Arabic news article classifier

Description

This project implements a sophisticated Arabic text classification system designed to automatically categorize Arabic news articles across multiple domains. Using advanced natural language processing techniques and machine learning algorithms, the system efficiently processes and classifies Arabic text while handling the unique challenges of Arabic language processing.

Built on Flask and powered by Scikit-learn's robust machine learning capabilities, the project combines TF-IDF vectorization with Stochastic Gradient Descent classification to achieve accurate text categorization. The system is trained on a comprehensive corpus from Hespress, a major Arabic news outlet, ensuring broad coverage across various news categories.

Key Features of the Project

• Arabic Text Processing: Implements specialized preprocessing techniques for Arabic text, including custom stop word removal, tokenization, and character reshaping.
• Machine Learning Pipeline: Utilizes Scikit-learn's TF-IDF vectorization and SGD classifier for efficient text categorization with online learning capabilities.
• Web Interface: Features a user-friendly Flask web application for easy text submission and classification results visualization.
• Modular Architecture: Employs a clean, modular design that separates preprocessing, classification, and web interface components for easy maintenance and updates.
• Multi-Category Support: Handles multiple news categories including politics, sports, society, culture, and economics, providing comprehensive classification capabilities.
• Production-Ready Design: Incorporates error handling, proper model validation, and efficient processing techniques suitable for production deployment.