WebExplorer: A TOOL FOR ONTOLOGY-BASED INFORMATION EXPLORATION
Feza Baskaya, Anne Keskimaa, Jaana Kekäläinen, Kalervo Järvelin · 2009
Vocabulary mismatch and muddled search goals plague users of traditional information retrieval (IR) systems. The former problem affects both monolingual IR and cross-language IR in particular, even if the user is able to express her information need clearly. This is due to the number of ways the search concepts may be expressed in texts and also due to the user’s incomplete skills in using foreign languages. The latter problem is inherent in information seeking and retrieval: information needs evolve as users learn about their information problems. Traditional approaches to vocabulary mismatch have been query expansion, and document indexing through thesauri, or recently, ontologies. The latter two assume document indexing (or annotation) and suffer from costs, inconsistencies and the great variety of ontologies in use. For muddled search goals there has not been proper support: users need to examine retrieved documents and reformulate their need statements. In the present paper we describe and evaluate a novel tool, WebExplorer, for ontology-based multilingual document exploration. It provides an aid for multiple-source searching and document analysis through summarization, clustering, and classification, among others.