摘要： 提出一种基于信任机制的协同过滤推荐算法，其中，直接信任度基于共同评价项目得出，推荐信任度通过对项目的预测得出。借鉴社会网络中人与人之间的信任评价方法，使用户之间的相似度计算更加准确，从而为目标用户提供更好的推荐结果。实验结果表明，该模型提高了信任度预测的准确性及系统的推荐质量。

关键词: 电子商务, 推荐系统, 协同过滤, 信任模型, 推荐算法

Abstract: This paper proposes a collaborative filtering recommendation algorithm based on trust mechanism. Direct trust is based on common rating data and indirect trust is based on the predict data. By considering the social network between Person’s trust assessment method, the similarity between the users is more accurate, so that better neighbors for a target user and the better recommendation result can be gained. Experimental result indicates that based on the trust model’s recommendation system, can consistently achieve better prediction accuracy and improves system’s recommendation quality effectively.

Key words: E-commerce, recommendation system, collaborative filtering, trust model, recommendation algorithm

中图分类号: 

• TP391