Questioning Query Expansion: An Examination of Behaviour and Parameters *

Abstract

In information retrieval, queries can fail to find documents due to mismatch in terminology. Query expansion is a well-known technique addressing this problem, where additional query terms are automatically chosen from highly ranked documents, and it has been shown to be effective at improving query performance. However, current techniques for query expansion use fixed values for key parameters, determined by tuning on test collections. In this paper we show that these parameters may not be generally applicable , and more significantly that the assumption that the same parameter settings can be used for all queries is invalid. Using detailed experiments with two test collections, we demonstrate that new methods for choosing parameters must be found. However, our experiments also demonstrate that there is considerable further scope for improvement to effectiveness through better query expansion.