计算机应用研究2016，Vol.33Issue(6)：1628-1632,5.DOI:10.3969/j.issn.1001-3695.2016.06.007
基于评论的隐式社交关系在推荐系统中的应用
Reviews based implicit social relations in recommender systems
赵亚辉 1刘瑞1
作者信息
- 1. 北京航空航天大学 软件开发环境国家重点实验室，北京 100191
- 折叠
摘要
Abstract
Conventional researches less utilize review data to enhance the recommendation systems.Reviews contain rich o-pinions of users on specific items,and can construct relations among users,as well as items.This paper proposed a model of review based implicit social matrix factorization.The model built neighbor relationship for users and items by similarity be-tween reviews,and integrated the implicit relations into the social recommendation framework SocialMF.Experiment shows that the proposed method reduces the RMSE of recommendation significantly by about 3% on Amazon review dataset,especial-ly when system sufferes from the cold start dilemma.As a result,it indicates that rich knowledge from active users’reviews can be transferred for improving recommender system.关键词
推荐系统/矩阵分解/评论数据/隐式社交关系Key words
recommender system/matrix factorization/review data/implicit social relations分类
信息技术与安全科学引用本文复制引用
赵亚辉,刘瑞..基于评论的隐式社交关系在推荐系统中的应用[J].计算机应用研究,2016,33(6):1628-1632,5.
