Using Suffix Tree to Detect Communities from Bipartite Network

This paper presents an algorithm for detecting communities in bipartite network based on the suffix tree structure. The algorithm, first extracts the adjacent node sequence for reach node in the network. Based on the node sequences of all the nodes, the algorithm constructs a suffix tree, where each node represents a complete bipartite sub-graph in the network G. Then the algorithm adjusts those cliques to form the initial communities. Then these initial communities are extended and adjusted to get the final community partitioning. Our proposed algorithm can detect the overlapping communities, and be able to get one-many relationship which can make the relationships in the bipartite more clearly. In order to verify the correctness and effectiveness of our algorithm, we test our algorithm on the real datasets. Experimental results demonstrate that our algorithm can extract communities from bipartite networks and obtain high quality of partitioning communities.