Privacy-Preserving Collaborative Filtering Protocol Based on Similarity between Items

A recommendation system enables us to take information from huge datasets about tastes effectively. Many cryptographical protocols for computing privacy-preserving recommendation without leaking the privacy of users are proposed. However, the current issue is the large computational overhead depending the number of users. Hence, the application of the protocol is limited within small communities. In this paper, we address the issue of scalability by replacing the similarity between users by that of between items. Since the similarities between items can be publicly available, the recommendation steps are processed without dealing with confidential information such as the similarities between users. We propose an efficient scheme by using item-item similarities for providing a prediction of arbitrary values of rating. We show the performance and the accuracy evaluation of our proposed scheme based on a numerical experiment.