PARNT : A statistic based approach to extract non-taxonomic relationships of on...

• Autor(es): Serra, Ivo ; Girardi, Rosário ; Novais, Paulo
• Data: 2013
• Identificador Persistente: http://hdl.handle.net/1822/26654
• Origem: RepositóriUM - Universidade do Minho
• Assunto(s): Learning non-taxonomic relationships; Ontology; Ontology learning; Natural language processing; Machine learning
• Tipo de Documento: Documento de conferência
• Idioma: Inglês

Descrição

Learning Non-Taxonomic Relationships is a subfield of Ontology learning that aims at automating the extraction of these relationships from text. This article proposes PARNT, a novel approach that supports ontology engineers in extracting these elements from corpora of plain English. PARNT is parametrized, extensible and uses original solutions that help to achieve better results when compared to other techniques for extracting non-taxonomic relationships from ontology concepts and English text. To evaluate the PARNT effectiveness, a comparative experiment with another state of the art technique was conducted.