标题：A novel method for matrix factorization in recommender system using item's information
作者：Zhao, Jianli ;Wu, Wenmin ;Zhang, Chunsheng ;Meng, Fang
作者机构：[Zhao, Jianli ;Wu, Wenmin ;Zhang, Chunsheng ;Meng, Fang ] School of Information Science and Engineering, Shandong University of Science and Technology 更多
来源：Journal of Computational Information Systems
摘要：Collaborative Filtering has been thoroughly investigated these years and one of the most popular approaches to it is matrix factorization. Traditional matrix factorization models only rely on rating data. In this paper, we take the content of items into account to improve the matrix factorization. We proposed a novel action-content cluster method with fuzzy logic, and calculate user's similarities based on user's different memberships to different groups (fuzzy set). We use user's similarities to optimize the matrix factorization during the learning process. Experiments with Movielens datasets show that our proposed model significantly outperforms the baseline and original model. ©, 2015, Binary Information Press. All right reserved.