Research On Methods Of Finding Community In Edge Heterogeneous Networks

Complex networks are widely used in system modeling,and heterogeneous networks can accurately describe complex relationships in real world.As one of the characteristics of complex networks,community structure has important significance for analyzing the structure and function of the network.Traditional community detection methods focus on finding community structure based on the topological information of network,ignoring the heterogeneous information of networks,which makes the difference between topological communities and functional communities.In this paper,the problem of finding community structure in edge heterogeneous networks is studied and the community detection method based on typed-edge similarity clustering(TESC)is proposed.This method is based on the local topological characteristics and local heterogeneous information,and merges the vertexes to get the hierarchical community structure of edge heterogeneous network by the idea of hierarchical clustering.Based on TESC method,the clustering objects are extended to dense node sets,and the community detection method based on node-density(NDC)is proposed.This method considers both the global topology information and heterogeneous information of the network,and takes the dense node sets for clustering object,so as to obtain the community structure of heterogeneous network edge and improve the efficiency of finding conununity structure.In the end,this paper validates the TESC and NDC methods in the use of experiments on real networks and synthetic networks,and analyzes the community structures through the indicators of normalized mutual information,homogeneity and others.Experimental results demonstrate that the community structures obtained by the TESC and NDC methods are closer to the real results,and the function of community structures are more consistent its topology.