Analysis of the query expansion for medical collections using mutual information, information gain and MeSH ontology
José M. Perea‐Ortega, Arturo Montejo‐Ráez, Manuel C. D ́ iaz-Galiano, M.D. Miguel, A Alejandro Garc · 2011
In this paper we show several experiments related to query expansion in medical information retrieval systems, using ImageCLEFmed as a evaluation fra- mework. We have evaluated di!erent query expansion techniques such as PRF and others based on the use of the medical ontology MeSH. Specifically, in this latter case, di!erent types of filtering have been tested, based on a previous selection of MeSH categories and the application of concepts such as information gain and mu- tual information. The results show that the PRF expansion slightly improves the base case without expanding the query, while the desirability of the expansion with MeSH terms is dicult to determine. However, we can deduce that the expansion with MeSH terms introduced too much noise during the retrieval process, at least by the techniques applied in this work.