Rumor Propagation Analysis Model Inspired by Gravity Theory for Online Social Networks

The influence of rumor propagation in online social networks (OSN) could result in great damage to social life, and it has been a hot topic to discover rumor propagation pattern. Traditional Epidemic-like rumor propagation models based on SIR, are generally coarse-grained for OSN but do not fully consider the features of OSN, such as personalization dimensions of users' behavior and information attributes. Inspired by gravity theory, this paper proposes a novel rumor propagation analysis model named gravity-inspired rumor propagation model (GRPModel), and tries to find a new pattern of rumor propagation from the perspectives both of users' properties and rumors' attributes. In GRPModel, user influence and rumor influence are modeled mathematically by user relations and information attributes, and fully consider their personalized features. We collect experimental real data from Sina Weibo, which is a famous OSN in China, and investigate features of users and real rumors. Experiments prove the effectiveness and efficiency.