Identifying influential individuals in microblogging networks using graph partitioning. (15th July 2018)

• Record Type: Journal Article
• Authors: Huang, Mingqing; Zou, Guobing; Zhang, Bofeng; Gan, Yanglan; Jiang, Susu; Jiang, Keyuan
• Is Part Of: Expert systems with applications. Volume 102(2018)
• Page Start: 70
• Page End: 82
• Publication Date: 2018-07-15
• Subjects: Influential individual -- Microblogging network -- Graph partitioning -- Latent Dirichlet Allocation model -- LeaderRank; Expert systems (Computer science) -- Periodicals; Systèmes experts (Informatique) -- Périodiques; Electronic journals; 006.33
• DOI: 10.1016/j.eswa.2018.02.021
• Languages: English
• ISSNs: 0957-4174

Abstract

Highlights: A directed and weighted network is constructed to depict influence dissemination. Sociability and interest are combined to calculate spreading probability. Influential users are selected in social media using modularity structure analysis. Modularity and size of the community are integrated to promote propagation. The proposed approach outperforms three methods in effectiveness and efficiency. Abstract: Identifying influential individuals who lead to faster and wider spreading of influence in social networks is of theoretical significance and practical value to either accelerating the speed of propagation in the case of product promotion, or hindering the pace of diffusion involved in rumors. Conventional methods, ranging from centrality indices to diffusion-based processes, already take into account the number and influences of followers, but fail to make full use of the characteristics of social media. A novel approach called PartitionRank for finding a pre-fixed number of influential individuals in microblogging scenarios is proposed in this study to maximize the impact; it combines interest similarity with social interaction between users via graph partitioning. Experimental results on artificial and real-world microblogging networks illustrate that our scheme outperforms the other state-of-the-art methods in effectiveness and efficiency.