A privacy-preserving framework for cross-domain recommender systems. (July 2021)

• Record Type: Journal Article
• Authors: Ogunseyi, Taiwo Blessing; Bo, Tang; Yang, Cheng
• Is Part Of: Computers & electrical engineering. Volume 93(2021)
• Publication Date: 2021-07
• Subjects: Privacy-preserving -- Cross-domain -- Recommender systems -- Cryptography -- Homomorphic encryption; Computer engineering -- Periodicals; Electrical engineering -- Periodicals; Electrical engineering -- Data processing -- Periodicals; Ordinateurs -- Conception et construction -- Périodiques; Électrotechnique -- Périodiques; Électrotechnique -- Informatique -- Périodiques; Computer engineering; Electrical engineering; Electrical engineering -- Data processing; Periodicals; Electronic journals; 621.302854
• DOI: 10.1016/j.compeleceng.2021.107213
• Languages: English
• ISSNs: 0045-7906

Abstract

Highlights: A generic framework for privacy-preserving cross-domain recommender systems that satisfies some privacy requirements: user privacy, secure knowledge transfer, and secure recommendations. The framework consists of two different protocols for privacy-preserving cross-domain recommender systems. The protocols use homomorphic encryption and have the ability to protect users' privacy both at the data site level and user-level. The steps involved and the security of the proposed protocols were mathematically described in an honest but curious model. The computational and communication complexity of the two protocols were also analyzed. Abstract: User privacy in the recommender systems have received much attention over the years. However, much of this attention has been on privacy protection in single-domain recommender systems and not on cross-domain recommender systems. The privacy-preserving cross-domain recommender systems not only encourages collaboration of data between different domains to solve the problem of data sparsity but also ensures the users' privacy and secure transfer of auxiliary information between domains. However, existing studies are not suitable for privacy protection in a cross-domain scenario. To this end, we propose a novel privacy-preserving framework for cross-domain recommender systems that provides a generic template for other secure cross-domain recommender systems. Employing a homomorphic encryption scheme, the framework consists of two protocols for users' privacy in cross-domain recommender systems. We mathematically described every step involved in each protocol, proved that the two protocols are secure against a semi-honest adversary, and compared the complexity of the protocols. Graphical abstract: The architecture of the proposed privacy-preserving cross-domain model Image, graphical abstract