A polyrepresentational approach to interactive query expansion

Abstract

Interactive Query Expansion (IQE) presents suggested terms to the user during their search to enable better Information Retrieval (IR). However, IQE terms are poorly used, and tend to lack information meaningful to the user. The lack of cognitive and functional support during query refinement is a well documented problem, and despite the work carried out, it is still an under researched area. This stagnation in progress has been partly due to the long held belief that users are able to make good IQE term selections, and that the de facto way IQE terms are presented is effective. In this paper, we introduce a novel method to improve the presentation of IQE terms by providing supplementary information alongside them. We describe a user study that compared our novel polyrepresentational approach to IQE against a conventional IQE system and a baseline system. Our findings have shown that a polyrepresentational approach to IQE can address the ambiguity and uncertainty surrounding IQE, and improve the perceived usefulness of the terms. Copyright 2009 ACM.