Research On Knowledge Acquisition Method Supported By Semantic Web Rough Ontology

The semantic Web rough ontology is a kind of rough ontology applied to the semantic Web, and it is usually described by ontology description language. It is an expansion of the semantic Web ontology. As the basic part of rough ontology, rough concepts and rough relations between them are obtained from the precise concepts and precise concept relations between them respectively combining with upper approximation set and lower approximation set in the rough set theory. As kind of knowledge model, the semantic Web rough ontology can express not only precise classified knowledge, but also the rough classified knowledge represented approximatively by the precise classified knowledge.Because the semantic Web rough ontology which is constructed by the method of concept lattice contains only the ontology schema, and can’t form a complete knowledge base. So this thesis pointedly puts forward the knowledge acquisition method supported by semantic Web rough ontology.Firstly, the semantic Web rough ontology is defined. The expansion from semantic Web ontology to semantic Web rough ontology is described, the composition model of semantic Web rough ontology is put forward and the OWL description method of semantic Web rough ontology is summarized.Secondly, a knowledge acquisition method supported by semantic Web rough ontology is put forward. The main steps of the method is:Rules defining, according to the actual of domain application, some rules are defined; Semantic Web rough ontology parsing, the classes, properties, instances and axioms defined in the semantic Web rough ontology are parsed out and stored into a relational database; Document preprocessing, documents of domain are collected and transformed into text documents, and are stored into the text document base; Grammar analysing, lexical analysis, syntactic analysis and sentence transformation are conducted to the text documents, and noun phrase structure and verb phrase structure that can be directly used for knowledge acquisition are obtained; Semantic annotation, words in text documents are mapped to the classes or properties of semantic Web rough ontology; Semantic reasoning, by using semantic annotation results, semantic reasoning is conducted based on rules, and the automatic classification of knowledge is realized, and precise knowledge and imprecise knowledge are obtained.Finally, the knowledge acquisition system supported by semantic Web rough ontology is designed and realized. A maritime accident semantic Web rough ontology of maritime domain is constructed, and it is applied to the maritime accident knowledge acquisition. The rationality and validity of knowledge acquisition method supported by semantic Web rough ontology is verified through the total of acquired knowledge, recall ratio and precision ratio. And the results show that compared to the knowledge acquisition method supported by semantic Web ontology, this method can acquire not only precise knowledge but also imprecise knowledge, and the recall ratio is improved obviously.