A Distributed Anonymization Scheme for Privacy-preserving Recommendation Systems

Recommendation systems need to collect user personal data for predicting user preferences. The privacy of user personal data becomes a main concern when users are served by the recommendation system. In this paper, we study the problem of privacy-preserving recommendation in the case that the user data are locally stored in a distributed manner. We present a distributed anonymization scheme based on the proposed anonymization map. The proposed scheme allows users individually anonymize their own data without accessing each other's data. The experiment results show that the proposed scheme can preserve the privacy of collaborative users and outperform the perturbation-based scheme.