Community Expansion Model Based on Charged System Theory

Abstract

Recently, the phenomenon of influence propagation becomes a hot topic in social networks. However, few existing influence models study the influence from communities, which has a large range of applications. In this paper, we use the charged system model to represent the social influence. This model provides a natural description about the behaviors of influence and explains why the influence makes communities expand. Based on this physical model, we propose two objective functions for choosing proper candidates to enlarge a community, considering of the cost and benefit issue. Then a linear programming approach is given to maximize those two objective functions. We validate our ideas and algorithm using two real-world networks. The results demonstrate that our model can choose excellent propagation candidates for a specific community, comparing to other two algorithms.