NLP_test_4inetnts.ipynb

nlp_test_4inetnts.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "NLP_test_4inetnts.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "authorship_tag": "ABX9TyMeZR0cm5DPoaMna2sOuXu5",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/gist/alyferryhalo/1a7c2d45901bc8fced43afd25bdc2dee/nlp_test_4inetnts.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "aaZO70UE18DA"
      },
      "source": [
        "# Тестовое задание “Классификация интентов”\n",
        "\n",
        "Для разработки бота, надо решить подзадачу классификации интентов. Необходимо определить намерение пользователя.\n",
        "\n",
        "Интенты могут быть:\n",
        "\n",
        "*   **Положительное намерение** - готовность пользователя поддерживать разговор, заинтересованность в предложение итд.\n",
        "*   **Отрицательное намерение** - нежелание общаться, незаинтересованность в предложении итд.\n",
        "*   **Желание общения с живым человеком** - пользователь просит общения с оператором итд.\n",
        "*   **Нейтральный интент** - когда не подходит ничего из вышеперечисленного.\n",
        "\n",
        "Датасет можно получить [по ссылке.](https://drive.google.com/file/d/1CH9rocsHcsykBCfu6wFBwVbedC0M3noT/view)\n",
        "\n",
        "Текст, получен из системы распознавания разговорной речи пользователей.\n",
        "\n",
        "### Классы меток:\n",
        "1.   положительный интент\n",
        "2.   отрицательный интент\n",
        "3.   перевод на оператора\n",
        "4.   нейтральный интент (другое)\n",
        "\n",
        "Нужно решить задачу многоклассовой классификации текста. При решении задачи, предоставляется свобода при выборе модели/моделей, анализе данных. Суть заключается в  демонстрации подхода к решению задачи. "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "tGLlvtf631I_"
      },
      "source": [
        "# Импорты"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Xknbm0_j34A0"
      },
      "source": [
        "import io\n",
        "from io import BytesIO\n",
        "from google.colab import files\n",
        "\n",
        "import re\n",
        "import nltk\n",
        "import keras\n",
        "import codecs\n",
        "import sklearn\n",
        "import itertools\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib\n",
        "import matplotlib.pyplot as plt\n",
        "import matplotlib.patches as mpatches\n",
        "\n",
        "from nltk.tokenize import sent_tokenize, word_tokenize\n",
        "\n",
        "from sklearn.metrics import confusion_matrix\n",
        "from sklearn.decomposition import PCA, TruncatedSVD\n",
        "from sklearn.linear_model import LogisticRegression\n",
        "from sklearn.metrics import accuracy_score, f1_score, precision_score, recall_score, classification_report\n",
        "\n",
        "from keras.preprocessing.text import Tokenizer\n",
        "from keras.preprocessing.sequence import pad_sequences\n",
        "from tensorflow.keras.utils import to_categorical\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.feature_extraction.text import CountVectorizer, TfidfVectorizer"
      ],
      "execution_count": 20,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "NVL0z_h2IErB",
        "outputId": "3fc8cbd1-f5a1-4ae2-90cc-340e4b93226d"
      },
      "source": [
        "nltk.download('punkt')"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[nltk_data] Downloading package punkt to /root/nltk_data...\n",
            "[nltk_data]   Unzipping tokenizers/punkt.zip.\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "True"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 2
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2SnSYCS33KnO"
      },
      "source": [
        "# Загрузка данных"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "IMTSH6hCM2Vq"
      },
      "source": [
        "Прежде всего я открыла датасет у себя, чтобы посмотреть, как он выглядит. Предполагалось, что будут только дв еколонки: данные и метка класса, однако в некоторых наблюдениях случилась проблема, из-за которой данные нормально не загружались: колонок оказалось более двух. Я испраивла это вручную. Полагаю, есть способ быстрее.\n",
        "\n",
        "Также на строке **3950 ваш а ничего не забили\t24.0** обнаружилась ошибка заполнения: метки 24 не существует. Я не была уверена, какой это класс, и решила убрать наблюдение.\n",
        "\n",
        "Ещё замечу, что некоторые наблюдения имели класс 0. Основываясь на своих соображениях, я распределила нули по имеющимся классами.\n",
        "\n",
        "Некоторая часть работы с датасетом была проведена в гугл таблицах."
      ]
    },
    {
      "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": "OK"
            }
          },
          "base_uri": "https://localhost:8080/",
          "height": 72
        },
        "id": "kxccygUf3OoD",
        "outputId": "a39ca600-8da2-4a4f-8aa9-9ae1d5a2025c"
      },
      "source": [
        "uploaded = files.upload()"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              "\n",
              "     <input type=\"file\" id=\"files-68be58f7-a962-4eab-9ced-d86fadd54314\" name=\"files[]\" multiple disabled\n",
              "        style=\"border:none\" />\n",
              "     <output id=\"result-68be58f7-a962-4eab-9ced-d86fadd54314\">\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 data.csv to data.csv\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Hz0vBATP3XbU"
      },
      "source": [
        "data = pd.read_csv(io.BytesIO(uploaded['data.csv']))"
      ],
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1miHq9oO6cJx",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "outputId": "fc5d7a85-ca76-4634-9954-60b54e161a1e"
      },
      "source": [
        "data.head()"
      ],
      "execution_count": 5,
      "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>text</th>\n",
              "      <th>label</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>удалить мой телефон номер у себя не надо</td>\n",
              "      <td>2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>вы о чем вообще алексей вы кто</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>в смысле для чего это</td>\n",
              "      <td>4</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>на счет чего на счет чего</td>\n",
              "      <td>4</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>а у меня минимальная пенсия она не подлежит ув...</td>\n",
              "      <td>4</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                                                text  label\n",
              "0           удалить мой телефон номер у себя не надо      2\n",
              "1                     вы о чем вообще алексей вы кто      1\n",
              "2                              в смысле для чего это      4\n",
              "3                          на счет чего на счет чего      4\n",
              "4  а у меня минимальная пенсия она не подлежит ув...      4"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XcZakFDb6tJk",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 300
        },
        "outputId": "bd574209-b782-47d6-baef-dc879dc74be5"
      },
      "source": [
        "data.describe()"
      ],
      "execution_count": 6,
      "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>label</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>count</th>\n",
              "      <td>99995.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>mean</th>\n",
              "      <td>2.190460</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>std</th>\n",
              "      <td>1.225115</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>min</th>\n",
              "      <td>1.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25%</th>\n",
              "      <td>1.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>50%</th>\n",
              "      <td>2.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>75%</th>\n",
              "      <td>4.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>max</th>\n",
              "      <td>4.000000</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "              label\n",
              "count  99995.000000\n",
              "mean       2.190460\n",
              "std        1.225115\n",
              "min        1.000000\n",
              "25%        1.000000\n",
              "50%        2.000000\n",
              "75%        4.000000\n",
              "max        4.000000"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "JG00NWDSisMW"
      },
      "source": [
        "Проверим балансировку классов:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "o02E5QSp6poj",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "outputId": "cebdc7cf-4f27-41b8-f31f-27c0c43d22ca"
      },
      "source": [
        "data.groupby(\"label\").count()"
      ],
      "execution_count": 7,
      "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>text</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>label</th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>39443</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>29950</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>2713</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>27887</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "        text\n",
              "label       \n",
              "1      39443\n",
              "2      29950\n",
              "3       2713\n",
              "4      27887"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 7
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hqYk-UW9jBtp"
      },
      "source": [
        "В целом, заметно, что класс 3 имеет в 10 раз меньше наблюдений, чем остальные классы. Попробуем пока что обойтись без балансировки классов и продолжим."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1WwPruTN0e0E"
      },
      "source": [
        "В идеале сейчас надо почистить немного датасет и убрать спецсимволы. Пока я смотрела датасет, я заметила, что в нём присутствуют астериксы, которые маскируют нецензурную лексику, однако она является важным показателем для определения класса 2, насколько я тоже заметила. Поэтому, как мне кажется, стоит убрать прочие спецсимволы, включая точки, а поскольку весь текст уже приведён к нижнему регистру, то так и оставим:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "V5SHxsMi0tc7",
        "outputId": "2b9ea42e-4116-48e4-8451-747b3aceec51"
      },
      "source": [
        "def standart_text(df, text_field):\n",
        "    df[text_field] = df[text_field].str.replace(r\"[.,-]\", \"\")\n",
        "    return df\n",
        "\n",
        "data = standart_text(data, \"text\")\n",
        "\n",
        "data.to_csv(\"clean_data.csv\")\n",
        "data.head()"
      ],
      "execution_count": 8,
      "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>text</th>\n",
              "      <th>label</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>удалить мой телефон номер у себя не надо</td>\n",
              "      <td>2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>вы о чем вообще алексей вы кто</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>в смысле для чего это</td>\n",
              "      <td>4</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>на счет чего на счет чего</td>\n",
              "      <td>4</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>а у меня минимальная пенсия она не подлежит ув...</td>\n",
              "      <td>4</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                                                text  label\n",
              "0           удалить мой телефон номер у себя не надо      2\n",
              "1                     вы о чем вообще алексей вы кто      1\n",
              "2                              в смысле для чего это      4\n",
              "3                          на счет чего на счет чего      4\n",
              "4  а у меня минимальная пенсия она не подлежит ув...      4"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "w9PH8p1-0eYR",
        "outputId": "5f48c847-4b4e-4c89-bd0b-1418c774b708"
      },
      "source": [
        "clean_data = pd.read_csv(\"clean_data.csv\")\n",
        "clean_data.tail()"
      ],
      "execution_count": 9,
      "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>Unnamed: 0</th>\n",
              "      <th>text</th>\n",
              "      <th>label</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>99990</th>\n",
              "      <td>99990</td>\n",
              "      <td>а скидки вы можете звонить уже даным давно заб...</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>99991</th>\n",
              "      <td>99991</td>\n",
              "      <td>у думаю что да</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>99992</th>\n",
              "      <td>99992</td>\n",
              "      <td>алло девушка вам нужен торт</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>99993</th>\n",
              "      <td>99993</td>\n",
              "      <td>хорошо я поняла вас ясно поняла передам обязат...</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>99994</th>\n",
              "      <td>99994</td>\n",
              "      <td>ну давайте посмотрим как я мне восемьдесят вос...</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "       Unnamed: 0                                               text  label\n",
              "99990       99990  а скидки вы можете звонить уже даным давно заб...      1\n",
              "99991       99991                                     у думаю что да      1\n",
              "99992       99992                        алло девушка вам нужен торт      1\n",
              "99993       99993  хорошо я поняла вас ясно поняла передам обязат...      1\n",
              "99994       99994  ну давайте посмотрим как я мне восемьдесят вос...      1"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wYAXqZAj4377"
      },
      "source": [
        "# Токенизация"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "EdFVOuSl5Ptt",
        "outputId": "4621e3d2-44c3-43e3-be6e-4076c348fb9f"
      },
      "source": [
        "clean_data.head()"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Unnamed: 0</th>\n",
              "      <th>text</th>\n",
              "      <th>label</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0</td>\n",
              "      <td>удалить мой телефон номер у себя не надо</td>\n",
              "      <td>2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1</td>\n",
              "      <td>вы о чем вообще алексей вы кто</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>2</td>\n",
              "      <td>в смысле для чего это</td>\n",
              "      <td>4</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>3</td>\n",
              "      <td>на счет чего на счет чего</td>\n",
              "      <td>4</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>4</td>\n",
              "      <td>а у меня минимальная пенсия она не подлежит ув...</td>\n",
              "      <td>4</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "   Unnamed: 0                                               text  label\n",
              "0           0           удалить мой телефон номер у себя не надо      2\n",
              "1           1                     вы о чем вообще алексей вы кто      1\n",
              "2           2                              в смысле для чего это      4\n",
              "3           3                          на счет чего на счет чего      4\n",
              "4           4  а у меня минимальная пенсия она не подлежит ув...      4"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "S2SMM3x85VU3",
        "outputId": "ad382cac-c5ff-485f-f283-60f710c30aa3"
      },
      "source": [
        "del clean_data['Unnamed: 0']\n",
        "clean_data.head()"
      ],
      "execution_count": 11,
      "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>text</th>\n",
              "      <th>label</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>удалить мой телефон номер у себя не надо</td>\n",
              "      <td>2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>вы о чем вообще алексей вы кто</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>в смысле для чего это</td>\n",
              "      <td>4</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>на счет чего на счет чего</td>\n",
              "      <td>4</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>а у меня минимальная пенсия она не подлежит ув...</td>\n",
              "      <td>4</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                                                text  label\n",
              "0           удалить мой телефон номер у себя не надо      2\n",
              "1                     вы о чем вообще алексей вы кто      1\n",
              "2                              в смысле для чего это      4\n",
              "3                          на счет чего на счет чего      4\n",
              "4  а у меня минимальная пенсия она не подлежит ув...      4"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 11
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4PF9G-_4jrFk"
      },
      "source": [
        "Проверим, остались ли пустые строки, и удалим их. При токенизации могут возникнуть проблемы с NaN-нами в текста."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "_pEEXe6i6AXG",
        "outputId": "88c661f7-a835-4e99-c3d5-e1e3eda40001"
      },
      "source": [
        "clean_data.isna().sum()"
      ],
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "text     3\n",
              "label    0\n",
              "dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 423
        },
        "id": "elJOlzfq6Sa5",
        "outputId": "e6fcfd86-82c7-4c80-d8f2-0ae963e26e92"
      },
      "source": [
        "clean_data.dropna()"
      ],
      "execution_count": 13,
      "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>text</th>\n",
              "      <th>label</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>удалить мой телефон номер у себя не надо</td>\n",
              "      <td>2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>вы о чем вообще алексей вы кто</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>в смысле для чего это</td>\n",
              "      <td>4</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>на счет чего на счет чего</td>\n",
              "      <td>4</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>а у меня минимальная пенсия она не подлежит ув...</td>\n",
              "      <td>4</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>99990</th>\n",
              "      <td>а скидки вы можете звонить уже даным давно заб...</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>99991</th>\n",
              "      <td>у думаю что да</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>99992</th>\n",
              "      <td>алло девушка вам нужен торт</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>99993</th>\n",
              "      <td>хорошо я поняла вас ясно поняла передам обязат...</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>99994</th>\n",
              "      <td>ну давайте посмотрим как я мне восемьдесят вос...</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>99992 rows × 2 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "                                                    text  label\n",
              "0               удалить мой телефон номер у себя не надо      2\n",
              "1                         вы о чем вообще алексей вы кто      1\n",
              "2                                  в смысле для чего это      4\n",
              "3                              на счет чего на счет чего      4\n",
              "4      а у меня минимальная пенсия она не подлежит ув...      4\n",
              "...                                                  ...    ...\n",
              "99990  а скидки вы можете звонить уже даным давно заб...      1\n",
              "99991                                     у думаю что да      1\n",
              "99992                        алло девушка вам нужен торт      1\n",
              "99993  хорошо я поняла вас ясно поняла передам обязат...      1\n",
              "99994  ну давайте посмотрим как я мне восемьдесят вос...      1\n",
              "\n",
              "[99992 rows x 2 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 13
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vIpJqFV_IYet"
      },
      "source": [
        "clean_data['text'] = clean_data['text'].astype(str)"
      ],
      "execution_count": 14,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "r_H2Qe60jz5x"
      },
      "source": [
        "Добавим, наконец, токены:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "MJVBmV6WracJ"
      },
      "source": [
        "clean_data['tokens'] = clean_data['text'].apply(word_tokenize)"
      ],
      "execution_count": 15,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "Ik4en1YTIh1K",
        "outputId": "3ad81856-3faf-4457-f878-62e898d86854"
      },
      "source": [
        "clean_data.head()"
      ],
      "execution_count": 16,
      "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>text</th>\n",
              "      <th>label</th>\n",
              "      <th>tokens</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>удалить мой телефон номер у себя не надо</td>\n",
              "      <td>2</td>\n",
              "      <td>[удалить, мой, телефон, номер, у, себя, не, надо]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>вы о чем вообще алексей вы кто</td>\n",
              "      <td>1</td>\n",
              "      <td>[вы, о, чем, вообще, алексей, вы, кто]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>в смысле для чего это</td>\n",
              "      <td>4</td>\n",
              "      <td>[в, смысле, для, чего, это]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>на счет чего на счет чего</td>\n",
              "      <td>4</td>\n",
              "      <td>[на, счет, чего, на, счет, чего]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>а у меня минимальная пенсия она не подлежит ув...</td>\n",
              "      <td>4</td>\n",
              "      <td>[а, у, меня, минимальная, пенсия, она, не, под...</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                                                text  ...                                             tokens\n",
              "0           удалить мой телефон номер у себя не надо  ...  [удалить, мой, телефон, номер, у, себя, не, надо]\n",
              "1                     вы о чем вообще алексей вы кто  ...             [вы, о, чем, вообще, алексей, вы, кто]\n",
              "2                              в смысле для чего это  ...                        [в, смысле, для, чего, это]\n",
              "3                          на счет чего на счет чего  ...                   [на, счет, чего, на, счет, чего]\n",
              "4  а у меня минимальная пенсия она не подлежит ув...  ...  [а, у, меня, минимальная, пенсия, она, не, под...\n",
              "\n",
              "[5 rows x 3 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 16
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Jmy8p3cRkDDu"
      },
      "source": [
        "Посмотрим, что получилось со словами, и изучим их немного:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "T9Pt7PXkInGw",
        "outputId": "83814d55-3387-4dfd-81a3-5adddb392b60"
      },
      "source": [
        "all_words = [word for tokens in clean_data[\"tokens\"] for word in tokens]\n",
        "sentence_lengths = [len(tokens) for tokens in clean_data[\"tokens\"]]\n",
        "VOCAB = sorted(list(set(all_words)))\n",
        "print(\"%s words total, with a vocabulary size of %s\" % (len(all_words), len(VOCAB)))\n",
        "print(\"Max sentence length is %s\" % max(sentence_lengths))"
      ],
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "658642 words total, with a vocabulary size of 28559\n",
            "Max sentence length is 34\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 606
        },
        "id": "5wzlTHlkJIIU",
        "outputId": "349ffc77-e0f6-4508-e7e5-7f200a845cfd"
      },
      "source": [
        "fig = plt.figure(figsize=(10, 10)) \n",
        "plt.xlabel('Sentence length')\n",
        "plt.ylabel('Number of sentences')\n",
        "plt.hist(sentence_lengths)\n",
        "plt.show()"
      ],
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x720 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "fkEY_rxmkKm7"
      },
      "source": [
        "Теперь, думаю, можно приступатьк  машинному обучению."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mdmzhGJ0kNPO"
      },
      "source": [
        "# Машинное обучение"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Nf_vVzOAklnL"
      },
      "source": [
        "Для компьютерного представления используются несколько моделей, однако я решила оставиться на мешке слов.\n",
        "\n",
        "**Мешок слов** – это модель представления текста в виде вектора (набора слов). Каждому слову в тексте сопоставляется число его вхождений. Модель связывает индекс с каждым словом в нашем словаре и представляет каждое предложение как нули, где 1 в каждом индексе соответствует слову, присутствующему в предложении."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JIymCQDbJNCt"
      },
      "source": [
        "def cv(data):\n",
        "    count_vectorizer = CountVectorizer()\n",
        "    emb = count_vectorizer.fit_transform(data)\n",
        "    return emb, count_vectorizer\n",
        "\n",
        "list_corpus = clean_data[\"text\"].tolist()\n",
        "list_labels = clean_data[\"label\"].tolist()\n",
        "\n",
        "X_train, X_test, y_train, y_test = train_test_split(list_corpus, list_labels, test_size=0.2, random_state=40)\n",
        "\n",
        "X_train_counts, count_vectorizer = cv(X_train)\n",
        "X_test_counts = count_vectorizer.transform(X_test)"
      ],
      "execution_count": 19,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 918
        },
        "id": "ndeI46g2lPxu",
        "outputId": "3addf848-8953-41d6-bb47-df3d6248fa15"
      },
      "source": [
        "def plot_LSA(test_data, test_labels, savepath=\"PCA.csv\", plot=True):\n",
        "        lsa = TruncatedSVD(n_components=2)\n",
        "        lsa.fit(test_data)\n",
        "        lsa_scores = lsa.transform(test_data)\n",
        "        color_mapper = {label:idx for idx,label in enumerate(set(test_labels))}\n",
        "        color_column = [color_mapper[label] for label in test_labels]\n",
        "        colors = ['orange', 'blue', 'green', 'red']\n",
        "        if plot:\n",
        "            plt.scatter(lsa_scores[:,0], lsa_scores[:,1], s=8, alpha=.8, c=test_labels, cmap=matplotlib.colors.ListedColormap(colors))\n",
        "            orange_patch = mpatches.Patch(color='orange', label='class 1')\n",
        "            blue_patch = mpatches.Patch(color='blue', label='class 2')\n",
        "            green_patch = mpatches.Patch(color='green', label='class 3')\n",
        "            red_patch = mpatches.Patch(color='red', label='class 4')\n",
        "            plt.legend(handles=[orange_patch, blue_patch, green_patch, red_patch], prop={'size': 30})\n",
        "\n",
        "\n",
        "fig = plt.figure(figsize=(16, 16))          \n",
        "plot_LSA(X_train_counts, y_train)\n",
        "plt.show()"
      ],
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1152x1152 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_SH0x6FbmFk9"
      },
      "source": [
        "Что же, выглядит немного невкусно и грустно, однако попробуем обучить логистическую регрессию. Зелёные так вообще потерялись :("
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "bHSZ4sZpmpD5"
      },
      "source": [
        "## Логистическая регрессия"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "18HLbz_tNYUX",
        "outputId": "0f716324-65ea-4fdd-c9e1-c8d22054c099"
      },
      "source": [
        "clf = LogisticRegression(C=30.0, class_weight='balanced', solver='newton-cg', multi_class='multinomial', n_jobs=-1, random_state=40)\n",
        "clf.fit(X_train_counts, y_train)\n",
        "y_predicted_counts = clf.predict(X_test_counts)\n",
        "\n",
        "def get_metrics(y_test, y_predicted):  \n",
        "    precision = precision_score(y_test, y_predicted, pos_label=None, average='weighted')             \n",
        "    recall = recall_score(y_test, y_predicted, pos_label=None, average='weighted')\n",
        "    f1 = f1_score(y_test, y_predicted, pos_label=None, average='weighted')\n",
        "    accuracy = accuracy_score(y_test, y_predicted)\n",
        "    return accuracy, precision, recall, f1\n",
        "\n",
        "accuracy, precision, recall, f1 = get_metrics(y_test, y_predicted_counts)\n",
        "print(\"accuracy = %.3f, precision = %.3f, recall = %.3f, f1 = %.3f\" % (accuracy, precision, recall, f1))"
      ],
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "accuracy = 0.567, precision = 0.599, recall = 0.567, f1 = 0.580\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QJd4co8Am01J"
      },
      "source": [
        "Метрики — это хорошо, но важно понимать, где ошибается классификатор, какие классы предсказывает хорошо, а какие — уже не очень. Поэтому надо построить **confusion matrix**:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9aEe8KpbnMeX"
      },
      "source": [
        "def get_confusion_matrix(cm, classes, normalize=False, title='Confusion matrix', cmap=plt.cm.winter):\n",
        "    if normalize:\n",
        "        cm = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]\n",
        "    plt.imshow(cm, interpolation='nearest', cmap=cmap)\n",
        "    plt.title(title, fontsize=30)\n",
        "    plt.colorbar()\n",
        "    tick_marks = np.arange(len(classes))\n",
        "    plt.xticks(tick_marks, classes, fontsize=20)\n",
        "    plt.yticks(tick_marks, classes, fontsize=20)\n",
        "    \n",
        "    fmt = '.2f' if normalize else 'd'\n",
        "    thresh = cm.max() / 2.\n",
        "\n",
        "    for i, j in itertools.product(range(cm.shape[0]), range(cm.shape[1])):\n",
        "        plt.text(j, i, format(cm[i, j], fmt), horizontalalignment=\"center\", \n",
        "                 color=\"white\" if cm[i, j] < thresh else \"black\", fontsize=40)\n",
        "    \n",
        "    plt.tight_layout()\n",
        "    plt.ylabel('True label', fontsize=30)\n",
        "    plt.xlabel('Predicted label', fontsize=30)\n",
        "\n",
        "    return plt"
      ],
      "execution_count": 23,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 798
        },
        "id": "hlyve__dnZV_",
        "outputId": "c77a8cab-fc16-45eb-8090-781fa63bbc56"
      },
      "source": [
        "cm = confusion_matrix(y_test, y_predicted_counts)\n",
        "fig = plt.figure(figsize=(10, 10))\n",
        "plot = get_confusion_matrix(cm, classes=['Class 1','Class 2','Class 3', 'Class 4'], normalize=False, title='Confusion matrix')\n",
        "plt.show()\n",
        "print(cm)"
      ],
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x720 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "[[4619  819  779 1621]\n",
            " [ 756 4160  320  872]\n",
            " [ 156   60  192  110]\n",
            " [1736  857  570 2372]]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "oW0NFApynuMq"
      },
      "source": [
        "Как и следовало ожидать, класс 3 предсказывается не очень хорошо: изначально его как раз было меньше. Пожалуй, в следующий раз надо будет воспользоваться аугментаций!"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zRUq1u9bntsM"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}