A Survey of Big Data Approaches for Fake News Detection on Social Media

The widespread dissemination of fake news on social media has emerged as a critical challenge with profound societal consequences. As traditional content-based approaches often fall short in detecting deceptive or biased information, researchers have increasingly turned to big data methods that incorporate linguistic features, user interactions, and network dynamics. This survey reviews and compares three representative studies that utilize big data for fake news and media bias detection. The first study constructs a multidimensional dataset through crowdsourced annotations of bias dimensions such as subjectivity and framing. The second employs geometric deep learning on social graphs to identify fake news based on propagation patterns. The third provides a comprehensive overview of data mining strategies, emphasizing the integration of content and social signals. We analyze the methodologies, datasets, and results presented in these works and highlight their contributions to the evolving field of misinformation detection. This article aims to provide a structured understanding of current big data approaches and identifies open challenges and future research directions in building robust, scalable, and interpretable fake news detection systems.