Preliminary Investigations into Ontology-based Collection Selection
John D. King, Peter Bruza · 2006
Abstract This article tackles the collection selection problem from the query side. Queries are enhanced by mapping them to subjects in an ontology; the associated subject classification terms are then employed to retrieve collections. An experimental comparison was performed with the state of the art ReDDE system, which relies on estimates of collection size to rank collections. Although the research is preliminary, there is some support to the hypothesis that this approach mitigates the need for collection size estimates in collection selection.