Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.489250
Title: An information flow frame work of representation and its application to semantic integration
Author: Wang, Yang
ISNI:       0000 0000 5497 897X
Awarding Body: University of the West of Scotland
Current Institution: University of the West of Scotland
Date of Award: 2008
Availability of Full Text:
Access through EThOS:
Abstract:
In information systems, a fundamental notion is representation. And yet, it would seem that in the literature of information systems, this notion has been treated intuitively at least informally, and in particular, how a possible representational relationship between two sets of objects may be systematically identified, formalized and reasoned about has not been fully addressed. Such a relationship should be formalizable because information flow channels underpin representational relationships between systems. A good understanding of such a relationship would help advance our knowledge regarding information systems. To this end, this thesis builds an information flow based representation framework whereby to capture, formalize, analyze and reason about representation for distributed systems. This is based on semantic information theories, which not only lay solid philosophical foundations, but also contribute to forming a mathematical foundation for the framework. Taking regularities as a fundamental principle, an information flow channel is constructed by taking both token and type levels into consideration. We introduce the notion of representation link, which connects separate contexts through type level constraints and token level connections. The framework enables an analysis of information flow within a distributed system both qualitatively and quantitatively. The channel hierarchy and link hierarchy take into account background information, which enhances inference on representation links. The framework allows agents to choose a channel that is the most appropriate for their purposes without violating the objectivity of information flow. Representation links also facilitate the identification of the information bearing capability of a system in terms of what and how information flows. This framework is then applied to the semantic integration of systems. A semantic integration model is thus built. With this model, the process of semantic integration is examined by using the mathematical mechanism of the framework. Important issues in semantic integration are investigated including how initial semantic correspondences affect integration results, how good and flexible an integration process is, and to what extent systems that are involved in integration may be changed and yet the result remains the same and thus reusable. The validity of the framework is also verified by analyzing results of integration from both content and quantity perspectives. This thesis shows that information flow based inference is applicable to analysing results of integration, and also channel and representation link hierarchy help us reason over different but relevant integrations. All these show the value of the proposed framework for information systems.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.489250  DOI: Not available
Share: