Private Recommendations Generation for Vertically Partitioned Datasets

With the rapid expansion of cloud services and the development of online trading and social network sites, various recommender systems are becoming prevalent and have been accepted well by the clients. It is necessary to provide accurate and reliable recommendations for users so that they trust the system and accept the recommendations. The more utility data is available to the system, the more accurate recommendations can be generated. There exist such systems where the data are distributed into multiple organizations and to generate satisfactory recommendations, these data need to be integrated. However, data related to user preferences are usually confidential and it is unlikely that any organization will agree to share its own data with other entities. To provide advanced facilities for securing privacy of data in distributed recommender systems, we propose a new system called Privacy-preserving Weighted Slope One Predictor for Vertically Partitioned Data (PWSOP-VPD). It generates recommendations in a privacy-preserving manner by integrating users' data distributed among several different organizations. In this protocol, private records are encrypted by means of homomorphic encryption and secure recommendations are provided based on the encrypted data. Our paper examines the PWSOP-VPD protocol for a real world dataset and shows that the proposed method is secure and outperforms previous solutions, which addressed more restricted special cases of this general scenario.