Amalia Todiraşcu, Dan Gâlea, François de
Beuvron
Using Ontologies and WordNet
Abstract.
This paper presents an information retrieval (IR) system integrating semantic
resources for searching a base of documents. The documents cover a limited
domain (the prototype has been tested for three small French corpora: heart
surgery – 70 000 words, NLP articles – 250 000 words). It uses a semantic-based
indexing method, using domain ontologies (built and extended dynamically), as
well as a general-purpose thesaurus (WordNet [12]). Domain ontologies are
representing in description logics (DL), a family of knowledge representation
formalisms. The paper focuses on methods relating general semantic resources
(EuroWordNet, an extension of WordNet) with domain-specific ontologies. |