A New Algorithm and Its Application in Detecting Community of the Bipartite Complex Network

The community division of bipartite networks is one frontier problem on the research of complex networks today. In this study, we propose a model of community detection of the bipartite network, which is based on the generalized suffix tree algorithm. First, extract the adjacent node sequences from the matrix of relation and use the obtained adjacent node sequences to build a generalized suffix tree; second, traverse the established generalized suffix tree to obtain the bipartite cliques; third, adjust the bipartite cliques; finally, dispose the isolated edges, get the communities, and complete the division of the bipartite network. This algorithm is different from the traditional community mining one since it uses edges as the community division medium and does not need to specify the number of the division of communities before the experiment. Furthermore, we can find overlapping communities by this new algorithm which can decrease the time complexity.

Automated summary

The study introduces a model of community detection for bipartite networks based on the generalized suffix tree algorithm, which involves steps like extracting adjacent node sequences, building the tree, obtaining bipartite cliques, adjusting them, processing isolated edges, and deriving communities to divide the network. This algorithm stands out from traditional methods by utilizing edges as the medium for community division and not requiring a pre-specification of the number of communities, as well as being able to identify overlapping communities and improve time complexity.