A privacy-preserving friend recommendation scheme in online social networks. (April 2018)

• Record Type: Journal Article
• Authors: Zhang, Shiwen; Li, Xiong; Liu, Haowen; Lin, Yaping; Sangaiah, Arun Kumar
• Is Part Of: Sustainable cities and society. Volume 38(2018)
• Page Start: 275
• Page End: 285
• Publication Date: 2018-04
• Subjects: Online social networks -- Privacy -- k-Degree anonymous -- Friend recommendation; Sustainable urban development -- Periodicals; Sustainable buildings -- Periodicals; Urban ecology (Sociology) -- Periodicals; 307.76
• DOI: 10.1016/j.scs.2017.12.031
• Languages: English
• ISSNs: 2210-6707

Abstract

Highlights: A novel scheme can be used to protect user's identity and relationship privacy. The privacy-preserving friend recommendation service for users is proposed in OSNs. The security analysis and performance evaluation of this scheme has been conducted. The results show that the proposed scheme is one of the best for applications. Abstract: Recently, online social networks (OSNs), which can offer many innovative services, are on the rise. As making friends is a basic way to create user's social relationship, friend recommendation is proposed to help users expand their social circles in OSNs. However, traditional friend recommendation process poses several crucial privacy breaches in OSNs, such as identity theft and relationship privacy leakage. Aimed to solve this problem, different from traditional friend recommendation schemes, based on the common interests by characterizing user's social behaviors/activities, we identify the threat model, and then propose a k -degree anonymous friend recommendation (KFR) scheme. Firstly, we abstract OSN as a hypergraph and then propose an edge segmentation algorithm to hide user's identity privacy and social relationship privacy. Subsequently, based on users' common interests, we design a similarity calculation algorithm. Finally, combined the similarity calculation algorithm with the segmentation tree (ST) technique, a novel k -degree anonymous friend recommendation scheme is proposed. The experiments carried out on real datasets show that the proposed scheme is scalable and effective.