Employing query expansion models to help patients diagnose themselves
Fangmei Lu · CLEF (Working Notes) · 2015
In the paper we use two query expansion models, Kullback-Liebler Divergence(KLD) model and parameter-free Bose-Einstein statistics-based (Bo1) model, to improve effectiveness of information retrieval systems and help lay people search relevant medical information for diagnosing themselves. Compared with Baseline BM25, the results of Bo1 models with 3 feedback documents are better than baseline but are not statistically significant, and the performance of Bo1 is generally better than KLD.