A New Framework based on Learning Automata for User Community Detection in Social Networks

Recently, social networks provide some rich resources of heterogeneous data which their analysis can lead to discover unknown information and relations within such networks. Users in online social networks tend to form community groups based on common location, interests, occupation, etc. Hence, communities play special roles in the structure–function relationship. Therefore, detecting significant and densely connected user communities from social networks has become one of the major challenges that help to understand some behavioral characteristics of users in social networks. Moreover, discovered communities can be a way to describe and analyze such networks. Most recent works on user community detection has focused on analyzing either user-friendship structure or user-generated contents but not both at the same time. In this paper, we propose a new framework based on distributed learning automata for detecting user community that considers user-friendship structure and user content information simultaneously. Finally we have evaluated our framework on the Twitter dataset. The evaluation results indicate that this framework is able to discover substantial user communities in which there are dense relationships among members.