文章模拟用户的打分过程,利用信任来改进用户评分。同时研究用户声誉在资源选择过程中的作用,与项目的声誉结合来解决同质资源泛滥的问题。实验结果显示,本文提出的基于声誉的协同过滤推荐方法能更准确地反映用户的偏好情况和资源的质量,从而提高推荐的准确率。
The paper simulates users' grading process and improves scores with trust. The paper also researches on the influence of user reputation in the process of selecting resources and combines the item reputation to solve the overflow problem of homogeneous resources. The experimental result shows that the reputation-based collaborative filtering recommendation method can reflect users' preference as well as resource' s quality accurately, thereby improving the precision of recommendation.