TF-IDF and Machine Learning Based Approach for Fake News Classification

Authors: Gobika Sree B

Abstract: Fake news has become a major issue in today’s digital world, where information spreads rapidly through social media platforms. Identifying and filtering such misleading content is important to maintain public trust and prevent misinformation. This paper proposes a Machine Learning-based approach for fake news classification using TF-IDF as a feature extraction technique. TF-IDF helps convert text into numerical values by highlighting important words within a document. Various Machine Learning models such as Logistic Regression, Support Vector Machine (SVM), and Naïve Bayes were trained and evaluated. Among them, Logistic Regression achieved the highest accuracy and consistency in classification. Experimental results show that traditional ML models, combined with proper preprocessing, can effectively detect misleading content. The proposed approach provides a strong baseline for automated fake news detection.

DOI: https://doi.org/10.5281/zenodo.17667878