基于差分隐私的Slope One协同过滤推荐算法

The Slope One algorithm is a collaborative filtering recommendation algorithm that is simple, efficient, and highly recommended. However, it is difficult to provide a strict privacy guarantee. The potential attacker can steal the user\'s rating record by observing the user\'s recommendation result, which will pose a great threat to the user\'s privacy. To solve this problem, a DP-Slope One collaborative filtering recommendation algorithm that satisfies differential privacy protection is proposed. First, construct a user-item rating similarity matrix, and then use the Laplace noise mechanism to perturb the user-item rating similarity matrix to satisfy the differential privacy model. Finally, it is verified on the real data set. Experimental results show that the DP-Slope One recommendation algorithm based on differential privacy protection can protect user privacy based on the recommendation accuracy.