A Game Theoretic Framework for Data Privacy Preservation in Recommender Systems
Report a problem or upload filesIf you have found a problem with this lecture or would like to send us extra material, articles, exercises, etc., please use our ticket system to describe your request and upload the data.
Enter your e-mail into the 'Cc' field, and we will keep you updated with your request's status.
We address the fundamental tradeoff between privacy preservation and high-quality recommendation stemming from a third party. Multiple users submit their ratings to a third party about items they have viewed. The third party aggregates the ratings and generates personalized recommendations for each user. The quality of recommendations for each user depends on submitted rating profiles from all users, including the user to which the recommendation is destined. Each user would like to declare a rating profile so as to preserve data privacy as much as possible, while not causing deterioration in the quality of the recommendation he would get, compared to the one he would get if he revealed his true private profile. We employ game theory to model and study the interaction of users and we derive conditions and expressions for the Nash Equilibrium Point (NEP). This consists of the rating strategy of each user, such that no user can benefit in terms of improving its privacy by unilaterally deviating from that point. User strategies converge to the NEP after an iterative best-response strategy update. For a hybrid recommendation system, we find that the NEP strategy for each user in terms of privacy preservation is to declare false rating only for one item, the one that is highly ranked in his private profile and less correlated with items for which he anticipates recommendation. We also present various modes of cooperation by which users can mutually benefit.
Link this pageWould you like to put a link to this lecture on your homepage?
Go ahead! Copy the HTML snippet !