Journal of South China University of Technology(Natural Science Edition) ›› 2011, Vol. 39 ›› Issue (12): 70-74,99.

• Computer Science & Technology • Previous Articles     Next Articles

Online Diversified Ranking Algorithm Based on Clustering and User Clicks

Ma Qian-li  Lin Gu-li   

  1. School of Computer Science and Engineering,South China University of Technology,Guangzhou 510006,Guangdong,China
  • Received:2011-06-17 Revised:2011-08-25 Online:2011-12-25 Published:2011-11-04
  • Contact: 马千里(1980-) ,男,博士,讲师,主要从事机器学习、数据挖掘等的研究. E-mail:qianlima@scut.edu.cn
  • About author:马千里(1980-) ,男,博士,讲师,主要从事机器学习、数据挖掘等的研究.
  • Supported by:

    广东省教育部产学研结合项目( 2011B090400032) ; 教育部高等学校博士学科点专项科研基金资助项目( 20110172120027) ; 广东省自然科学基金资助项目( 9451064101003233) ; 广东省电子商务市场应用技术重点实验室开放基金资助项目( 2011GDECOF01) ; 华南理工大学中央高校基本科研业务费专项资金资助项目( 2009ZM0125,2009ZM0189,2009ZM0255)

Abstract:

In the information retrieval on the Internet,diversified ranking methods are used to provide top diverse results for users. This paper proposes an online diversified ranking algorithm CRBA based on clustering and user clicks. CRBA utilizes the similarity of documents to user feedbacks and provides diverse ranking results according to the continuous interaction of users. With the combination of the online method and the offline one,CRBA takes advantage of the topic clustering so that the convergence can be speeded up by preliminarily dividing candidate documents according to their topics. Moreover,it utilizes the merits of online ranking algorithms so that more accurate and complete estimation of users' purposes can be obtained from user clicks. Experimental results show that,as compared with the other online diversified ranking algorithms,CRBA converges more quickly and adapts well to the ranking of documents with a large amount.

Key words: information retrieval, diversification, clustering, online ranking, ranking algorithm