Extraction And Organizational Domain Ontology Hyponymy Research

The domain ontology backbone is the domain concepts hyponymy, namely the domain concepts of hierarchy relation. At present the organizational domain concepts hyponymy is mainly in a manual way, meanwhile, and need to participate in the domain knowledge engineers and generate the backbone of the domain ontology. And there are a lot of domain knowledge text on the Internet, the concept of domain knowledge text contains a large number of the domain concepts hyponymy, using the manually organized domain ontology backbone is time-consuming, so An automated build ontology backbone is the future trends and an important research direction. This article from the domain concept of hyponymy extraction, the domain ontology concept of hyponymy of an organization and the application in the field of tourism, has summarized as follows:1) The domain ontology of concepts hyponymy of the extraction. This paper proposed a method to extracted domain ontology concept hyponymy. It based on free text extracted object, it extracted domain ontology concept hyponymy with the cascaded conditional random fields algorithm, in low-level conditional random fields model firstly to identify the domain concepts, and extract concepts in order composition, then obtained many the paired concepts with the characteristics of the template definition, in high-level model the paired concepts use the ’UD’ label with annotating semantic, the hyponymy relation between the two concepts can be identified.2) The domain ontology concept of hyponymy organization. This paper proposed an automatic method of organization domain ontology concepts hyponymy, the method is based on the extraction of concept hyponymy objects in web pages, the extracted concept hyponymy appeared in the form of discrete, and training into the model, namely the concept of hyponymy model. The training model and OWL on the relationship between the concepts description is consistent, namely the concept hyponymy is to use the class relations representation, and finally generate the basic backbone of the ontology model.3) The construction of the test system. The system can be taken for this article the concept hyponymy to provide technology, such as the segments, corpus tagging, automatic processing of the condition random fields (CRFs) training format, automatically generate conditions for random fields (CRFs) training model, automatic extracted concept hyponymy, automatic organization domain ontology backbone.