A novel trust-based community detection algorithm used in social networks

Mining the community structure is an important subject in the area of social network analysis, and detecting the hidden communities within the social networks will help to better understand the topological properties of the real-life networks. Meanwhile, community detection will be also helpful to monitor the public opinion, identify the opinion leaders and perform the personalized recommendation. In comparison with the simplex user ties or contents, considering the trust features from multiple users will provide a more comprehensive account of the linking relationship between users. To this end, we propose a novel non-overlapping community detection algorithm, which is based on the trust mechanism, to recognize the community structure in this paper. At first, we propose several definitions with regard to trust relationship between users to depict the trust strength, which includes the direct, indirect and mutual trust, and then the specific trust calculation method is provided to quantitatively describe the extent of trust. Secondly, starting from the trust relationship, we integrate the edge fitness and community fitness into the non-overlapping community detection and propose a novel trust-based algorithm to comprehensively leverage the trust among nodes to further mine the communities within the networks. Finally, to deeply analyze the analyze the performance, we take use of Lesmis and Gemo data sets to carry out extensive experiments, and the results show that, compared with other classical algorithms, the community based on the newly proposed algorithm features the higher trust cohesion on the condition that the structural cohesiveness of social network is fully satisfied. The current methods will be of significance to deeply understand and effectively find out the communities within realistic networks. (C) 2018 Elsevier Ltd. All rights reserved.