A Community-Gravity-Directed Algorithm for Showing Community Structure of Complex Networks

Since most of traditional force-directed algorithms cannot present the community structure of complex networks, this paper proposes a new graph layout method, called community-gravity-directed algorithm, to effectively visualize the community structure of complex networks. Based on the force-directed algorithm, the proposed algorithm introduces community-gravity force to each node, and employs k-means algorithm to make nodes in the same community near to the center of the community. Different from traditional approaches which have to cluster network before visualization, the proposed algorithm can perform the two steps at the same time, without any node pre-classification. In the experiment, modularity is used to evaluate the strength of community structures. Experimental results show that the proposed algorithm can present a significant clustering effect. It is simple, easy to implement, and has a better convergence rate.