一种改进的系统间隐私保持协同过滤推荐算法

Improved privacy-preserving collaborative filtering recommendation algorithm between systems

针对系统间协同过滤推荐过程中的隐私泄露问题,以RSA公钥密码系统和安全多方计算SMC理论为基础,提出一个安全计算模型SCM,将安全计算模型SCM应用到系统间协同过滤中,得到一个有效的隐私保持协同过滤推荐算法。算法利用安全矢量积计算用户的相似度,防止了第三方的恶意串通。实验表明,该算法不但可以保护用户的隐私不泄露给协同合作的系统,而且提高了推荐算法的精度,特别是对用户数据稀疏的小站点。

To solve the privacy disclosure problem of the recommendation algorithm between systems, this paper addresses a secure computation model based on RSA public key cryptosystem and secure multi-party computation. Applying this model to the collaborative filtering between systems, an efficient privacy-preserving collaborative filtering recommender algorithm is pro-posed. The algorithm uses secure vector product to calculate the similarity of users, prevents the untrusted third party from col- luding. Experimental results show that algorithm not only has stronger ability to protect the user＇ s privacy disclosing to the sys-tem which is cooperated, but also has better quality of recommendation, especially for the small system of sparse data.