Semi-automatic Construction of a General Purpose Ontology
Andrew Philpot, Michael Fleischman, Eduard H. Hovy · 2003
Ontologies are hierarchically organized networks of conceptual information. Used to systematize and model domain knowledge, they play an important role in artificial intelligence, natural language processing, information integration, electronic commerce, and related fields. In this paper, we briefly describe the characteristics of ontologies, and then discuss the Omega ontology, a large general-purpose ontology created semi-automatically from WordNet, Mikrokosmos, and a newly created upper semantic model, and note the relationship between Omega and its predecessor Sensus. We detail the procedures used to create the Omega base ontology and to merge in new ontological information, such as custom ontologies derived from glossaries and instances obtained by data mining. We survey some of the applications in which we have applied Omega. We highlight the tools used and characteristics of the Lisp environment which facilitate the development and maintenance of the ontological structure. We conclude by discussing current and future work.