Author(s)

Chouchak Chan, Ao Liu

Corresponding Author

Ao Liu

Abstract

In the scenario of multi-source data fusion, the issues of privacy protection and utility improvement of recommendation systems have increasingly attracted attention. Traditional recommendation algorithms often rely on users' personal information and behavioral data. However, in the big data environment, the risk of user privacy leakage has significantly increased. It is of great significance to study how to improve the performance of recommendation systems on the basis of ensuring user privacy. This paper discusses the recommendation algorithm based on differential privacy mechanism and homomorphic encryption technology. Through the fusion and analysis of multi-source data, a novel privacy protection framework is proposed. By integrating collaborative filtering with deep learning models, a balance between utility and privacy has been achieved. The experimental results show that the proposed method has good feasibility in terms of user privacy protection and is superior to traditional methods in recommendation effect, providing new ideas and directions for future research on recommendation systems.

Keywords

multi-source data fusion; recommendation system; privacy protection; utility improvement