Use this URL to cite or link to this record in EThOS:
Title: Knowledge Discovery in Enterprise Architecture: An Ontology-driven Approach
Author: El Kourdi, Mohamed
ISNI:       0000 0001 3443 1691
Awarding Body: Staffordshire University
Current Institution: Staffordshire University
Date of Award: 2008
Availability of Full Text:
Access from EThOS:
With the divergent information within the underlying enterprise architecture (EA) repositories, automated techniques for supporting enterprise architecture planning (EAP) activities has become a crucial task in order to support enterprise architects. Pragmatically, manual approaches for supporting impact and similarity analysis, which are the main EAP activities, are error-prone, time consuming, and cumbersome. This is due to the fact that enterprise architects need to drill down different architectural models in the EA repository databases in order to assess the impact of change and the similarity between distinct models. Consequently, knowledge discovery techniques are considered to be imperative in terms of supporting impact and similarity analysis. In this research, a knowledge management model has been conceptualized in order to identify the knowledge resources that are needed to support enterprise architects in aligning business and IT resources. The knowledge resources that are needed to support impact and similarity analysis have been addressed from a technical perspective using an agent based approach for knowledge discovery in enterprise architect~re. Such an agent employs an embedded ontology as a knowledge base in order to serve as a vocabulary through which the agent and enterprise architects will communicate without making assumptions about each others language. The inference engine of the agent consists of two ontologically driven knowledge discovery mechanisms, which are association rules mining and hierarchical agglomerative clustering (HAC). Association rules mining is a means to derive the hidden patterns and the relationships between different EA components to support: impact analysis in EA. HAC, on the other hand, is mainly concerned with clustering similar EA models in an EA repository database in order to support EA similarity analysis. Prior to the application of association rules mining and HAC techniques to EA repository databases, the latter are typically pre-processed, indexed, and converted to tabular structures that are ready for quantitative analysis. The association rules and HAC techniques have been implemented and evaluated based on empirical experiments on an illustrative example repository database, which is a part of the ARIS (architecture for integrated information systems) toolset, and represents the different architectural models of a real world automobile company.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available