Design and Implementation of Privacy-preserving Recommendation System Based on MASK

In view of the issue of internal data in the construction of recommendation system may be leaked by external person. A privacy-preserving recommendation system based on MASK is designed to address it. The system consists of two parts: privacy-preserving subsystem and recommendation subsystem. First, disrupt original data to form new privacy-preserving data according to a certain probability. Second, conduct data mining with the new privacy-preserving data in recommendation subsystem. Third, its mining results will be used to help to make recommendation for users. Experimental results show that it can protect the original data without decreasing the accuracy of recommendation system.