Automatic Generation of Glosses in the OntoLearn System
Alessandro Cucchiarelli (1), Roberto Navigli (2), Francesca Neri (1), Paola Velardi (2)
(1) DIIGA - Universita' Politecnica delle Marche - via Brecce Bianche 12, I60131 Ancona - Italy; (2) Dipartimento di Informatica - UniversitÓ di Roma "La Sapienza"- via Salaria 113, I00196 Roma - Italy
OntoLearn is a system for automatic acquisition of specialized ontologies from domain corpora, based on a syntactic pattern matching technique for word sense disambiguation, called structural semantic interconnection (SSI). We use SSI to extract from corpora complex domain concepts and create a specialized version of WordNet. In order to facilitate the task of domain specialists who inspects and evaluate the newly acquired domain ontology, we defined a method to automatically generate glosses for the learned concepts. Glosses provide an informal description, in natural language, of the formal specifications of a concept, facilitating a per-concept evaluation of the ontology by domain specialists, who are usually unfamiliar with the formal language used to describe a computational ontology. The proposed evaluation framework has been tested in a financial domain.
Ontology learning, ontology evaluation, gloss generation.