期刊
PHYSICA A-STATISTICAL MECHANICS AND ITS APPLICATIONS
卷 436, 期 -, 页码 236-245出版社
ELSEVIER
DOI: 10.1016/j.physa.2015.05.060
关键词
Collaborative filtering; Interevent time distribution; Bipartite network
资金
- Zhejiang Provincial Natural Science Foundation of China [LY12A05003, LQ14F030009, LQ13F030015]
- National Natural Science Foundation of China [11305042, 61403114, 61305148, 61304150]
- Hangzhou Normal University
- China Postdoctoral Science Foundation [2013M540493]
Recently, many scaling laws of interevent time distribution of human behaviors are observed and some quantitative understanding of human behaviors are also provided by researchers. In this paper, we propose a modified collaborative filtering algorithm by making use the scaling law of human behaviors for information filtering. Extensive experimental analyses demonstrate that the accuracies on MovieLensand Last.fm datasets could be improved greatly, compared with the standard collaborative filtering. Surprisingly, further statistical analyses suggest that the present algorithm could simultaneously improve the novelty and diversity of recommendations. This work provides a creditable way for highly efficient information filtering. (C) 2015 Elsevier B.V. All rights reserved.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据