Overlapping Community Identification Algorithm in Directed Network

In the actual complex networks, nodes belong to multiple communities and the connection relationship between nodes is often directional or asymmetrical. New algorithm based overlapping community identification in directed network was proposed. Firstly, the research converted a directed path in a directed graph to a weighted undirected graph with conversion strategy; Then, according to the distributional and aggregated characteristics of nodes in community, the core node was sought; Furthermore, the algorithm calculated the memberships of newly added nodes and the existing small community formed by the core node and neighbor nodes, and achieved a gradual expansion of the local community to the global community; Finally, the overlapping degree and modularity were used to divide the community effectively. Experimental results showed: Community identification algorithm in this research realized a conversion from directed graph to weighted undirected graph, and extended the entire overlapping network community from local communities. The method improved the effect of aggregation of community nodes, and can effectively identify the complex network community.