Fake News Detection Using Temporal Snapshots in Graph Neural Networks

Authors:

Abstract
The widespread propagation of misinformation in social media has been a significant barrier to information credibility and trust among the general public. It is crucial to address this problem in order to stop misinformation from twisting public conversations. Despite many deep learning methods having been employed for the detection of fake news, one task remains challenging: capturing complex and ever-evolving interactions present in social media. In this paper, we use dynamic Graph Neural Networks for fake news detection on Twitter. We study news articles’ content and the evolving patterns of user engagement to provide a more accurate and context-aware classification of fake news, and its early detection, and stop misinformation from spreading. Our results illustrate that the combination of temporal dynamics with network structure creates a more complete approach to improving information reliability on social media and combating misinformation.

This publication has 13 references indexed in Scilit: