Movie Recommendation in Heterogeneous Information Networks
Recommender systems,as we all know,have gained tremendous popularity over the past few years and been widely used in e-commerce.Recent researches have improved recommendation performance combining additional user and item relationships with hybrid recommender systems.However,most of these studies only consider a single type of relationship while in application recommendation problems always exist in heterogeneous information networks.In this paper,we combine the model-based collaborative filtering with heterogeneous information networks to create an efficient recommendation model.We adopt meta-path to denote multiple types of entities and relationships in heterogeneous information networks and use PathSim as the similarity measurement.We employ a nonnegative matrix factorization based collaborative filtering recommendation method under each meta-path.Furthermore,we cluster users or items into subgroups and our method shows advantages through empirical studies.
heterogeneous information network meta-path NMF recommendation
Yannan Chen Ruifang Liu Weiran Xu
School of Information and Communication Engineering Beijing University of Posts and Telecommunications Beijing,China
国际会议
重庆
英文
637-640
2016-03-20（万方平台首次上网日期，不代表论文的发表时间）