Automatic discovering latent communities of users from observed textual content and their relationships is vital for understanding the cooperation and interaction patterns of users on large scale. In this paper, a novel probabilistic generative model was proposed to detect latent communities in a social network based on semantic information and the social relationships between users. In this model, it was assumed that users from the same community tend to share similar interests, and those who engage in common topics should be closely connected to each other on the topology structure of the social network. Users can have multiple interests and participate in multiple communities. Further, heterogeneous relationship strength was used in this paper to improve community discovery. The research indicated that the probabilistic generative model present in this paper has a good capability of discovering meaningful communities and topics on real-world data from Twitter. © 2014 Springer-Verlag Berlin Heidelberg.
CITATION STYLE
Bi, J., Huang, J., & Qin, Z. (2014). A relationship strength-aware topic model for communities discovery in online social networks. In Lecture Notes in Electrical Engineering (Vol. 279 LNEE, pp. 709–715). Springer Verlag. https://doi.org/10.1007/978-3-642-41674-3_101
Mendeley helps you to discover research relevant for your work.