Ontology-Based Information Interoperability In Heterogeneous Resources

Information interoperability is an extensive topic which attracts a broad range oftheories, methodologies and techniques. It had being an active research area for nearlytwenty years and recent interests are concerned on semantic interoperability amongindependent resources. Because of features of heterogeneous, autonomous anddistributed of independent web systems, their interoperation is still difficult.Our approaches focus on a particular problem, conceptual heterogeneity, which occursin diversified representations about the same real world object. We founded ourapproaches on ontology, a logical theory account for the intension meaning of aformal vocabulary or language, and assume that multiple ontologies could help usfind more agreements among different conceptualizations.First, we establish a describing model for ontological engineering of informationresources. This model combines representation of objects-level, classes-level andcontexts-level in an integrated framework in order to increase the agreed area amongmultiple conceptualizations.Second, we propose an Active Mediator Mode for ontology based resources toachieve interoperability. The key component of this mode is the semantic resolutionability which supported by intelligent information processing methods and algorithms.This effect of mode requires the co-effort of information provider, mediator andinformation requester, so it is also called closely coupled mode.Third, we propose a Loosely Coupled mode for web based information interoperation.This mode is featured by the components of XML Schema, a web document form ofontology and wrapper, a XML parser powered extraction program. The finalproduction of this mode is an extracted and resolved schema, which is be consideredas a machine understandable Schema for heterogeneous XML data.Fourth, we present a supposition of common vocabulary retrieval algorithm based onrefined ontology. This supposition aims for automatic generation of commonvocabularies, especially thesauruses, among structured semantic spaces. We consider ontological descriptions as semantic spaces and try to find appropriate algorithm to cluster terminologies into different groups according to the relationships among them. In our opinion, ontological description explicitly answers the question of K-NN means. Further comparison is grounded on the basis of Tversky Similarity measure and Matrix computations.Finally, we introduce the software component and structure of a prototype tool, ODIMAC TOOLS, which is developed for ontological description development.