一种基于时间特性的二部图推荐算法

广西师范大学学报（自然科学版） ›› 2013, Vol. 31 ›› Issue (3): 81-86.

一种基于时间特性的二部图推荐算法

周俊临, 傅彦, 孔祥迎, 丁建勇

电子科技大学互联网科学中心,四川成都611731

收稿日期:2013-06-05 出版日期:2013-09-20 发布日期:2018-11-26
通讯作者: 周俊临(1981—),男(回族),河南潢川人,电子科技大学副教授,博士。E-mail:jlzhou@uestc.edu.cn
基金资助:
国家自然科学基金资助项目(61103109,11105024,61003231);中央高校基本科研业务费基金资助项目(ZYGX2011J057,ZYGX2012J071,ZYGX2012J085);四川省科技基金资助项目(2010HH0002,2011GZ0106,20112Z0001,2012RZ0002,2012RZ0003);高等学校博士学科点专项科研基金资助项目(20120185120017)

Recommendation Based on Bipartite Graph with Time Property

ZHOU Jun-lin, FU Yan, KONG Xiang-ying, DING Jian-yong

Web Sciences Center,University of Electronic Science and Technology of China,Chengdu Sichuan 611731,China

Received:2013-06-05 Online:2013-09-20 Published:2018-11-26

摘要/Abstract

摘要： 随着互联网的发展,大量商品信息不断涌现,从而产生了信息过载问题。推荐系统作为解决此问题的有效手段,近年来得到快速发展。现存方法大多以用户行为和商品内容相似性为基础,利用用户购买记录和商品描述信息来产生推荐结果。事实上,用户的购买行为与时间也有着密切的联系。例如,最近购买的商品往往更能体现用户的当前兴趣。因此,在传统基于相似性推荐的基础上,本文提出一种基于时间特性的二部图推荐方法,通过调整初始资源权重分布体现用户兴趣随时间的变化趋势。实验证明,本文提出的方法在面向时间的Top-N命中率上有较大幅度提升。本文工作不仅对现有推荐算法的效果提高具有实际意义,对推荐系统在真实商业环境中的应用也有很大促进作用。

关键词: 推荐系统, 时间特性, 二部图, 资源分配

Abstract: With the rapid development of the World Wide Web,a large amount of trading information is available on the Internet.This abundance of information has created the need to help users find resources that match their individual goals and interests.This problem has been in the focus of recent research and an approach of recommendation system has been proposed.In the traditional system,users are provided with assistance in making selections according to the similarity of users' behavior and the products' information.Actually,users' interests vary gradually over time and their recent activities reveal their current interests.In this paper,a bipartite graph recommendation method is proposed based on time property,in which the time variation is taken into account and a new approach of initial resource adjustment is adopted to endow the recommended result with timeliness.Experiments show notable improvement on the Top-N hits metric.This method is not only of real value for improving the performance of recommendation system,but also has an active effect on the applications of recommendation system in E-Commerce.

Key words: recommender system, time property, bipartite graph, resource allocation

中图分类号:

TP311

周俊临, 傅彦, 孔祥迎, 丁建勇. 一种基于时间特性的二部图推荐算法[J]. 广西师范大学学报（自然科学版）, 2013, 31(3): 81-86.

ZHOU Jun-lin, FU Yan, KONG Xiang-ying, DING Jian-yong. Recommendation Based on Bipartite Graph with Time Property[J]. Journal of Guangxi Normal University(Natural Science Edition), 2013, 31(3): 81-86.

参考文献

[1] 刘建国,周涛,汪秉宏.个性化推荐系统的研究进展[J].自然科学进展,2009,19(1):1-15.
[2] ZHANG Yi-cheng,BLATTNER M,YU Yi-kuo.Heat conduction process on community networks as a recommendation model[J].Phys Rev Lett,2007,99(15):154301.
[3] SARWAR B M,KARYPIS G,KONSTAN J A,et al.Application of dimensionality.reduction in recommender system:a case study[EB/OL].(2000-08-20)[2013-05-20].http://robotics.stanford.edu/~ronnyk/WEBKDD2000/papers/sarwar.pdf.
[4] 胡海波,王科,徐玲,等.基于复杂网络理论的在线社会网络分析[J].复杂系统与复杂性科学,2008,5(2):1-14.
[5] ZHOU Tao,REN Jie,MEDO M,et al.Bipartite network projection and personal recommendation[J].Physical Review E,2007,76(4):046115.
[6] HUANG Zan,CHEN Hsin-chun,ZENG D.Applying associative retrieval techniques to alleviate the sparsity problem in collaboriative filtering[J].ACM Transactions on Information Systems,2004,22(1):116-142.
[7] 杨杰,陈恩红.推荐系统中的用户兴趣模式检测[EB/OL].(2008-12-18)[2013-5-20].http://www.paper.edu.cn/files/resultshare/chenenhong/C.pdf.
[8] KOREN Y.Collaborative filtering with temporal dynamics[C]//Proceedings of the 15th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining.New York:ACM Press,2009:447-456.
[9] LATHIA N,HAILES S,CAPRA L,et al.Temporal diversity in recommender systems[C]//Proceedings of ACM SIGIR International Conference on Research and Development in Information Retrieval.New York:ACM Press,2010:210-217.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed