Research Of Representation And Process Of Semantic Information Of Web Data Based On Ontology

Taking the key techniques of semantic Web as research background, this thesis makes a focal research on the methods of domain knowledge modeling, the semantic information representation method of Web data based on ontology, the implementation technologies of XML data storage and dissemination, and Web personalized information services based on domain ontology. The primary contributions of this thesis include:With respect to the requirements of domain knowledge modeling on the Web, the author puts forwards a new method for domain ontology modeling, named Domain Knowledge Ontology for Web (DKOW). Based on the foundational theory of Ontology, DKOW integrates the methods of semantic network, object-oriented modeling and semantic data modeling. DKOW is capable of modeling and reasoning static domain knowledge, including domain structures, concepts and properties of the concepts, and the relations among concepts, etc.In order to apply the power of domain ontology to express the semantic information of Web data, the author presents an algorithm named DTD Generating From Domain Ontology(DTDGDO) that maps a given ontology to a Document Type Definition (DTD) based on the transformation rules. The transformation rules include rearrange rule, transition rule and cardinality rule. The algorithm makes it possible that the computer can efficiently process the self-describing information of semi-structured data by mapping between model of semi-structured data and model described by Description Logic.The author applies model-driven method to design and implement the transaction middleware between XML and relational database, named XML-Relational Database Middleware(XRDBM). XRDBM automatically generates the mapping model from the initial DTD and then stores the XML documents into relational database. XRDBM improves XML-DBMS at following aspects: the consistency maintaining method during transforming semi-strctured data model into relation model, the maintenance of XML document and the problem of different subelements with the same name, etc.Based on XRDBM, the author proposes an algorithm of XML publishing to deal with the input SQL statements and generate valid XML document for the certain DTD automatically, named Valid XML Generating from SQL(VXG -SQL). Contrasting with other analogical algorithms, this algorithm doesn't require artificially generating XML segment model beforehand so as to reduce the workload of XML publishing.In order to provide personal information services effectively, the author proposes a personal user model named User Model Based on Ontology(UMOnto), which is composed of Personal Domain Ontology(PDO) and Personal InformationRetrievalS(PIRS). The key problem in this model is how to map domain ontology to personal domain ontology. To handle this problem, an algorithm of how to model UMOnto dynamically during user browsing the Web and retrieving information is proposed. This model has been implemented in Personal Assistant Based On Domain Ontology(PADO), which provides personal browse, retrieval, navigation and recommendation in Personal Information Services Experiment System(PISES). PISES is an experiment entironment for the semantic information representation, storage and management of Web data, and provides personalized information services.The achievements of the thesis have great theoretic and realistic significance in deepening the methodology of domain knowledge modeling, expression and management, which promotes the development of new Web application and advances the research of Semantic Web.