Antony-6487/Swahili_News_Classification

This repository contains a machine learning pipeline for multi-class classification of Swahili news articles. The goal is to categorize news content into five topics: kitaifa, michezo, biashara, kimataifa, and burudani. # Swahili_News_Classification This repository contains a machine learning pipeline for multi-class classification of Swahili news articles. The goal is to categorize news content into five topics: kitaifa, michezo, biashara, kimataifa, and burudani. The project demonstrates an end-to-end approach including: Data cleaning and preprocessing for Swahili text TF-IDF vectorization (word and bigram features) Multi-class Logistic Regression classifier Stratified K-Fold cross-validation for reliable evaluation Out-of-fold predictions and test submission generation Objective: The model is designed for the Zindi Swahili News Classification Competition and emphasizes both accuracy and responsible AI practices, while supporting African language representation in technology. Files included: swahili_news_classification.ipynb – Jupyter/Colab notebook with full code submission.csv README.md – Project documentation Technologies used: Python, scikit-learn, pandas, NumPy, Google Colab