Community detection in networks using new update rules for label propagation

Detecting community structure clarifies the link between structure and function in complex networks and is used for applications in many disciplines. The Label Propagation Algorithm (LPA) has the benefits of nearly-linear running time and easy implementation, but it returns multiple resulting partitions over multiple runs. Following LPA, some new updating rules are proposed to detect communities in networks, which are based mainly on the almost strong definition of communities and the topological similarity. Experiments on more artificial and real social networks have demonstrated better performance of the proposed method compared with that of the community detection algorithms CNM, Cfinder and MEP on the quality of communities.