Acquiring taxonomic conceptual relationships using a machine readable dictionary and text corpus

摘要

In order to reduce a user's effort in constructing a domain ontology, we propose a method to support the construction of domain ontologies by using a machine-readable dictionary (MRD) and text corpus. We construct a domain ontology with large numbers of concepts based on the hierarchical structure of the MRD. Furthermore, the domain-oriented relationships of taxonomy is obtained from text corpus. We take context similarity based on WordSpace to extract taxonomic relationships from text corpus. Taxonomic and non-taxonomic relationships are distinguished with respect to the distance between terms in the text. Through the case study with CISG, we make sure that our system can work to support the process of construction domain ontologies with a MRD and text corpus.