An Overlapping Community Detection Algorithm Based on Multistep Greedy Strategy

Considering the problem of overlapping community detection in directed and weighted complex networks, we proposed an overlapping community detection algorithm named MSG-OCD based on the multistep greedy strategy. The MSG-OCD integrates the weighting mechanism into the traditional modularity function, uses the improved multistep greedy strategy to detect the initial community structure, and analyzes the overlapping community structure in accordance with the overlapping community strategy. The MSG-OCD algorithm not only has a lower time complexity but also can prevent premature condensation into few larger communities. We verify the validity of the MSG-OCD algorithm through benchmark named Enron email set, which is a directed and weighted network.