LAUSR

Abstract A Community detection in signed networks is a partition on nodes such that the intra-community edges are positive and the inter-community edges are negative. The communities detection had been solved by Harary and Davis when a signed graph is balanced or weak balanced. While communities detection become much complex when a signed network is imbalanced. In this paper, a novel evolution algorithm is presented on community detection in imbalanced signed networks which can be modeled as an optimal partition problem. And the evolving mechanism of nodes is updated by its neighbors’ information which leads to form optimal community structure. The effectiveness of the algorithm is proved by experiments both on real-world and synthetic networks. The comparison with other algorithms by some parameters showed that the evolution algorithm is effective and accurate.