Improved Boosting Algorithms Using Confidence-rated Predictions
Machine Learning - The Eleventh Annual Conference on computational Learning Theory
Optimizing search engines using clickthrough data
Proceedings of the eighth ACM SIGKDD international conference on Knowledge discovery and data mining
An efficient boosting algorithm for combining preferences
The Journal of Machine Learning Research
AdaRank: a boosting algorithm for information retrieval
SIGIR '07 Proceedings of the 30th annual international ACM SIGIR conference on Research and development in information retrieval
Query dependent ranking using K-nearest neighbor
Proceedings of the 31st annual international ACM SIGIR conference on Research and development in information retrieval
Knowledge transfer for cross domain learning to rank
Information Retrieval
Learning to rank only using training data from related domain
Proceedings of the 33rd international ACM SIGIR conference on Research and development in information retrieval
Semi-supervised ranking for document retrieval
Computer Speech and Language
Transfer learning with one-class data
Pattern Recognition Letters
Hi-index | 0.00 |
Learning to rank for information retrieval needs some domain experts to label the documents used in the training step. It is costly to label documents for different research areas. In this paper, we propose a novel method which can be used as a cross-domain adaptive model based on importance weighting, a common technique used for correcting the bias or discrepancy. Here we use "cross-domain" to mean that the input distribution is different in the training and testing phases. Firstly, we use Kullback-Leibler Importance Estimation Procedure (KLIEP), a typical method in importance weighing, to do importance estimation. Then we modify AdaRank so that it becomes a transductive model. Experiments on OHSUMED show that our method performs better than some other state-of-the-art methods.