Community Structure Detection Algorithm in Multi-relationships Complex Network Based on Clustering

Abstract

There are frequently many kinds of relationships in various real networks, which can be represented by multi-relationships complex network. It is a fundamental function for detecting the community structure for identifying the edges between structures in complex networks. Previous community structure detection algorithms often limited by the limitation of network topology structure arising from one relationship, but our algorithm based on a semi-supervised clustering algorithm and multi-subnet composited complex network can overcome the restriction. Through experiment analysis, our algorithm is compared to classical spectral clustering algorithm and nonnegative matrix factorization (NMF) by using artificial generated datasets. The Experimental results show that community structures which are divided by our proposed algorithm are more obscure.