A privacy-protecting architecture for collaborative filtering via forgery and suppression of ratings

Abstract

Recommendation systems are information-filtering systems that help users deal with information overload. Unfortunately, current recommendation systems prompt serious privacy concerns. In this work, we propose an architecture that protects user privacy in such collaborative-filtering systems, in which users are profiled on the basis of their ratings. Our approach capitalizes on the combination of two perturbative techniques, namely the forgery and the suppression of ratings. In our scenario, users rate those items they have an opinion on. However, in order to avoid privacy risks, they may want to refrain from rating some of those items, and/or rate some items that do not reflect their actual preferences. On the other hand, forgery and suppression may degrade the quality of the recommendation system. Motivated by this, we describe the implementation details of the proposed architecture and present a formulation of the optimal trade-off among privacy, forgery rate and suppression rate. Finally, we provide a numerical example that illustrates our formulation.