Probabilistic Diffusion of Social Influence with Incentives

With explosive growth of social media, social computing becomes a new IT feature. A core functionality of social computing is social network analysis, which studies dynamics of social connectivity among people, including how people influence one another and how fast information diffuses in a social network and what factors stimulate influence diffusion. One of the models for information diffusion is the heat diffusion model. Although it is simple in capturing the basic principle of social influence, there are several limitations. First, the uniform heat diffusion is no longer hold in social networks. Second, high degree nodes are not always most influential in all contexts. We propose a probabilistic approach of social influence diffusion model with incentives. Our approach has three features. First we define an influence diffusion probability for each node instead of uniform probability. Second, we categorize nodes into two classes: active and inactive. Active nodes have chances to influence inactive nodes but not vice versa. Third, we utilize a system defined diffusion threshold to control how influence is propagated. We study how incentives can be utilized to boost the influence diffusion.