| As the model of semantic description and organization for information and knowledge, ontology is the key technique to solve semantic information sharing and interoperability. Especially in the network with increasingly distributed and heterogeneous Web resources, it is a focus research to provide users information services intelligently by ontology in the Semantic Web. While the semantic information processing strategy plays an increasingly important role in the new information service techniques of Semantic Web Service, Cloud Computing and the Internet of Things, ontology modeling as the basis for semantic information processing and application has becoming particularly significant. Therefore, it is an important issue with theoretical and practical value to study on ontology modeling methods and application strategies for semantic information processing.The ambiguous and user preference information of semantic meaning brings up great challenges for ontology modeling. The existing ontology models cannot handle uncertain and ambiguous semantic information。To provide intelligent information services according to the ambiguity and personal preferences of user semantic information, existed modeling method of ontology and fuzzy ontology are deeply analyzed in this thesis, and fuzzy theory is applied to build semantic model of fuzzy top ontology and fuzzy domain ontology respectively, and ontology description language and tool are used to describe fuzzy semantic information directly. By using the built fuzzy ontology, semantic information processing mechanisms of semantic reasoning, semantic information mapping and semantic information matching are designed in the thesis. The content and contributions are as following:(1) Semantic modeling method of fuzzy top ontology based on fuzzy theoryTo handle uncertain and ambiguous semantic information in ontology, a modeling method of fuzzy top ontology based on fuzzy theory is proposed, and the semantic model is created to describe fuzzy membership function and fuzzy modifier concepts. By the semantic model of fuzzy membership functions, fuzzy concepts and fuzzy relations are described by defined typical fuzzy membership functions. As the same, by the semantic model of fuzzy modifier concepts, fuzzy semantic information is described by defined fuzzy modifiers. The results of semantic modeling and application in the ontology tools of Protege show that the fuzzy top ontology can not only describe fuzzy concepts and fuzzy relations in ontology, but also provides a general ontology semantic model for sharing by other fuzzy domain ontology and fuzzy task ontology.(2) Semantic modeling method of fuzzy domain ontology based on5W2HThere is not a general and perfect modeling method of fuzzy domain ontology, so a semantic modeling method of fuzzy domain ontology based on5W2H is presented in this thesis. Ontology concepts in a certain domain are defined from the seven aspects of5W2H, such as Who, When, Where, What, Why, How and How Much, and then semantic relations between those concepts are defined. In which, the fuzzy concepts in the ontology are within the aspect of How Much, and those fuzzy concepts and fuzzy relations are described by the defined typical fuzzy membership functions in fuzzy top ontology. The fuzzy domain ontology of expert information domain is constructed by this semantic modeling method, and the application shows that the ontology structure is clear and reusable. The comparison with the modeling method of object-oriented method and hierarchical method shows that the modeling method based on5W2H is more significant in the capabilities of aided analysis, fuzzy knowledge modeling and scalability.(3) Semantic mapping model and semantic matching strategy based on fuzzy ontologyTo deal with vagueness and preference in the user semantic information, a semantic mapping model and a semantic matching strategy are proposed based on fuzzy ontology. According to the user demand of the semantic information and personal preference, as well as the application scene and restrictions, the two levels of semantic reasoning mechanism is presented. Combined with semantic reasoning of ontology description logic and fuzzy reasoning of fuzzy logic, reasoning rules are defined and realized, such as filtering reasoning rules,1W1H reasoning rules, expansible reasoning rules and fuzzy reasoning rules. Semantic mapping models and semantic matching strategies of three kinds of user information queries are designed, such as the precise query, the fuzzy query and preference query. Through semantic reasoning, semantic mapping and semantic matching, query services with high satisfaction are provided to the users. Practical application results in the Expert Information Service System shows that the semantic-based information query has better performance of recall and precision than keyword-based information query. Meantime, ranking service of the query results is provided for user by the calculation of query matching value using semantic matching strategy, and it is helpful for users to select experts by the ranking results.
|
PDF Full Text Request