A Robust Ranking Methodology Based on Diverse Calibration of AdaBoost
In subset ranking, the goal is to learn a ranking function that approximates a gold standard partial ordering of a set of objects (in our case, relevance labels of a set of documents retrieved for the same query). In this paper we introduce a learning to rank approach to subset ranking based on multi-class classification. Our technique can be summarized in three major steps. First, a multi-class classification model (AdaBoost.MH) is trained to predict the relevance label of each object. Second, the trained model is calibrated using various calibration techniques to obtain diverse class probability estimates. Finally, the Bayes-scoring function (which optimizes the popular Information Retrieval performance measure NDCG), is approximated through mixing these estimates into an ultimate scoring function. An important novelty of our approach is that many different methods are applied to estimate the same probability distribution, and all these hypotheses are combined into an improved model. It is well known that mixing different conditional distributions according to a prior is usually more efficient than selecting one “optimal” distribution. Accordingly, using all the calibration techniques, our approach does not require the estimation of the best suited calibration method and is therefore less prone to overfitting. In an experimental study, our method outperformed many standard ranking algorithms on the LETOR benchmark datasets, most of which are based on significantly more complex learning to rank algorithms than ours.
KeywordsLearning-to-rank AdaBoost Class Probability Calibration
Unable to display preview. Download preview PDF.
- 1.Busa-Fekete, R., Kégl, B., Éltető, T., Szarvas, G.: Ranking by calibrated AdaBoost. In: JMLR W&CP, vol. 14, pp. 37–48 (2011)Google Scholar
- 2.Cao, Z., Qin, T., Liu, T., Tsai, M., Li, H.: Learning to rank: from pairwise approach to listwise approach. In: Proceedings of the 24rd International Conference on Machine Learning, pp. 129–136 (2007)Google Scholar
- 4.Chapelle, O., Chang, Y.: Yahoo! Learning to Rank Challenge Overview. In: Yahoo Learning to Rank Challenge (JMLR W&CP), Haifa, Israel, vol. 14, pp. 1–24 (2010)Google Scholar
- 5.Chapelle, O., Metlzer, D., Zhang, Y., Grinspan, P.: Expected reciprocal rank for graded relevance. In: Proceeding of the 18th ACM Conference on Information and Knowledge Management, pp. 621–630. ACM, New York (2009)Google Scholar
- 10.Herbrich, R., Graepel, T., Obermayer, K.: Large margin rank boundaries for ordinal regression. In: Smola, B., Schoelkopf, S. (eds.) Advances in Large Margin Classifiers, pp. 115–132. MIT Press, Cambridge (2000)Google Scholar
- 11.Kégl, B., Busa-Fekete, R.: Boosting products of base classifiers. In: International Conference on Machine Learning, Montreal, Canada, vol. 26, pp. 497–504 (2009)Google Scholar
- 12.Li, P., Burges, C., Wu, Q.: McRank: Learning to rank using multiple classification and gradient boosting. In: Advances in Neural Information Processing Systems, vol. 19, pp. 897–904. The MIT Press, Cambridge (2007)Google Scholar
- 13.Mease, D., Wyner, A.: Evidence contrary to the statistical view of boosting. Journal of Machine Learning Research 9, 131–156 (2007)Google Scholar
- 14.Niculescu-Mizil, A., Caruana, R.: Obtaining calibrated probabilities from boosting. In: Proceedings of the 21st International Conference on Uncertainty in Artificial Intelligence, pp. 413–420 (2005)Google Scholar
- 18.Valizadegan, H., Jin, R., Zhang, R., Mao, J.: Learning to rank by optimizing NDCG measure. In: Advances in Neural Information Processing Systems, vol. 22, pp. 1883–1891 (2009)Google Scholar
- 20.Xu, J., Li, H.: AdaRank: a boosting algorithm for information retrieval. In: SIGIR 2007: Proceedings of the 30th Annual International ACM SIGIR Conference on Research and Development in Information Retrieval, pp. 391–398. ACM, New York (2007)Google Scholar