Tracing the fake news propagation path using social network analysis

Nowadays, people rely mostly on social media for any kind of information sharing and also started acquiring information through social media platforms for e-news mostly related to politics via Twitter, Facebook, and YouTube. Fake news detection and identifying its propagation path are technically very challenging. In this work, we have implemented a novel method to learn discriminative features from tweets content, Facebook posts and followed their non-sequential propagation structure, and generated more powerful representations for identifying fake news and its propagation by constructing a social network graph. We proposed level order traversal up to three levels based on top-down tree structured networks for fake propagation learning and detected the neighbors of the fake news source and removed them from the network which naturally confirms the reduction of their propagation. We have considered the benchmark data set LIAR and used PolitiFact user data for our research work. The main objective of our work is to identify the propagation path of the fake news content by collecting news and verifying its authenticity using fact-checking websites, namely “www.politifact.com”, and creating a network among the users who have high similarity in their contents posted. Now it will be easier to trace the path if the source identified has fake content, then its neighbors can be tracked and moving forward the same idea can be iterated up to bottom levels.