基于信用矩阵的推荐机制  

Credit Martix Based Information Recommendation Mechanism

作　　者：潘红艳[1] 史晓燕[1] 

机构地区：[1]浙江工商职业技术学院信息工程学院,宁波315012

出　　处：《图书情报工作》2010年第14期50-52,125,共4页Library and Information Service

摘　　要：如何准确高效地提供给用户需要的信息,是信息推荐研究的核心。提出一种推荐机制——基于信用矩阵的推荐机制,借鉴协同推荐的实现思想,引入一个信用矩阵,在向目标用户推荐时,不是基于最近邻产生推荐,而是基于用户之间的信用产生对目标用户的推荐。在MovieLens上的实验证明,这种算法在预测精度上较传统的推荐算法有很大的提高。How to accurately and effectively provide the information which the user needs, is the core of the information recommendation research. This paper proposes a recommendation mechanism which is based on credit matrix. It uses the collaborative filtering for reference and uses a credit matrix. It recommends the information to users based on the credits among users. The test in the MovieLens proves that this method can obtain a better predictive precision, compared with traditional recommendation algorithm.

关 键 词：信用矩阵 信用管理 协同推荐 

分 类 号：G350[文化科学—情报学]