Optimal Query Expansion (QE) Processing Methods with Semantically Encoded Structured Thesauri Terminology

While researchers have explored the value of structured thesauri as controlled vocabularies for general information retrieval (IR) activities, they have not identified the optimal query expansion (QE) processing methods for taking advantage of the semantic encoding underlying the terminology in these tools. The study reported on in this article addresses this question, and examined whether QE via semantically encoded thesauri terminology is more effective in the automatic or interactive processing environment. The research found that, regardless of end-usersʹ retrieval goals, synonyms and partial synonyms (SYNs) and narrower terms (NTs) are generally good candidates for automatic QE and that related (RTs) are better candidates for interactive QE. The study also examined end-usersʹ selection of semantically encoded thesauri terms for interactive QE, and explored how retrieval goals and QE processes may be combined in future thesauri-supported IR systems.