A Stochastic Game Model for Analysis of Rumor and Anti-rumor Propagation in Social Networks*

摘要

In this paper, we first propose an Ignorant Spreader Vacciner(Immune) Fence-sitter (ISVF) diffusion model, which focuses on the symbiosis and confrontation of rumor and antirumor in social networks. Based on it, a stochastic game model is proposed to overcome the incomplete knowledge and the uncertainty in user’s choice of behavior. We design a utility function that matches the characteristics of user diversity to interpret the confrontation between rumors and anti-rumors. In order to learn neighbor strategies effectively, a Deep QNetwork (DQN) based approach is proposed. Simulation results based on the real datasets reveal that this model significantly accelerates positive information propagation.