Community Detection based on Structural and Attribute Similarities

The study of social networks has gained much interest from the research community in recent years. One important challenge is to search for communities in social networks. A community is defined as a group of users such that they interact with each other more frequently than with those outside the group. Being able to identify the community structure can facilitate many tasks such as recommendation of friends, network analysis and visualization. In real-world networks, in addition to topological structure (i.e., links), content information is also available. Existing community detection methods are usually based on the structural features and do not take into account the attributes of nodes. In this paper, we propose two algorithms that use both structural and attribute information to extract communities. Our methods partition a graph with attributes into communities so that the nodes in the same community are densely connected as well as homogeneous. Experimental results demonstrate that our methods provide more meaningful communities than conven- tional methods that consider only relationship information. Keywords-social network; community detection; clustering;