Conference proceeding
TransRank: A Novel Algorithm for Transfer of Rank Learning
2008 IEEE International Conference on Data Mining Workshops, pp.106-115
12/2008
DOI: 10.1109/ICDMW.2008.42
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.
Details
- Title: Subtitle
- TransRank: A Novel Algorithm for Transfer of Rank Learning
- Creators
- Depin Chen - Univ. of Science & Technology of China, HefeiJun Yan - MicrosoftGang Wang - Virginia TechYan Xiong - Univ. of Science & Technology of China, HefeiWeiguo Fan - Virginia TechZheng Chen - Microsoft
- Resource Type
- Conference proceeding
- Publication Details
- 2008 IEEE International Conference on Data Mining Workshops, pp.106-115
- Publisher
- IEEE
- DOI
- 10.1109/ICDMW.2008.42
- ISSN
- 2375-9232
- eISSN
- 2375-9259
- Language
- English
- Date published
- 12/2008
- Academic Unit
- Business Analytics
- Record Identifier
- 9984380392702771
Metrics
2 Record Views