A domain ontology represents a shared understanding of a given sector of reality (for instance, mathematics, economics, tourism etc.). Many application fields like Information Retrieval, Information Extraction and so on, as long as the Semantic Web [2], the next Web generation, need this kind of structured domain knowledge in order to add the missing semantic layout. However, an accurate search through the Internet shows the lack of large domain ontologies available to the community. In fact, building such knowledge resources requires big efforts in terms of time, costs and work due to the difficulty in identifying and properly defining domain concepts and their inter-relationships. One primary problem in this process is to establish an appropriate is - a hierarchy for the ontology. To this end, general-purpose lexical resources like WordNet [3] can be of help because they code a massive, although nonspecific, quantity of knowledge. This paper shows an original solution to the problem of building an is-a hierarchy for a domain ontology. This is achieved through the automatic enrichment and reorganization of the WordNet hierarchy by properly adding domain knowledge structured in the form of concept trees.
展开▼