News Category Classification

This project classifies news articles into four categories — World, Sports, Business, and Sci/Tech — using the AG News dataset. It features text preprocessing, TF-IDF vectorization, and classification using Logistic Regression and a feedforward Neural Network.

---

Features

• Text preprocessing: tokenization, stopword removal, lemmatization
• TF-IDF vectorization for feature extraction
• Classification using Logistic Regression and a feedforward Neural Network
• Visualizations: class distribution and word clouds per category
• Model evaluation with accuracy, classification report, and confusion matrix

Table of Contents

• Project Overview
• Dataset
• Installation
• Usage
• Models
• Results
• Visualizations
• License

---

Project Overview

The goal is to build a classifier to predict the category of a news article based on its title and description.

• Data cleaning and preprocessing include tokenization, stopword removal, and lemmatization.
• Feature extraction using TF-IDF vectorization (uni-grams and bi-grams).
• Classification using:
  • Logistic Regression
  • Feedforward Neural Network with dropout and L2 regularization

---

Dataset

The dataset consists of CSV files (train.csv and test.csv) with the following columns:

• class_index: Numeric class label (1 to 4)
• title: News article title
• description: News article description

Class mapping:

class_index	category_name
1	World
2	Sports
3	Business
4	Sci/Tech

---

Installation

1. Clone the repository:

git clone https://github.com/your-username/news-category-classification.git
cd news-category-classification

2.(Optional) Create and activate a virtual environment:

python -m venv venv
# On Linux/macOS
source venv/bin/activate
# On Windows
venv\Scripts\activate

3.Install the required packages directly with pip:

pip install pandas numpy matplotlib seaborn wordcloud nltk scikit-learn tensorflow

---

Usage

1. Update the dataset file paths inside the Python script:

Open the main script file (main.py or your script filename) and replace the following variables with the paths to your local dataset files:

train_path = r"YOUR_LOCAL_PATH_TO_train.csv"
test_path = r"YOUR_LOCAL_PATH_TO_test.csv"

---

Models

Logistic Regression

• Trained with TF-IDF features (up to 10,000 features, uni-grams and bi-grams).
• Maximum iterations set to 1000.
• Uses scikit-learn implementation with a fixed random seed for reproducibility.

Neural Network

• Input layer size matches TF-IDF feature size.
• Two hidden layers with 512 and 256 neurons respectively.
• Uses ReLU activation functions.
• Includes Dropout (0.5) and L2 regularization (0.01) to reduce overfitting.
• Output layer with softmax activation for multi-class classification.
• Optimizer: Adam.
• Loss: Sparse categorical crossentropy.
• Early stopping with patience of 3 epochs.

---

Results

The models produced the following outputs and metrics:

1. Class Distribution

2. Logistic Regression Accuracy

3. Logistic Regression Confusion Matrix

4. Neural Network Accuracy Over Epochs

5. Neural Network Training History

6. Successful Predictions

7. Word Cloud - Business

8. Word Cloud - Sci/Tech

9. Word Cloud - Sports

10. Word Cloud - World

---

Visualizations

• Bar plot for class distribution (class_distribution.png).
• Word clouds for each category (wordcloud_<category>.png).
• Confusion matrix heatmap for Logistic Regression (logreg_confusion_matrix.png).
• Neural network training accuracy and loss over epochs (nn_training_history.png).

All visualizations are saved automatically when you run the script.

---

License

This project is licensed under the MIT License - see the LICENSE file for details.