Electronic Thesis and Dissertation Repository


Doctor of Philosophy


Electrical and Computer Engineering


Hamada H. Ghenniwa, Weiming Shen


In an open computing environment, such as the World Wide Web or an enterprise Intranet, various information systems are expected to work together to support information exchange, processing, and integration. However, information systems are usually built by different people, at different times, to fulfil different requirements and goals. Consequently, in the absence of an architectural framework for information integration geared toward semantic integration, there are widely varying viewpoints and assumptions regarding what is essentially the same subject. Therefore, communication among the components supporting various applications is not possible without at least some translation. This problem, however, is much more than a simple agreement on tags or mappings between roughly equivalent sets of tags in related standards. Industry-wide initiatives and academic studies have shown that complex representation issues can arise. To deal with these issues, a deep understanding and appropriate treatment of semantic integration is needed. Ontology is an important and widely accepted approach for semantic integration. However, usually there are no explicit ontologies with information systems. Rather, the associated semantics are implied within the supporting information model. It reflects a specific view of the conceptualization that is implicitly defining an ontological view. This research proposes to adopt ontological views to facilitate semantic integration for information systems in open environments. It proposes a theoretical foundation of ontological views, practical assumptions, and related solutions for research issues. The proposed solutions mainly focus on three aspects: the architecture of a semantic integration enabled environment, ontological view modeling and representation, and semantic equivalence relationship discovery. The solutions are applied to the collaborative intelligence project for the collaborative promotion / advertisement domain. Various quality aspects of the solutions are evaluated and future directions of the research are discussed.