Improve Diversity-oriented Biomedical Information Retrieval using Supervised Query Expansion

In the paper, we propose a novel diversity-oriented biomedical information retrieval method based on supervised query expansion. Our method aims to obtain the most relevant and diversified terms to enrich user queries for better interpreting the information needs. We first propose a diversity-oriented labeling strategy to annotate the usefulness of candidate expansion terms. We then extract both the context-based and resource-based term features to represent terms as feature vectors. In model training, we propose a diversity-oriented group sampling method to modify the loss function of learning-to-rank for accurate biomedical term ranking. Experimental results on TREC Genomics datasets show that our method is effective in improving the performance of biomedical information retrieval in terms of both the relevance and the diversity.