607142857142857\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.66      0.96      0.78       103\n",
            "           1       0.00      0.00      0.00        14\n",
            "           2       0.75      0.33      0.46         9\n",
            "           3       0.75      0.23      0.35        13\n",
            "           4       0.75      0.17      0.27        18\n",
            "           5       1.00      0.27      0.43        11\n",
            "\n",
            "    accuracy                           0.66       168\n",
            "   macro avg       0.65      0.33      0.38       168\n",
            "weighted avg       0.65      0.66      0.59       168\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Pipeline(memory=None,\n",
              "         steps=[('vect',\n",
              "                 CountVectorizer(analyzer='word', binary=False,\n",
              "                                 decode_error='strict',\n",
              "                                 dtype=<class 'numpy.int64'>, encoding='utf-8',\n",
              "                                 input='content', lowercase=True, max_df=1.0,\n",
              "                                 max_features=None, min_df=1,\n",
              "                                 ngram_range=(1, 1), preprocessor=None,\n",
              "                                 stop_words=None, strip_accents=None,\n",
              "                                 token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\n",
              "                                 tokenizer=None, vocabulary=None)),\n",
              "                ('tfidf',\n",
              "                 TfidfTransformer(norm='l2', smooth_idf=True,\n",
              "                                  sublinear_tf=False, use_idf=True)),\n",
              "                ('clf',\n",
              "                 SVC(C=1.0, break_ties=False, cache_size=200, class_weight=None,\n",
              "                     coef0=0.0, decision_function_shape='ovr', degree=3,\n",
              "                     gamma='scale', kernel='rbf', max_iter=-1,\n",
              "                     probability=False, random_state=None, shrinking=True,\n",
              "                     tol=0.001, verbose=False))],\n",
              "         verbose=False)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "C1OUq-LiAe-g",
        "outputId": "869c0b1c-ac68-420d-d739-a1ef1584040b"
      },
      "source": [
        "#SGD\n",
        "def svc_train(X_train, X_test, y_train, y_test):\n",
        "    sgd = Pipeline([('vect', CountVectorizer()),\n",
        "                ('tfidf', TfidfTransformer()),\n",
        "                ('clf', SGDClassifier(loss='hinge', penalty='l2',alpha=1e-3, random_state=42, max_iter=5, tol=None)),\n",
        "               ])\n",
        "    sgd.fit(X_train, y_train)\n",
        "    y_pred = sgd.predict(X_test)\n",
        "    print('Accuracy %s' % accuracy_score(y_pred, y_test)) #test_size=0.2, random_state=45-Multinomial Naive Bayes model\n",
        "    print(classification_report(y_test, y_pred,target_names=tags))\n",
        "    return sgd\n",
        "\n",
        "svc_train(X_train, X_test, y_train, y_test)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Accuracy 0.6726190476190477\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.68      0.92      0.79       103\n",
            "           1       0.00      0.00      0.00        14\n",
            "           2       0.67      0.67      0.67         9\n",
            "           3       0.57      0.31      0.40        13\n",
            "           4       0.57      0.22      0.32        18\n",
            "           5       0.67      0.36      0.47        11\n",
            "\n",
            "    accuracy                           0.67       168\n",
            "   macro avg       0.53      0.41      0.44       168\n",
            "weighted avg       0.60      0.67      0.61       168\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/metrics/_classification.py:1272: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Pipeline(memory=None,\n",
              "         steps=[('vect',\n",
              "                 CountVectorizer(analyzer='word', binary=False,\n",
              "                                 decode_error='strict',\n",
              "                                 dtype=<class 'numpy.int64'>, encoding='utf-8',\n",
              "                                 input='content', lowercase=True, max_df=1.0,\n",
              "                                 max_features=None, min_df=1,\n",
              "                                 ngram_range=(1, 1), preprocessor=None,\n",
              "                                 stop_words=None, strip_accents=None,\n",
              "                                 token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\n",
              "                                 tokenizer=None, vocabulary=Non...\n",
              "                ('clf',\n",
              "                 SGDClassifier(alpha=0.001, average=False, class_weight=None,\n",
              "                               early_stopping=False, epsilon=0.1, eta0=0.0,\n",
              "                               fit_intercept=True, l1_ratio=0.15,\n",
              "                               learning_rate='optimal', loss='hinge',\n",
              "                               max_iter=5, n_iter_no_change=5, n_jobs=None,\n",
              "                               penalty='l2', power_t=0.5, random_state=42,\n",
              "                               shuffle=True, tol=None, validation_fraction=0.1,\n",
              "                               verbose=0, warm_start=False))],\n",
              "         verbose=False)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 63
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "SY2nw05u3lnL",
        "outputId": "5f4da968-1efe-428d-ae54-1ea549572107"
      },
      "source": [
        "from sklearn import decomposition, ensemble\n",
        "sgd = Pipeline([('vect', CountVectorizer()),\n",
        "                ('tfidf', TfidfTransformer()),\n",
        "                ('clf', linear_model.LogisticRegression()),\n",
        "               ])\n",
        "sgd.fit(X_train, y_train)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Pipeline(memory=None,\n",
              "         steps=[('vect',\n",
              "                 CountVectorizer(analyzer='word', binary=False,\n",
              "                                 decode_error='strict',\n",
              "                                 dtype=<class 'numpy.int64'>, encoding='utf-8',\n",
              "                                 input='content', lowercase=True, max_df=1.0,\n",
              "                                 max_features=None, min_df=1,\n",
              "                                 ngram_range=(1, 1), preprocessor=None,\n",
              "                                 stop_words=None, strip_accents=None,\n",
              "                                 token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\n",
              "                                 tokenizer=None, vocabulary=None)),\n",
              "                ('tfidf',\n",
              "                 TfidfTransformer(norm='l2', smooth_idf=True,\n",
              "                                  sublinear_tf=False, use_idf=True)),\n",
              "                ('clf',\n",
              "                 LogisticRegression(C=1.0, class_weight=None, dual=False,\n",
              "                                    fit_intercept=True, intercept_scaling=1,\n",
              "                                    l1_ratio=None, max_iter=100,\n",
              "                                    multi_class='auto', n_jobs=None,\n",
              "                                    penalty='l2', random_state=None,\n",
              "                                    solver='lbfgs', tol=0.0001, verbose=0,\n",
              "                                    warm_start=False))],\n",
              "         verbose=False)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 64
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VQ_mbeez3hit",
        "outputId": "6caf23c5-36af-494f-cc90-5cdeb660cb0a"
      },
      "source": [
        "sgd.predict(X_test)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([3, 0, 0, 0, 0, 0, 0, 4, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 4, 0, 0,\n",
              "       0, 0, 0, 3, 0, 0, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2,\n",
              "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 0, 0, 0, 0, 0, 3,\n",
              "       0, 0, 0, 0, 0, 4, 0, 0, 5, 2, 0, 0, 0, 0, 0, 0, 0, 3, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 0, 0, 4, 0, 2, 0, 0, 0, 0, 0, 0, 0, 2, 0, 5, 0, 0,\n",
              "       0, 0, 0, 0, 0, 4, 5, 0, 0, 0, 0, 0, 0, 0])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 65
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ccX4crCGJUEU",
        "outputId": "a67a2b54-d5fd-4abf-b7db-585d14a182b4"
      },
      "source": [
        "from pickle import load, dump\n",
        "dump(sgd, open(\"model.pkl\", \"wb\"))\n",
        "model = load(open(\"model.pkl\", \"rb\"))\n",
        "\n",
        "y_pred = model.predict(X_test)\n",
        "\n",
        "print('Accuracy %s' % accuracy_score(y_pred, y_test))\n",
        "\n",
        "print(classification_report(y_test, y_pred,target_names=tags))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Accuracy 0.7023809523809523\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.69      0.99      0.82       103\n",
            "           1       0.00      0.00      0.00        14\n",
            "           2       0.71      0.56      0.63         9\n",
            "           3       0.75      0.23      0.35        13\n",
            "           4       0.71      0.28      0.40        18\n",
            "           5       1.00      0.27      0.43        11\n",
            "\n",
            "    accuracy                           0.70       168\n",
            "   macro avg       0.65      0.39      0.44       168\n",
            "weighted avg       0.66      0.70      0.63       168\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/metrics/_classification.py:1272: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "GZdsNQZd4JN6",
        "outputId": "4af36b1f-e1bc-44d9-9327-16ecfafc9146"
      },
      "source": [
        "y_pred"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([3, 0, 0, 0, 0, 0, 0, 4, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 4, 0, 0,\n",
              "       0, 0, 0, 3, 0, 0, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2,\n",
              "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 0, 0, 0, 0, 0, 3,\n",
              "       0, 0, 0, 0, 0, 4, 0, 0, 5, 2, 0, 0, 0, 0, 0, 0, 0, 3, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 0, 0, 4, 0, 2, 0, 0, 0, 0, 0, 0, 0, 2, 0, 5, 0, 0,\n",
              "       0, 0, 0, 0, 0, 4, 5, 0, 0, 0, 0, 0, 0, 0])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 67
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ptl2CCCN4Npy"
      },
      "source": [
        "labels = ['Sexism','sports','Political','Racist','Religion','Neutral']"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "iItGvjSW4Rt7",
        "outputId": "9e51bf2c-ee25-4bcb-a9d5-5edd0724bcf4"
      },
      "source": [
        "text = \"මහ බැංකුවේ ලයිට් දාලා බයියෝ චූන් කරලා අදට හරියටම අවුරුද්දයි\"\n",
        "\n",
        "pred = model.predict([text])\n",
        "print(labels[pred[0]-1])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Neutral\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 229
        },
        "id": "1184KHbu6BQ4",
        "outputId": "0a5f33b1-6bb8-4dd9-ec87-8516a2b3dfa7"
      },
      "source": [
        "import seaborn as sn\n",
        "sn.heatmap(confusion_matrix(y_test,y_pred), annot=True)\n",
        "plt.xlabel('Predicted')\n",
        "plt.ylabel('Actual')\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "error",
          "ename": "NameError",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-1-ed027ef06462>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mseaborn\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0msn\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 2\u001b[0;31m \u001b[0msn\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mheatmap\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mconfusion_matrix\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0my_test\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0my_pred\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mannot\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mTrue\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      3\u001b[0m \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mxlabel\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'Predicted'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mylabel\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'Actual'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshow\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mNameError\u001b[0m: name 'confusion_matrix' is not defined"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 350
        },
        "id": "ph5X9IeABYXG",
        "outputId": "50dd1440-ca46-466d-e3f0-5380d23d4681"
      },
      "source": [
        "g=plt.scatter(y_test,y_pred)\n",
        "g.axes.set_yscale('log')\n",
        "g.axes.set_xscale('log')\n",
        "g.axes.set_xlabel('True Values ')\n",
        "g.axes.set_ylabel('Predictions ')\n",
        "g.axes.axis('equal')\n",
        "g.axes.axis('square')\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(0.8912509381337456,\n",
              " 11.220184543019636,\n",
              " 0.8912509381337456,\n",
              " 11.220184543019636)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 89
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        },
        "id": "rNNwk9qsBuSm",
        "outputId": "872d6097-16c6-4a39-bead-a688fb3ae96d"
      },
      "source": [
        "g=plt.plot(y_test - y_pred,marker='o',linestyle='')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXIAAAD4CAYAAADxeG0DAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAXGUlEQVR4nO3df4zkdX3H8dd7hzndBXT35ERYOA8ovQaLcLiRa84aq5ZDRLgiKKRUrU0vTdpEgy7hiomY2KDdFK2xqbmqiVYKBj1WjJoTK7ZJk0P3WI6TH8evIjiccggr1tvIsvfuH/P97n13dmZ2vjPfme/3M/N8JJub/e73x+fzuc+8du77/d73be4uAEC4hvJuAACgMwQ5AASOIAeAwBHkABA4ghwAAndMHgc94YQTfMOGDXkcGgCCtXfv3mfdfV3t8lyCfMOGDZqZmcnj0AAQLDP7Wb3lnFoBgMAR5AAQOIIcAAJHkANA4AhyAAhcLnetdGJ6tqKp3Qf09Ny8Th4d1uTWjdq2aTzvZgHIwfRsRZ/49v16/vCCJGl0uKwbLnldqkxIZsorh8syk+YOL6yaL0XKoqCCfHq2oh279mt+YVGSVJmb145d+yWJMAcGzPRsRZPf2KeFxaNPcJ2bX9DkbfsktZYJtZkyN7+w9LNm+VK0LArq1MrU7gNLAxebX1jU1O4DObUIQF6mdh9YFuKxhSPecibUy5SkRvlStCwKKsifnptPtRxA/2r2vm81E1pZr946RcuioIL85NHhVMsB9K9m7/tWM6GV9eqtU7QsCirIJ7du1HC5tGzZcLmkya0bc2oRgLxMbt2ocslWLC8PWcuZUC9TkhrlS9GyKKiLnfFFhKJcKQaQn/h938ldK7WZ0updK0XLIsujZufExITz0CwASMfM9rr7RO3yoE6tAABWIsgBIHAEOQAEjiAHgMAR5AAQOIIcAAJHkANA4AhyAAgcQQ4AgSPIASBwBDkABI4gB4DAZfb0QzMrSZqRVHH3i7PabzNFqpmHYpmereiGO+5fVrprbKSsj78rXT3HQZXFeyu5j9GRstylX8+vXguzV+r1UereEw27mVeZPf3QzK6RNCHpFasFeRZPP6ytmSdVnwd842Vn5z5BkK/p2Yomb9unhSMr53a5ZJq6/BzmSBNZvLfq7SMp7/dqvfaVh0wyLSsfl1U7s8qrrj790MxOkfROSV/MYn+tKFrNPBTH1O4DdUNcqr5JmSPNZfHearcWZq/Ua9/CEV9RAzSrdnY7r7I6R/5ZSddKOtJoBTPbbmYzZjZz6NChjg9YtJp5KI7V5gBzpLks3lvt1sLslaz70u4+shqDjoPczC6W9Iy77222nrvvdPcJd59Yt25dp4ctXM08FMdqc4A50lwW7612a2H2StZ9aXcfWY1BFp/It0i6xMyekHSrpLea2dcy2G9TRauZh+KY3Lqxer6zjnKp9XqOgyqL91a7tTB7pV77ykO2ogZoVu3sdl51fNeKu++QtEOSzOwtkj7q7ld3ut/VFK1mHoojngPctdKeLN5btfso2l0rjfpYb1kW7ex2XmVaszMR5F2/awUABk2ju1Yyu49cktz9R5J+lOU+AQDN8T87ASBwBDkABI4gB4DAEeQAEDiCHAACR5ADQOAIcgAIHEEOAIEjyAEgcAQ5AASOIAeAwGX6rJW8dbuGZ541QgelPmkr/RyUsUA6Wc6LejVfR4fLuuGSYj49s2+CvLYmXmVuXjt27ZekTAa+2/sv6rF7qZV+DspYIJ0s50Wjmq9z8wuavG1fW/vstr45tdLtmnh51ggdlPqkrfRzUMYC6WQ5L5rWfD1SzJqvfRPk3a6Jl2eN0EGpT9pKPwdlLJBOlvMixJqvfRPk3a6Jl2eN0EGpT9pKPwdlLJBOlvMixJqvfRPk3a6Jl2eN0EGpT9pKPwdlLJBOlvOiac3XoWLWfO2bi53dromXZ43QQalP2ko/B2UskE6W86JRzdci37WSac3OVlGzEwDSa1Szs29OrQDAoCLIASBwBDkABI4gB4DAEeQAEDiCHAACR5ADQOAIcgAIHEEOAIEjyAEgcAQ5AASu4yA3s1PN7C4ze8DM7jezD2XRMABAa7J4+uFLkj7i7veY2fGS9prZne7+QAb7BtCnpmcr+sS379fzh6tPGFzt6YJxTc7K3LxKZlp01zhPv5SUQZC7+0FJB6PXvzGzByWNSyLIAdQ1PVvR5Df2aWHx6NNXm9XErK3JuRg9tZWarVWZniM3sw2SNkm6O8v9AugvU7sPLAvxWKOamPVqcsao2ZphkJvZcZK+KenD7v5CnZ9vN7MZM5s5dOhQVocFEKBmdS/r/SzEOpq9lEmQm1lZ1RC/2d131VvH3Xe6+4S7T6xbty6LwwIIVLO6l/V+FmIdzV7K4q4Vk/QlSQ+6+02dNwlAv5vculHl0sq6mI1qYtaryRmjZms2n8i3SPoLSW81s3ujr4sy2C+APrVt07imLj9HYyPlpWWjw2VNXXFO3YuW2zaN68bLztZ49Mm7ZNVfAuOjw7rxsrMH+kKnRM1OAAgGNTsBoE8R5AAQOIIcAAJHkANA4AhyAAgcQQ4AgSPIASBwBDkABI4gB4DAEeQAEDiCHAACl0Wpt5752PR+3XL3U1p0V8lMV51/qj657exMjxGXk3p6bl4nU0YKXZKmzFm1Os59ml840tK+Wyl/VjvP/+QP1umuhw4x77uk27kSTJB/bHq/vrbnyaXvF92Xvs8qzGvLSVFGCt2QpszZ9GxF13z9XrUW4VWrzdt68zz53mLeZ6sXuRLMqZVb7n4q1fJ21CsnRRkpZC1NmbOp3QdShXis2bxtVjatle2RTi9yJZggX2zwuN1Gy9vRqFzUoJeRQrbSlDnrZO51Op+Z99noRa4EE+Txg+RbXd6ORuWiBr2MFLKVpsxZJ3Ov0/nMvM9GL3IlmCC/6vxTUy1vR71yUpSRQtbSlDmb3LqxrTdps3nbrGxaK9sjnV7kSjAXO+MLmt28ayW+8MBdK+imeD61ctdK/H2Wd63Um+fctdI9vcgVSr0BQCAo9QYAfYogB4DAEeQAEDiCHAACR5ADQOAIcgAIHEEOAIEjyAEgcAQ5AASOIAeAwBHkABA4ghwAApfJ0w/N7EJJ/yypJOmL7v6pLPab1KzmXW39w9jYSFnvfP1JuuuhQ6rMzctMip8RZpJc9Z8UV7u/kfKQ1hxT0q/nF5Y9Ka4yN6+SmRbdNTZS1u8WFnU4ekLdcHlILy+X9PzhhaVjxW064bg1euSZ3y4db8sZa3XFxPq6fYiNR8f9zn0HVzwxT5Kuv32/fvti46ovw+UhmbTUPklaUzK9WKdSzZC0rCpNeUg67uVlPX94QUMmHYk2aTYuteJxiv9sNO433HG/5uYXlsbqrJOO157Hn1/advPpY3riV/Mr5kHttseuKalcGlpq24ZXDS/tp7ZN4zVP/xtZU9LhFxfl0Trxo5LjJ2+aScPHDGl+4UjDfpu0tJ92n3aXps5jvbGL537t9o1q3yaXx8abjPHYSFkff9frNPOz53TznieVnEnxnE/Ol9p2Jd8/rdQZbXWsRkfKctfS3338uNgsnz5Y7+/mtpkn9T+PPbe0zonHr9Gz/7ewYs51o9Zwx08/NLOSpIcl/amkn0v6iaSr3P2BRtukffphbc07qfo83xsvqw5Gbf3DtOJ9xZO10/21Ixn2aQxFGx/p/UMsO7Zi3G/bp4WUHRkul/TuN4zr6z9+KvW2vZTsayuazfnafbQ6dsPlks5b/8plYRM789XHLvtwUbtdozGuDepOpB2jWL2xSiqXTHIta3u7x2rleK24evP6tsK8m08/fKOkR939cXd/UdKtki7NYL9LmtW8a1T/MI1k/bws9teOdo94RGGGuFRn3NvoyPzCom65u9ghLqWv0ZimzmOrYze/sFg3xCU1DPF4u0ZjnOWwt1vHcrUapAuLvqLtndTMbKXm6WqyrDUsZXNqZVxSslU/l3R+7Upmtl3Sdklav359qgP0ouZdvC/qFPZWFuOeZd3WbkrTxzRzvhdztldj3E5f2u1/r7dLyno8e3ax0913uvuEu0+sW7cu1bbNat5lVfcu3g91Cnsri3HPsm5rN6XpY5o6j72Ys70a43b60m7/e71dUtbjmUWQVyQlC2eeEi3LTLOad43qH6aRrJ+Xxf7a0e4Rh1Q9TxmiFePeRkeGyyVddf6pbW3bS2lrNKap89jq2A2XS9pyxtq6Pzvz1cc23a7RGGc57O3WsVytBmm5ZCva3knNzFZqnq4my1rDUjZB/hNJZ5rZaWa2RtKVku7IYL9Ltm0a142Xna3x0WGZqlfS4wsV2zaNa+ryczQ2Ul6x3dhIWVdvXq/x6Ddo8pdg/DK5r/hYtfsbKQ9pdLi8dOzkPuPfrGMjZY2Ujw7ncHloaR/JKTQ2Ul7xptlyxlp95r3n1u1DLD5ucp3R4bJueu+5uuk95+rYNasV0x1a1j6petdKPbWTojykpeMm3w/NxqVWPE7xn3XH/YpzNDp8tH9jI2VtOWPtsm23nLF2xTz45LazV2x77JrSsrYl91PbpmTbLdrWEutcvXm9rt68fml9s2rfm/U7uZ/avrai2Zyvt269sUv2Kd7+5r/+o2V9ift35zVvWbY81myMx0bKuuk95+rqzetXfBCJv68N+tr3ZKP5kEbtWI2NlJf93U9dfo6mrjinpbFs53jjo8P67HvPXfFL8sTj19Sdc+1e6Gwmk5qdZnaRpM+qevvhl939H5qtT81OAEiv0V0rmdxH7u7flfTdLPYFAEiH/9kJAIEjyAEgcAQ5AASOIAeAwBHkABA4ghwAAkeQA0DgCHIACBxBDgCBI8gBIHAEOQAELpNnrWAwrVZTMk3NyXbWB1BFkKMttXULK3Pz2rFrvyQt1eBs9vO0+wPQGKdW0JbVakqmqTnZzvoAjiLI0ZbVakqmrbPai7qsQL8iyNGW1WpKpqk52c5yAEcR5GjLajUl09ScbGd9AEdxsRNtiS9ANrrLZLWfp90fgMYyqdmZFjU7ASC9RjU7ObUCAIEjyAEgcAQ5AASOIAeAwBHkABA4ghwAAkeQA0DgCHIACBxBDgCBI8gBIHAEOQAErqMgN7MpM3vIzO4zs9vNbDSrhgEAWtPpJ/I7Jf2hu79e0sOSdnTepP41PVvRlk/9UKdd9x1t+dQPNT1bybtJAPpAR0Hu7t9395eib/dIOqXzJvWnuCZlZW5erqM1KQlzAJ3K8hz5ByV9L8P99RVqUgLollULS5jZDyS9ps6Prnf3b0XrXC/pJUk3N9nPdknbJWn9+vVtNTZk1KQE0C2rBrm7v73Zz83sA5IulvQ2b1Klwt13StopVQtLpGtm+E4eHValTmhTkxJApzq9a+VCSddKusTdD2fTpP5ETUoA3dJpzc7PS3qZpDvNTJL2uPvfdNyqPkRNSgDd0lGQu/vvZdWQQbBt0zjBDSBz/M9OAAgcQQ4AgSPIASBwBDkABI4gB4DAEeQAEDiCHAACR5ADQOAIcgAIHEEOAIEjyAEgcJ0+NCto07OVvn6IVb/3r4gYc+RhYIM8Lr0WV+2JS69J6os3Xr/3r4gYc+RlYE+t9HvptX7vXxEx5sjLwAZ5v5de6/f+FRFjjrwMbJA3KrHWL6XX+r1/RcSYIy8DG+T9Xnqt3/tXRIw58jKwFzv7vfRav/eviBhz5MWaFL7vmomJCZ+Zmen5cQEgZGa2190napcP7KkVAOgXBDkABI4gB4DAEeQAEDiCHAACR5ADQOAIcgAIHEEOAIEjyAEgcAQ5AASOIAeAwBHkABC4TILczD5iZm5mJ2SxPwBA6zoOcjM7VdIFkp7svDkAgLSy+ET+GUnXSur983ABAJ0FuZldKqni7vtaWHe7mc2Y2cyhQ4c6OSwAIGHVCkFm9gNJr6nzo+sl/b2qp1VW5e47Je2UqoUlUrQRANDEqkHu7m+vt9zMzpZ0mqR9ZiZJp0i6x8ze6O6/yLSVAICG2q7Z6e77Jb06/t7MnpA04e7PZtAuAECLuI8cAALX9ifyWu6+Iat9AQBaxydyAAgcQQ4AgSPIASBwBDkABI4gB4DAEeQAEDiCHAACR5ADQOAIcgAIHEEOAIEjyAEgcJk9a6UIpmcrmtp9QE/Pzevk0WFNbt2obZvG824WAHRV3wT59GxFO3bt1/zCoiSpMjevHbv2SxJhDqCv9c2plandB5ZCPDa/sKip3QdyahEA9EbfBPnTc/OplgNAv+ibID95dDjVcgDoF30T5JNbN2q4XFq2bLhc0uTWjTm1CAB6o28udsYXNLlrBcCg6Zsgl6phTnADGDR9c2oFAAYVQQ4AgSPIASBwBDkABI4gB4DAmbv3/qBmhyT9rMXVT5D0bBeb0y20u/dCbTvt7r1Q2/5ad19XuzCXIE/DzGbcfSLvdqRFu3sv1LbT7t4Lue31cGoFAAJHkANA4EII8p15N6BNtLv3Qm077e69kNu+QuHPkQMAmgvhEzkAoAmCHAACV9ggN7MLzeyAmT1qZtfl3Z4kMzvVzO4yswfM7H4z+1C0/AYzq5jZvdHXRYltdkR9OWBmW/NrvWRmT5jZ/qiNM9GytWZ2p5k9Ev05Fi03M/tc1Pb7zOy8nNq8MTGu95rZC2b24aKOuZl92cyeMbOfJpalHmMze3+0/iNm9v6c2j1lZg9FbbvdzEaj5RvMbD4x9l9IbPOGaI49GvXNcmh36rlR5Nxpyt0L9yWpJOkxSadLWiNpn6Sz8m5Xon0nSToven28pIclnSXpBkkfrbP+WVEfXibptKhvpRzb/4SkE2qW/aOk66LX10n6dPT6Iknfk2SSNku6uwDjX5L0C0mvLeqYS3qzpPMk/bTdMZa0VtLj0Z9j0euxHNp9gaRjotefTrR7Q3K9mv38OOqLRX17Rw7tTjU3ip47zb6K+on8jZIedffH3f1FSbdKujTnNi1x94Pufk/0+jeSHpTU7EHol0q61d1/5+7/K+lRVftYJJdK+kr0+iuStiWWf9Wr9kgaNbOT8mhgwtskPebuzf53cK5j7u7/Lem5Om1KM8ZbJd3p7s+5+/OS7pR0Ya/b7e7fd/eXom/3SDql2T6itr/C3fd4NTm/qqN97YoG491Io7lR6NxppqhBPi7pqcT3P1fzoMyNmW2QtEnS3dGiv4v+Cfrl+J/OKl5/XNL3zWyvmW2Plp3o7gej17+QdGL0umhtl6QrJd2S+D6EMZfSj3ER+/BBVT9hx04zs1kz+y8z++No2biqbY3l2e40c6OI492SogZ5EMzsOEnflPRhd39B0r9KOkPSuZIOSvqnHJvXzJvc/TxJ75D0t2b25uQPo09Rhbwv1czWSLpE0m3RolDGfJkij3EjZna9pJck3RwtOihpvbtvknSNpP8ws1fk1b46gpwb7ShqkFcknZr4/pRoWWGYWVnVEL/Z3XdJkrv/0t0X3f2IpH/T0X/KF6o/7l6J/nxG0u2qtvOX8SmT6M9notUL1XZVf/nc4+6/lMIZ80jaMS5MH8zsA5IulvTn0S8hRacmfhW93qvq+eXfj9qYPP2SS7vbmBuFGe+0ihrkP5F0ppmdFn0Cu1LSHTm3aUl0Bf5Lkh5095sSy5Pnjv9MUnwF/Q5JV5rZy8zsNElnqnoxqOfM7FgzOz5+reqFrJ9GbYzvini/pG9Fr++Q9L7ozorNkn6dOD2Qh6uUOK0SwpgnpB3j3ZIuMLOx6LTABdGynjKzCyVdK+kSdz+cWL7OzErR69NVHePHo7a/YGabo/fK+3S0r71sd9q5UejcaSrvq62NvlS9kv+wqr/lr8+7PTVte5Oq/yy+T9K90ddFkv5d0v5o+R2STkpsc33UlwPq8hX8Vdp+uqpX4/dJuj8eW0mvkvSfkh6R9ANJa6PlJulforbvlzSRY9uPlfQrSa9MLCvkmKv6y+agpAVVz7X+VTtjrOo56Uejr7/Mqd2PqnruOJ7rX4jWfXc0h+6VdI+kdyX2M6FqcD4m6fOK/hd5j9udem4UOXeaffFf9AEgcEU9tQIAaBFBDgCBI8gBIHAEOQAEjiAHgMAR5AAQOIIcAAL3/8E4wUwJ1PyCAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 621
        },
        "id": "DLDHp-1gSJFy",
        "outputId": "5b194bf3-bf8f-4662-ad0c-e78290a74e5d"
      },
      "source": [
        "import seaborn as sns\n",
        "from sklearn.metrics import roc_curve, auc\n",
        "\n",
        "def plot_multiclass_roc(clf, X_test, y_test, n_classes, figsize=(17, 6)):\n",
        "    y_score = clf.decision_function(X_test)\n",
        "\n",
        "    # structures\n",
        "    fpr = dict()\n",
        "    tpr = dict()\n",
        "    roc_auc = dict()\n",
        "\n",
        "    # calculate dummies once\n",
        "    y_test_dummies = pd.get_dummies(y_test, drop_first=False).values\n",
        "    for i in range(n_classes):\n",
        "        fpr[i], tpr[i], _ = roc_curve(y_test_dummies[:, i], y_score[:, i])\n",
        "        roc_auc[i] = auc(fpr[i], tpr[i])\n",
        "\n",
        "    # roc for each class\n",
        "    fig, ax = plt.subplots(figsize=figsize)\n",
        "    ax.plot([0, 1], [0, 1], 'k--')\n",
        "    ax.set_xlim([0.0, 1.0])\n",
        "    ax.set_ylim([0.0, 1.05])\n",
        "    ax.set_xlabel('False Positive Rate')\n",
        "    ax.set_ylabel('True Positive Rate')\n",
        "    ax.set_title('Receiver operating characteristic for multiclass classification')\n",
        "    for i in range(n_classes):\n",
        "        ax.plot(fpr[i], tpr[i], label='ROC curve (area = %0.2f) for label %i' % (roc_auc[i], i))\n",
        "    ax.legend(loc=\"best\")\n",
        "    ax.grid(alpha=.4)\n",
        "    sns.despine()\n",
        "    plt.show()\n",
        "\n",
        "plot_multiclass_roc(model, X_test, y_test, n_classes=6, figsize=(16, 10))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1152x720 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    }
  ]
}