A Hypergraph Clustering-based Technique for Detecting Fake News from Broadcasting Network

In today's digital world, people are aware of social media and consuming scoops or news directly from its news feed. It creates a strenuous situation as the authenticity of the news sources is unknown. Also, it is very difficult to judge whether a published or propagated report is genuine or not. In recent times, fake news or rumors are creating various social issues by deceiving people in several relevant contexts. Existing methods of fake news detection are relied heavily on manual leveling of data, fact-checking from relevant or authenticate sources, and supervised learning techniques. As the supervised learning approach is highly dependent on manual intervention, this manuscript presents an unsupervised approach for detecting any forgery news from online broadcasting network. We have studied inter-user behavior patterns in broadcasting networks based on hypergraph and labeled the news articles for fake news detection. The proposed method would terminate the fake news propagation through social media by eliminating them from broadcasting. The satisfactory inquisitive results of our approach on different real-world data set would inspire the interdisciplinary researchers to probe further in this area.