Community Influence Analysis Based on Social Network Structures

Modeling and measuring social influence is a major problem in Social Network Analysis. Existing models and methods could handle individual influence analysis conveniently, but they rarely estimate the social influence of communities which are ubiquitous in social networks. Based on the structures of online social networks, a community oriented influence analysis model is proposed. Then, we provide an algorithm called CommRank for calculating the social influence of communities. Since the algorithm combines both internal structural information and external interaction data of communities, it estimates community influence more precisely on multiple datasets. Experimental results also demonstrate that, at the cost of a little gain loss, CommRank can dramatically improve the efficiency when dealing with the influence maximization problem.