An agent-based approach for privacy-preserving recommender systems
Richard Cissée, Şahin Albayrak · 2007
Recommender Systems and Matchmaker Systems utilize Information Filtering technologies in order to provide personalized information in the form of recommendations of items or users with similar interests, based on a user’s long-term information needs, which are in turn derived from personal data and personal preferences. These systems are inherently privacy-critical because they essentially require personal data. Systems for sensitive domains in particular are not likely to be widely accepted by users unless they preserve the privacy of the user data they operate on. At the same time, Information Filtering technology providers as well as information providers have to be sufficiently motivated to develop and run privacy-friendly Recommender Systems and Matchmaker Systems. In the optimal case, these systems are multilaterally privacy-preserving in the sense that the privacy of all participating entities is preserved adequately. This work describes an approach for distributed multilateral Privacy-