Automatic Method Of Domain Ontology Construction based on Characteristics of Corpora POS-Analysis

Olena Orobinska (1)

1 DMD

It is now widely recognized that ontologies, are one of the fundamental cornerstones of knowledge-based systems. What is lacking, however, is a currently accepted strategy of how to build ontology; what kinds of the resources and techniques are indispensables to optimize the expenses and the time on the one hand and the amplitude, the completeness, the robustness of en ontology on the other hand. The paper offers a semi-automatic ontology construction method from text corpora in the domain of radiological protection. This method is composed from next steps: 1) text annotation with part-of-speech tags; 2) revelation of the significant linguistic structures and forming the templates; 3) search of text fragments corresponding to these templates; 4) basic ontology instantiation process

Mots clés: Information Extraction; NLP; Relation Recognition Rules; Instance Recognition Rules; Semantic Analysis

Domaines: Informatique et langage [cs.CL]

Olena Orobinska. Automatic Method Of Domain Ontology Construction based on Characteristics of Corpora POS-Analysis. XV Russian conference Internet and Modern Society, Oct 2012, Sent-Petersburg, Russia. pp.209-212. ⟨hal-00987456⟩