Research Of E-Government Resources Management System Based On Ontology

With quickly development of information technology, mass information invariety industries explodes. Huge information resources provide people abundantinformation, at the same time, people is submersed in the sea of information. It is hardfor people to choose or manage so much information. And the information has variousformats. In traditional information management mode, it is very hard for a user toretrieve useful information related to his requirement, because there isn't unifiedsemantic description for these network information resources. It is problem ininformation management systems to enable machine/computer to understand thesemantic information of resources and to give facilities for users to retrieve relatedinformation.Semantic Web is the representation of the development direction of the newgeneration technology of World Wide Web. Its target is to solve the great plenty ofproblems emerging in current WWW applications and to realize the resources sharingand application cooperation. The developing Semantic Web technology provides anew path to resolve the problems in traditional information resources management.Ontology is a formal, explicit specification of a shared conceptualization, andrepresents domain knowledge in an explicit and formal mode. It is useful to enhanceknowledge sharing and the cooperation feature between heterogeneous systems. OWLis the W3C recommendation Web Ontology Language and is developed based onDescription Logic. And it can describe ontology knowledge and reasons efficiently.For the situation of E-Government information resources management, it isresearched in this paper to setup a framework of resources management based ondomain knowledge and to build E-Government Ontology based on OWL andcorresponding developing process and method. In the framework, it is researchedincluding semantic distance and relatedness based on domain ontology,ontologyconcepts discrimination and semantic retrieval based on semantic relatedness of ontology concepts. At last, as an experiment platform, a prototype system isdeveloped in which an applied mode of knowledge ontology based on J2EEframework is proposed.Resources description and reasoning are the key technology to realize resourcessharing and application cooperation. Knowledge ontology is a useful method torepresent the great plenty of nonstandard knowledge in internet with concepts, and itcan greatly boost the knowledge sharing between applications. E-Governmentknowledge ontology based on OWL adopts the standard semantic descriptiontechnology recommended by W3C, and represents the domain knowledge by themode that can be read and understood by machine. It is the precondition to realizeresources sharing and applications cooperation in E-Government domain.EGO is an extendable ontology according to E-Government domain. EGOrepresents domain knowledge by setting up concepts,inheritance of concepts,relationship of instances and constrains. EGO can be used in many resources'semantic descriptions in network. EGO combines the reasoning ability of descriptionlogic to expose implicit knowledge and various relations between ontology concepts,and provides semantic information that can be understood by machine. EGO resolvesthe problem that knowledge in expert system can't be reused in network applications,and realizes highly sharing of domain knowledge.Computing semantic distance and relatedness between concepts is an importtechnology in EGO application. Semantic distance and relatedness can resolve theproblem that the relationship in traditional knowledge representation can't berepresented in quantity mode. The algorithm of computing semantic relatedness takesin consideration the density extension of concepts in hierarchy,the concept's depthand the relationship weight between concepts, and can represent the relatedness ofontology concepts. The semantic relatedness is the base to improve the precision ofretrieval and realize reasoning for the resource management framework.Concepts discrimination for user-keyword and user keywords combination canefficiently catch semantic context of user keywords, and represents the user'sintention. When several words are specified as user keywords, the semantic context between them can be decided and the corresponding meaning of a word can be fixedon, too. Ontology concept discrimination proposes a kind of concepts discriminationmethod based on semantic distance to improve the precision of retrieval. At first, asynonymous extension of ontology concepts is made. When a user specified akeyword, he may key in a synonym of the ontology concept. In this condition,synonymous extension is useful to improve the probability to match a suitable conceptfor one keyword. Secondly, this concept discrimination method not only considers thematching extent between user keyword and concept, but also considers therelationship between candidate concepts to represent the implicit semantic contextamong these keywords. So, this algorithm has a certain good performance.The semantic retrieval based on ontology relatedness transfers semanticreasoning in retrieving to database query, and sufficiently uses the advantage ofdatabase's robust and fast retrieving engine. Semantic annotation is the footstone ofsemantic retrieving. In the process of annotation, according to W3C recommendedAnnotea Protocols, an annotation file with RDF/XML format is stored in annotationdatabase alone. Then, the annotation body is extracted from annotation file, and it isparsed as ontology concepts and instances to be saved in database. Combing semanticrelatedness in ontology concepts, these ontology concepts and instances stored indatabase can be used to retrieve semantic information efficiently.As the experiment platform in this research, a prototype is developed as theresources management platform based on semantic distance of ontology concept.Through semantic description to information resources, this prototype can help user tomake use of more related resources as shared knowledge and services.