Fake news spreader detection using trust-based strategies in social networks with bot filtration

An important aspect of preventing fake news spreading in social networks is to proactively detect the users that are likely going to spread such news. Research in the domain of spreader detection is at a nascent stage compared to fake news detection. In this paper, we propose a graph neural network-based framework to identify nodes that are likely to become spreaders of false information. Using the community health assessment model and interpersonal trust (quantified using network topology and historical behavioral data), we propose an inductive representation learning framework to predict nodes of densely connected community structures that are most likely to spread fake news, thus making the entire community vulnerable to the infection. We also analyze the performance of our model in the presence and absence of bots detected using an existing state-of-the-art bot detection model. Using topology- and activity-based trust properties sampled and aggregated from neighborhood of nodes, we are able to predict false information spreaders better than refutation information spreaders. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1007/s13278-022-00890-z.