Self-Organization Of Domain Ontology Concept Hyponymy

This paper proposes a method of domain ontology concept hyponymy self-organization, which is able to automatically extract and organize the concepts hyponymy out of web pages. First, the corpus specific to the web pages in the field of tourisms is processed in order to retrieve text data, the text concept similarity of which was calculated by concept instances mapping. Second, the concept hyponymy model is trained by the discrete concept hyponymy extracted with the feature template. The model is represented in web ontology language (OWL), i.e., described by the relationship between classes. Finally the trunk of the ontology model is created. Through the closed and open tests on about 20,000 web pages in the field of tourism from the Internet, the method has achieved remarkable results.