Temporal Stable Community in Time-Varying Networks

Abstract

Identifying community structure of a complex network provides insight to the interdependence between the network topology and emergent collective behaviors of networks, while detecting such invariant communities in a time-varying network is more challenging. In this paper, we define the temporal stable community and use the temporal similarity of edges for judging whether temporal stable communities exist in time-varying networks. Furthermore, we newly propose the concept of dynamic modularity to evaluate the stable community structures in time-varying networks, which is robust against small changes as verified by several empirical time-varying network datasets. Besides, using the volatility features of temporal stable communities in functional brain networks, we successfully differentiate the ADHD (Attention Deficit Hyperactivity Disorder) patients and healthy controls efficiently.