一种结合用户可信度与相似度的鲁棒性推荐算法

协同过滤推荐系统面临着托攻击的安全威胁。研究抵御托攻击的鲁棒性推荐算法已成为一个迫切的课题。传统的鲁棒性推荐算法在算法稳定性与推荐准确度之间难以权衡。针对该问题,首先定义一种用户可信度指标,其次改进传统的相似度计算方法,通过结合用户可信度与改进的相似度,滤除攻击概貌,为目标用户作出推荐。实验表明,与传统算法相比,本文算法具备更强的稳定性,同时保持了良好的推荐准确度。

A robust recommender algorithm based on user reliability and improved user similarity

Shilling attacks pose a significant threat to the security of collaborative filtering recommender systems. It has come to be an important task to develop the attack-resistant techniques for robust collaborative recommendation. However, traditional collaborative filtering algorithms have weakness in the balance between stability and predictive accuracy. To address this problem, user reliability is proposed and user similarity is improved, and both the user reliability and similarity is incorporated into standard collaborative filtering framework. Experiments show that the proposed algorithm performs better than state-of-the-art recommender algorithms in stability and predictive accuracy.