AI / Machine Learning Developer · 2026
News Credibility Classifier
An NLP-based news credibility classifier that compares predictions from four machine learning models.
Engineered an NLP classification pipeline on the ISOT benchmark dataset of more than 44,000 articles. Applied TF-IDF vectorization and regex preprocessing while removing metadata to prevent data leakage. Trained and benchmarked Logistic Regression, Decision Tree, Random Forest, and Gradient Boosting models, achieving up to 98.6% evaluation accuracy. Built a real-time manual testing pipeline that transforms unseen raw news text and compares credibility predictions across all trained classifiers.