International Research Journal of Education and Technology

The rapid growth of social media has transformed how people access news. Although digital platforms enable fast information sharing, they also facilitate the widespread dissemination of fake news. Misinformation can distort public opinion, manipulate political processes, and erode trust in credible media. This paper proposes a binary classification framework for detecting fake and real news articles using Natural Language Processing (NLP) and Machine Learning (ML) techniques. The system integrates static text classification, dynamic keyword-based verification, and website credibility analysis. Multiple supervised learning algorithms— including Naïve Bayes, Random Forest, Logistic Regression, and Passive Aggressive Classifier—were evaluated using benchmark datasets. Experimental results indicate that Logistic Regression achieved the best performance in static classification after hyperparameter tuning, while the Passive Aggressive classifier performed effectively for dynamic verification. The proposed system demonstrates that ML-based approaches provide a practical solution to mitigate the impact of online misinformation.

Keywords: Fake News Detection, Machine Learning, NLP, Text Classification, TF-IDF, Logistic Regression, Website Credibility