Preference Learning to Rank with Sparse Bayesian

In this paper, we propose a sparse Bayesian approach to learn ranking function from labeled data. The ranking function can be used to define an ordering among documents according to their degree of relevance to the user query. This ranking function is more efficient and accurate than the function leaned by proposed approaches. Experimental results on document retrieval dataset show that the generalization performance of it is competitive with SVM-based ranking method and Gaussian process based method.