FAKE NEWS DETECTION USING AI

Abstract

The proliferation of online misinformation presents a significant threat to public discourse and democratic stability. Detecting fake news automatically using Natural Language Processing (NLP) and Machine Learning (ML) techniques has emerged as a vital research domain. This paper presents an NLP-based Fake News Classification system that employs a Progressive Aggressive Classifier (PAC), a variant of online learning models optimized for high-dimensional textual data. The proposed model uses word embeddings, tokenization, and syntactic features to classify news content as fake or real. Experimental evaluations on benchmark datasets such as LIAR and FakeNewsNet demonstrate that the PAC model achieves competitive performance compared to Support Vector Machines (SVM), Logistic Regression, and Neural Networks, while offering superior computational efficiency.