Trust-aware collaborative filtering Recommendation in Reputation level
we propose a novel method based on users double identities in the context of social network which people can consume information as well as generate content.We acquire relationship,which we call trust,between users from users activities that performed on the related items(i.e.resource)that authors have published.Then we improve users ratings on items with their relationship with authors.Moreover,both reputation on users and items are computed to incorporate into our recommendation model to improve predictive accuracy.Compared with the classic user-based collaborative filtering recommendation,the experiment shows that our method is better in predictive accuracy.
trust collaborative filtering recommendation reputation
Hong Zhou Qing Li Fang Zhou
Department of Electronic Commerce City College of Wuhan University of Science and Technology Wuhan,C Department of computer and Information Engineering Wuhan Institute of Biological engineering Wuhan,C
国际会议
重庆
英文
2452-2457
2017-03-25（万方平台首次上网日期，不代表论文的发表时间）