An Agent-Based Approach for Privacy-Preserving Recommender Systems
Recommender Systems are used in various domains to generate personalized information based on personal user data. Preserving the privacy of all participants is an essential requirement of the underlying Information Filtering architectures, because the deployed Recommender Systems have to be accepted by privacy-aware users as well as information and service providers. Existing approaches neglect to address privacy in this multilateral way. We have developed an approach for privacy-preserving Recommender Systems based on Multi-Agent System technology which enables applications to generate recommendations via various filtering techniques while preserving the privacy of all participants. We describe the main modules of our solution as well as an implemented application based on this approach.