A clique-superposition model for social networks

How does a social network evolve? Sociologists have studied this question for many years. According to some famous sociologists, social links are driven by social intersections. Actors who affiliate with the shared intersections tend to become interpersonally linked and form a cluster. In the social network, an actor cluster could be a clique or a group of several smaller-sized cliques. Thus we can conclude that a social network is composed of superposed cliques of different sizes. However, sociologists did not verify the theory in large scale data due to lack of computing ability. Motivated by this challenge, incorporated with the theory, we utilize data mining technologies to study the evolution patterns of large scale social networks in real world. Then, we propose a novel Clique-superposition generative model, which generates undirected weighted networks. By extensive experiments, we demonstrate that our model can generate networks with static and time evolving patterns observed not only in earlier literature but also in our work.