The Mapping Research Between WordNet And Chinese Classification Thesaurus

Along with the development of the network, the quantity of information increases sharply but lacks of designability and that makes it harder to find necessary knowledge for the user, and also the difficulty of maintaining the information system is greatly increased. Based on this, through the establishment and sharing of the ontology efficiently so as to achieve the sharing of information, knowledge and reuse attract the attention of more and more scholars and become research hot spot.In reality, different fields and even the same field tend to have more heterogeneous ontology, so the methods and systems that could be applied in all ontology and achieve the sharing of ontology perfectly do not see more. Ontology mapping is a method that can achieve re-use and interoperability between heterogeneous ontologies by establishing the semantic relations and outputting the mapping pairs of concept.Chinese resources is an important component of the global knowledge and information network, to achieve the mapping between Chinese ontologies or between Chinese and foreign ontologies is a important aspect in ontology sharing and reuse. At present, there are lots of researches about ontology mapping, but the quantities of mapping methods and systems about mapping between Chinese ontologies or between Chinese and foreign ontologies are limited, and some do not consider the semantic ambiguity between cross-language ontology. This paper is committed to the mapping research between WordNet and Chinese classification thesaurus, aiming to expand the vocabulary of Chinese classification thesaurus, thus to provide a mapping standard and foundation for the mapping between Chinese ontology. All of this make the result of mapping between Chinese ontology more accurate and reflect the structure and characteristics of Chinese more veracious.This paper adopted Chinese WordNet and Web version of Chinese classification thesaurus. The latter is not a standard ontology, but it contain a great quantity of vocabularies and semantic features, which characteristic are similar with the ontology. First, this article use Protege to describe the Chinese classification thesaurus with the noumenal language of OWL converting it to the standard form. Second, this article preprocesses the two ontology and extracts the candidate mapping pairs through the establishment of virtual documents. Third, it calculate the similarity of concepts, attribute and instances based on HowNet. Then it revise the mapping results on certain significance by using the method of reasoning. At last, I prove the validity of the mapping method through experiment analysis putted forward in this paper.