A comprehensive review of machine learning-based models for fake news detection

Abstract

The internet's role in the spread of fake news has made it a serious problem that calls for sophisticated techniques for its automatic detection. To counter the quick spread of misleading material on social media sites, including Facebook, Twitter, Instagram, and WhatsApp, this study explores the domain of deep learning methods and various classification strategies. In particular, the study investigates the application of transformer-based models like BERT, recurrent neural networks (RNNs), and convolutional neural networks (CNNs) for the detection of false news. To provide context for the results, the examination covers a wide range of historical and contemporary trends in the spread of fake news. The study uses strict criteria, including precision, recall, and F1 score, to assess how effective the suggested models are. The created methodologies are validated for effectiveness by using benchmark datasets. This study emphasizes how urgent it is to address the growing problem of fake news, especially as it relates to its use in psychological warfare and clickbait to generate revenue. Through the clarification of the methods, duration, and standards of evaluation used, this research adds to the current conversation on dispelling false information in the digital era.