Social influence minimization based on context-aware multiple influences diffusion model

With the increasing popularity of online social networks, online information sharing turns out to be pervasive. A variety of innovations simultaneously propagates through online social networks, including both positive and negative information. However, the spread of any undesirable influence potentially breeds threat of rumors and misinformation, which may arouse extensive attention from society. For example, adverse information or rumors inevitably lead public relation crisis for corporates; misinformation exerts negative impact and public panic in the society. In this research, we systematically studied the undesirable influence minimization problem in the context of multiple influences. The strategies of introducing extra influences are theoretically analyzed. A novel agent-based influence–diffusion model is proposed for handling the diffusion of multiple influences. We also developed two context-aware seeding algorithms to minimize the adverse impact of an undesirable influence. Within the context of our investigation, the experimental results not only demonstrate the feasibility and advantages of the proposed approach but also reveal several intriguing discoveries.