Retrieval feedback for query design in MEDLINE. A comparison with expert network and LLSF approaches

Abstract

Query design is a vibrant research focus in information retrieval. The objective is to modify a user's original query into one that is more effective for retrieval. Researchers have proposed and investigated a variety of strategies to support query design in MEDLINE. This paper examines a query design method built within the framework of retrieval feedback. In particular, the effectiveness of this method is compared with the effectiveness of competing methods that utilize query mapping functions constructed in an expert network or built using the Linear Least Squares Fit approach. The comparison indicates that retrieval feedback offers an approach that is just as effective as these alternative approaches. Moreover, it has the advantage of being simpler to implement.