Understanding Community Effects on Information Diffusion

In social network research, community study is one flourishing aspect which leads to insightful solutions to many practical challenges. Despite the ubiquitous existence of communities in social networks and their properties of depicting users and links, they have not been explicitly considered in information diffusion models. Previous studies on social networks discovered that links between communities function differently from those within communities. However, no information diffusion model has yet considered how the community structure affects the diffusion process. Motivated by this important absence, we conduct exploratory studies on the effects of communities in information diffusion processes. Our observations on community effects can help to solve many tasks in the studies of information diffusion. As an example, we show its application in solving one of the most important problems about information diffusion: the influence maximization problem. We propose a community-based fast influence (CFI) model which leverages the community effects on the diffusion of information and provides an effective approximate algorithm for the influence maximization problem.