word_embedding_classification.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "word_embedding_classification.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "authorship_tag": "ABX9TyOig2EHUzd7G7EmqpnPHukI",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/shraddha-an/nlp/blob/main/word_embedding_classification.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "P568CBV8nPxM"
      },
      "source": [
        "# **Case Study: NLP models for Text Classification**"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BmwRePl_nOB4"
      },
      "source": [
        "## **Introduction**\n",
        "\n",
        "In this project, I look at 2 different NLP models for classifying questions on Stack Overflow into 3 categories depending on their quality.\n",
        "\n",
        "This Case Study outlines 2 techniques to achieve the task of text classification:\n",
        "\n",
        "1.   Training Word Embedding\n",
        "2.   [Pretrained GloVe Word Embeddings](https://github.com/shraddha-an/nlp/blob/main/pretrained_glove_classification.ipynb)\n",
        "3.   [fastText Classifier](https://github.com/shraddha-an/nlp/blob/main/so_fasttext.ipynb)\n",
        "\n",
        "This Colab Notebook focusses on the first task, Training a Word Embedding on the chosen dataset.\n",
        "\n",
        "\n",
        "**Dataset**: [Stack Overflow Questions](https://www.kaggle.com/imoore/60k-stack-overflow-questions-with-quality-rate)\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-JJfG7f2ok57"
      },
      "source": [
        "## 1) **Data Preparation**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9TXLaHOnmdBY"
      },
      "source": [
        "# Importing libraries\n",
        "# Data Manipulation/ Handling\n",
        "import pandas as pd, numpy as np\n",
        "\n",
        "# Visualization\n",
        "import seaborn as sb, matplotlib.pyplot as plt\n",
        "\n",
        "# NLP libraries\n",
        "import re\n",
        "from nltk.corpus import stopwords\n",
        "from gensim.utils import simple_preprocess\n",
        "\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "P7X-jWkho0IT"
      },
      "source": [
        "# Importing training & testing datasets\n",
        "dataset = pd.read_csv('train.csv')[['Body', 'Y']].rename(columns = {'Body': 'question', 'Y': 'category'})\n",
        "ds = pd.read_csv('valid.csv')[['Body', 'Y']].rename(columns = {'Body': 'question', 'Y': 'category'})"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "YZPixtSjpU6W"
      },
      "source": [
        "## **2) NLP Preprocessing**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HYw0IZIOo70V"
      },
      "source": [
        "# Removing symbols, stopwords, punctuation\n",
        "stop_words = set(stopwords.words('english'))\n",
        "\n",
        "symbols = re.compile(pattern = '[/<>(){}\\[\\]\\|@,;]')\n",
        "tags = ['href', 'http', 'https', 'www']\n",
        "\n",
        "def text_clean(s: str) -> str:\n",
        "    \"\"\"\n",
        "    Removes unwanted symbols, punctuation and stop words from a given string.\n",
        "    \"\"\"\n",
        "    s = symbols.sub(' ', s)\n",
        "    for i in tags:\n",
        "        s = s.replace(i, ' ')\n",
        "    cleaned_text = ' '.join(word for word in simple_preprocess(s, deacc = True) if not word in stop_words)\n",
        "    return cleaned_text\n",
        "\n",
        "# Applying the function on the questions column\n",
        "dataset.iloc[:, 0] = dataset.iloc[:, 0].apply(text_clean)\n",
        "ds.iloc[:, 0] = ds.iloc[:, 0].apply(text_clean)\n",
        "\n",
        "# Train & Test subsets\n",
        "X_train, y_train = dataset.iloc[:, 0].values, dataset.iloc[:, 1].values.reshape(-1, 1)\n",
        "X_test, y_test = ds.iloc[:, 0].values, ds.iloc[:, 1].values.reshape(-1, 1)\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "aePnEQ11tokf"
      },
      "source": [
        "## **3) Categorical Encoding**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kfqV81XPqUJt"
      },
      "source": [
        "# One Hot Encoding the Categories Column\n",
        "from sklearn.preprocessing import OneHotEncoder as ohe\n",
        "from sklearn.compose import ColumnTransformer\n",
        "\n",
        "ct = ColumnTransformer(transformers = [('one_hot_encoder', ohe(categories = 'auto'), [0])],\n",
        "                       remainder = 'passthrough')\n",
        "\n",
        "y_train = ct.fit_transform(y_train)\n",
        "y_test = ct.transform(y_test)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "JWo1SnLXtzv3"
      },
      "source": [
        "## **4) Tokenization**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "c_BOIy-qtuP_"
      },
      "source": [
        "# Vectorizing our text corpus of questions\n",
        "# Setting some paramters\n",
        "vocab_size = 2100\n",
        "sequence_length = 300\n",
        "\n",
        "# Tokenization with keras\n",
        "from keras.preprocessing.text import Tokenizer\n",
        "\n",
        "tk = Tokenizer(num_words = vocab_size)\n",
        "tk.fit_on_texts(X_train)\n",
        "\n",
        "X_train = tk.texts_to_sequences(X_train)\n",
        "X_test = tk.texts_to_sequences(X_test)\n",
        "\n",
        "# Padding all questions with zeros\n",
        "from keras.preprocessing.sequence import pad_sequences\n",
        "\n",
        "X_train_seq = pad_sequences(X_train, maxlen = sequence_length, padding = 'post')\n",
        "X_test_seq = pad_sequences(X_test, maxlen = sequence_length, padding = 'post')\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "akQ5PmeYubxT"
      },
      "source": [
        "## **5) Embedding Model**\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WOETV7ivujNm"
      },
      "source": [
        "# Building & Training the model\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Embedding, Dense, Flatten\n",
        "\n",
        "model = Sequential()\n",
        "model.add(Embedding(input_dim = vocab_size, output_dim = 5 , input_length = sequence_length))\n",
        "model.add(Flatten())\n",
        "\n",
        "model.add(Dense(units = 3, activation = 'softmax'))\n",
        "\n",
        "model.compile(loss = 'categorical_crossentropy',\n",
        "              optimizer = 'rmsprop',\n",
        "              metrics = ['accuracy'])\n",
        "\n",
        "model.summary()\n",
        "\n",
        "# Compiling the model\n",
        "model.compile(optimizer = 'adam', metrics = ['accuracy'], loss = 'categorical_crossentropy')\n",
        "\n",
        "history = model.fit(X_train_seq, y_train, epochs = 20, batch_size = 512, verbose = 1)\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Rwj1yehSurIk"
      },
      "source": [
        "## **6) Evaluating Performance**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "b9yMSuLbuqdH"
      },
      "source": [
        "# Evaluating model performance on test set\n",
        "loss, accuracy = model.evaluate(X_test_seq, y_test, verbose = 1)\n",
        "print(\"\\nAccuracy: {}\\nLoss: {}\".format(accuracy, loss))\n",
        "\n",
        "# Save the model\n",
        "#model.save('mdl_80.h5')\n",
        "\n",
        "# Loading a saved model\n",
        "#from keras.models import load_model\n",
        "#model = load_model('mdl_80.h5')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7nPDD_MTu091"
      },
      "source": [
        "# Plotting Accuracy & Loss over epochs\n",
        "sb.set_style('darkgrid')\n",
        "\n",
        "# 1) Accuracy\n",
        "plt.figure(figsize = (8, 8))\n",
        "plt.plot(history.history['accuracy'], label = 'training', color = '#003399')\n",
        "plt.legend(shadow = True, loc = 'lower right')\n",
        "plt.title('Accuracy Plot over Epochs')\n",
        "plt.show()\n",
        "\n",
        "# 2) Loss\n",
        "plt.figure(figsize = (8, 8))\n",
        "plt.plot(history.history['loss'], label = 'training loss', color = '#FF0033')\n",
        "plt.legend(shadow = True, loc = 'upper right')\n",
        "plt.title('Loss Plot over Epochs')\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xhnaX4HtOjve"
      },
      "source": [
        "Accuracy & Loss Metrics:\n",
        " ![Screen Shot 2020-10-24 at 20.17.57.png](data:image/png;base64,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)\n",
        "\n",
        " Loss over Training Epochs: ![loss.png](data:image/png;base64,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)\n",
        " \n",
        " Accuracy over Training Epochs:![accuracy.png](data:image/png;base64,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)"
      ]
    }
  ]
}