Date of Completion

8-20-2013

Embargo Period

8-20-2013

Keywords

Software Engineering, Software Modeling, Ontology Design, Ontology Modeling, Ontology Patterns

Major Advisor

Steven A. Demurjian

Associate Advisor

Dong-Guk Shin

Associate Advisor

Xiaoyan Wang

Field of Study

Computer Science and Engineering

Degree

Doctor of Philosophy

Open Access

Campus Access

Abstract

Ontologies have gained increasing usage to augment an application with semantic knowledge, particularly in biomedical informatics, where ontologies are employed to represent knowledge ranging from bioinformatics data to clinical informatics. Over the past decade, the federal government and medical organizations have pushed for the adoption of health information technology (HIT) in the form of an Electronic Health Record (EHR), personal health records (PHRs), patient portals, pharmacy systems, etc. The ontologies used in the current medical systems are development using an instance and construction based approach, creating ontologies for a specific purposes unique to the system, e.g., different for each HIT. As a result, for the same medical knowledge, every HIT can have a differently structured ontology in a HIT specific manner, resulting in an inability to easily integrate these ontologies, thereby hindering the HIE process. To address this deficiency, the work presented in this dissertation focuses on the creation of an ontology modeling, design, and development that emphasizes software engineering and its process, moving ontologies away from instance-based to modeling based approach. Towards this objective, first, the work explores different approaches (UML, ERD, XML, and OWL) in terms of their modeling capabilities; this assisted in identifying features of the more conceptual design models that may be suitable for inclusion into an ontology design process. Using this as a basis, the second part of this work extends the OWL with features that include Attribute to allow the definition of a class similar to UML, Domain Profile to capture abstract entities, and, Schema Associations to represent user-defined relationships among ontology models. The third part of our work takes our proposed extensions and casts them as software engineering artifacts to allow them to be placed in a software development process (SDP) that constructs ontologies in a more structured and organized workflow. Referred to as the Hybrid Ontology Design & Development Model with Lifecycle, HOD2MLC proposes an agile software process for ontology design and development. To illustrate our research in depth, we leverage a realistic medical scenario that is broadly defined to demonstrate the utility of our OWL extensions and the usage of the HOD2MLC in a realistic setting.

COinS