| The current Web has reached a new milestone called the "Semantic Web", which is characterized by two fundamental techniques: XML (extensible Markup Language) and RDF (Resource Description Framework). XML, as the seed technology within the progress of the Web, is the universal applicable declarative syntax in the future Web, and RDF providing a data model for resource description and a set of primitives at semantic level is the technical foundation of semantic Web. Some matured techniques in KE (Knowledge Engineering) and software engineering, especially ontological techniques and OO (Object-Oriented) techniques, have played important roles in pushing the current Web towards the semantic one. Even today, people are still facing the problems how to elegantly merge and customize these traditional techniques into the Web computing mechanism. In addition, there are other important issues to be addressed, such as how to represent formal knowledge and to markup logic formula for querying and inference in semantic Web, and how to improve retrieving efficiency and accuracy of querying via ontology learning, and etc.As the Web is thought as a popular distributed platform, into which more and more traditional applications are been transitioned, there are other two promising fields for applying and further developing concepts and methods of KE in semantic Web worth paying attention to two fields: III (Intelligent Information Integration) and KM (Knowledge Management) in the context of semantic Web. However, the connotation of these techniques in semantic Web might differ from traditional ones (e.g. KM in Artificial Intelligence), consequently, there is a need for reviewing and ratification to allow them to be deployed in semantic Web.This dissertation is devoted to some key issues of semantic Web and to KMS (Knowledge Management System) in the context of semantic Web. The author's contributions of presented in this thesis include:Defining the Ontology Semantics for Description Primitives of Logical Formula and Relational AxiomsMarkup languages are employed to describe the objects and resources in a Web environment. In order to represent knowledge, logical formula, relational axioms and rules, there is a need for primitives specifying prepositional formula, predictive formula and Horn clause, and relations between properties and relational axiomx. However, existing ontology languages are either unsuitable for Web with markuplanguages, or short of of generic support to representation of knowledge, logical formula, axioms and rules.The author has made extensions on logical and relational primitives based on some influential techniques, like RDFS (Resource Description Framework Schema) and OIL (Ontology Inference/Interchange Layer), so as to enable knowledge, rules and axioms to be syntactically seriliazed with ordinary XML. Such extension has also enhanced the compatability to a great extent among applications on the semantic Web platform, for example, between an RDF parser and an XML parser.Defining Ontology Semantics for User Querying/Inference Primitives and Enveloping Query/Inference Formula in SOAPQuery Languages for current Web either lack of structural characteristics, or are restricted to queries of syntactic structure level of XML documents, while in the semantic Web environment, a mechanism for semantic query is essential. The author provides a solution to this problem by defining request/response primitives for query/inference, based on the defined primitives for logic formula and relational axioms and the key inferring technologies in KE inference engines. The query/inference primitives are enveloped into SOAP (Simple Object Access Protocol) messages and bound to HTTP (Hyper-Text Transfer Protocol) transfer protocol. All these works result in a querying system structure, suitable for ontology/RDF data and possessing simple inference function.Proposing an RDF-Annotation-based Algorithm for Ontology LearningTo improve query efficiency and semantic accuracy, semantic Web needs to learn...
|
PDF Full Text Request