SCARLET: Explainable Attention Based Graph Neural Network for Fake News Spreader Prediction

False information and true fact checking it, often co-exist in social networks, each competing to influence people their spread paths. An efficient strategy here contain false is proactively identify if nodes the path are likely endorse (i.e. further it) or refutation (thereby help spreading). In this paper, we propose SCARLET (truSt andCredibility bAsed gRaph neuraLnEtwork model using aTtention) predict action of path. We aggregate trust credibility features from a node’s neighborhood historical behavioral data network structure explain how spreader’s vary. Using real world Twitter datasets, show that able spreaders with an accuracy over 87%.