Concept Lattice Factorization in Social Networks

Fei Hao, Chao Zhang, Bo Li, Jianmei Cheng

Abstract: Cohesive subgraph is regarded as an important topological structure in social network analysis. It is critical to understand the organization behavior of users and supervise the users work well together in order to achieve goals within a social network. Therefore, cohesive subgraph mining is becoming a critical research issue for social network analysis. However, there exist many challenges for mining cohesive subgraphs from massive social networks due to its large scale property. Aiming to reduce the size of social networks for fast analytic, this paper formulates the problem on concept lattice factorization in a social network and further devises a factorization algorithm. Importantly, an efficient community detection approach based on concept lattice factors is then proposed considering the concept lattice factors can well characterize the topology skeleton of a given social network.