Community Detection Based on Variable Vertex Influence

In recent years community detection has been a hot research topic in network science, which helps to explain the characteristics of the network structure. This paper analyzed the effect of vertices with high influence in community detection, and found that such vertices have different roles in different networks. A variable influence community detection algorithm based on PageRank is proposed in this paper, which can resize the community involving vertices with high influence. A variable influence local community detection algorithm based on a single initial vertex is also proposed, which introduces the mechanism of influence decay and is able to find communities with different sizes according to users' requirements.