Affiliations: [a] College of Computer Science and Engineering, Shandong Province Key laboratory of Wisdom Mine Information Technology, Shandong University of Science and Technology, Qingdao, China
| [b] Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, Shenzhen, China
| [c] Guangxi Key Lab of Trusted Software, Guilin University of Electronic Technology, Guilin, China
| [d] Center for Computational Intelligence, Faculty of Technology, De Montfort University, Leicester, UK
Correspondence:
[*]
Corresponding author. Zhongying Zhao and Chao Li, College of Computer Science and Engineering, Shandong University of Science and Technology, Qingdao 266590, China. Tel.: +86 532 86057524; Fax: +86 532 86057758; E-mails: [email protected] (Z. Zhao) and [email protected] (C. Li).
Abstract: Community detection aims to discover cohesive groups in which people connect with each other closely in social networks. A variety of methods have been proposed to detect communities in social networks. However, there is still few work to make a comparative study on those methods. In this paper, we first introduce and compare several representative methods on community detection. Then we implement those methods with python and make a comparative analysis on different real world social networking data sets. The experimental results have shown that GN algorithm is suitable for small networks, while LPA algorithm has a better scalability. FU algorithm is of the best stability. This work could help researchers to understand the ideas of community detection methods better and select appropriate method on demand more easily.
Keywords: Social network mining, community detection, comparative analysis
