Dynamic behavioral mixedmembership model for large evolving networks

Abstract. The majority of real-world networks are dynamic and ex-tremely large (e.g., Internet Traffic, Twitter, Facebook,...). To under-stand the structural behavior of nodes in these large dynamic networks, it may be necessary to model the dynamics of behavioral roles represent-ing the main connectivity patterns over time. In this paper, we propose a dynamic behavioral mixed-membership model (DBMM) that captures the “roles ” of nodes in the graph and how they evolve over time. Unlike other node-centric models, our model is scalable for analyzing large dy-namic networks. In addition, DBMM is flexible, parameter-free, has no functional form or parameterization, and is interpretable (identifies ex-plainable patterns). The performance results indicate our approach can be applied to very large networks while the experimental results show that our model uncovers interesting patterns underlying the dynamics of these networks.