Content-based recommendations with approximate integer division

Recommender systems have become a vital part of e-commerce and online media applications, since they increased the profit by generating personalized recommendations to the customers. As one of the techniques to generate recommendations, content-based algorithms offer items or products that are most similar to those previously purchased or consumed. These algorithms rely on user-generated content to compute accurate recommendations. Collecting and storing such data, which is considered to be privacy-sensitive, creates serious privacy risks for the customers. A number of threats to mention are: service providers could process the collected rating data for other purposes, sell them to third parties, or fail to provide adequate physical security. In this paper, we propose a cryptographic approach to protect the privacy of individuals in a recommender system. Our proposal is founded on homomorphic encryption, which is used to obscure the private rating information of the customers from the service provider. Our proposal explores basic and efficient cryptographic techniques to generate private recommendations using a server-client model, which neither relies on (trusted) third parties, nor requires interaction with peer users. The main strength of our contribution lies in providing a highly efficient division protocol which enables us to hide commercially sensitive similarity values, which was not the case in previous works.