Virtual Community Detection Model and Related Research in Online Social Networks

In the field of virtual community detection in online social networks, most of existing methods often detect communities from a single perspective, and ignore the influence of related characteristics of the networks on community detection. All these reduce the interpretability and accuracy of the community partition results. In order to resolve this problem, a virtual community detection model framework for online social networks is proposed. The model framework considers three key factors that affect the community detection results: the structural characteristics, the attribute information and the nodes' influence levels of the network. The proposed model is not only a mapping of existing community detection models, but also a reference for designing more future models for community detection methods.