A vector partitioning approach to detecting community structure in complex networks

In recent years, the problem of community structure detection has attracted more and more attention and many approaches have been proposed. Recently, Newman pointed out that this issue can be transformed into the problem of constrained maximization of the assignment matrix over possible divisions of a network. He presents further that this maximization process can be written in terms of the eigenspectrum of the “modularity matrix”. On the basis of this work and the vector partition approach in computer science, we propose a kind of multiway division approach for detecting community structure of complex networks. Experimental results indicate that the algorithm works well and is effective at finding both good communities and the appropriate number of communities.