Overlapping Community Detection Algorithms: A Comparative Study

In complex networks a node may belong to many communities resulting in a highly overlapping community structure. This provides multiple information about such nodes by analyzing the communities they belong to. Recent advances in benchmarking haslead to the fact that most of the popular CAA (Community Assignment Algorithms) works only when the extent of overlap in a network is modest. GCE happens to be one such CAA with the ability to report communities in a network graph with huge accuracy. In this paper, we have discussed several existing state of art methods for detecting of overlapping communities with their approaches and disadvantages. Also we presented experimental evidence of how the extension of GCE algorithm (EGCE) outperforms the other existing overlapping community detection algorithm. At the same time we have analyzed the performances of the other existing algorithms.