DocumentCode :
2129538
Title :
TransRank: A Novel Algorithm for Transfer of Rank Learning
Author :
Chen, Depin ; Yan, Jun ; Wang, Gang ; Xiong, Yan ; Fan, Weiguo ; Chen, Zheng
Author_Institution :
Univ. of Sci. & Technol. of China, Hefei
fYear :
2008
fDate :
15-19 Dec. 2008
Firstpage :
106
Lastpage :
115
Abstract :
Recently, learning to rank technique has attracted much attention. However, the lack of labeled training data seriously limits its application in real-world tasks. In this paper, we propose to break this bottleneck by considering the cross-domain ldquotransfer of rank learningrdquo problem. Simultaneously, we propose a novel algorithm called TransRank, which can effectively utilize the labeled data from a source domain to enhance the learning of ranking function in the target domain. The proposed algorithm consists of three key steps. Firstly, we introduce a utility function to select the k-best queries from the source domain labeled data. Secondly, feature augmentation is performed on both source and target domain data, which can straightly adapt the ranking information from source domain to target domain. Finally, we utilize the classical ranking SVM to learn the enhanced ranking function on the augmented features. Experimental results on benchmark datasets well validate our proposed TransRank algorithm.
Keywords :
learning (artificial intelligence); query processing; support vector machines; utility theory; cross-domain transfer; feature augmentation; labeled training data; query selection; ranking SVM; ranking function; source domain labeled data; target domain rank learning; transfer rank learning technique; transrank algorithm; utility function; Asia; Collaborative work; Conferences; Data mining; Humans; Information retrieval; Machine learning; Machine learning algorithms; Support vector machines; Training data;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Data Mining Workshops, 2008. ICDMW '08. IEEE International Conference on
Conference_Location :
Pisa
Print_ISBN :
978-0-7695-3503-6
Electronic_ISBN :
978-0-7695-3503-6
Type :
conf
DOI :
10.1109/ICDMW.2008.42
Filename :
4733928
Link To Document :
بازگشت