Towards Methodological Principles for Ontology Engineering.

Abstract

The Internet and other open connectivity environments create a strong demand for the sharing of data semantics. Emerging ontologies are increasingly becoming essential for computer science applications. Organizations are looking towards them as vital machine-processable semantics for many application areas. An ontology in general, is an agreed understanding (i.e. semantics) of a certain domain, axiomatized and represented formally as logical theory in a computer resource. By sharing an ontology, autonomous and distributed applications can meaningfully communicate to exchange data and make transactions interoperate independently of their internal technologies.

The main goal of this thesis is to present methodological principles for ontology engineering to guide ontology builders towards building ontologies that are both highly reusable and usable, easier to build, and smoother to maintain.

First, we investigate three foundational challenges in ontology engineering (namely, ontology reusability, ontology application-independence, and ontology evolution). Based on these challenges, we derive six ontology-engineering requirements. Fulfilling these requirements is the goal and motivation of our methodological principles.

Second, we present two methodological principles for ontology engineering: 1) ontology double articulation, and 2) ontology modularization. The double articulation principle suggests that an ontology be built as separate domain axiomatizations and application axiomatizations. While a domain axiomatization focuses on the characterization of the intended meaning (i.e. intended models) of a vocabulary at the domain level, application axiomatizations mainly focus on the usability of this vocabulary according to certain application/usability perspectives. An application axiomatization is intended to specify the legal models (a subset of the intended models) of the application(s)' interest. The modularization principle suggests that application axiomatizations be built in a modular manner. Axiomatizations should be developed as a set of small modules and later composed to form, and be used as, one modular axiomatization. We define a composition operator for automatic module composition. It combines all axioms introduced in the composed modules.

Third, to illustrate the implementation of our methodological principles, we develop a conceptual markup language called ORM-ML, an ontology engineering tool prototype called DogmaModeler and a customer complaint ontology that serves as a real-life case study.

This research is a contribution to the DOGMA research project, which is a research framework for modeling, engineering, and deploying ontologies. In addition, we find we have benefited enormously from our participation in several European projects. It was through the CCFORM project (discussed extensively in chapter 7) that we were able to test and debug many ideas that resulted in this thesis. The Network of Excellence KnowledgeWeb has also proved to be a fruitful brainstorming environment that has undoubtedly improved the quality of the analyses performed and the results obtained.

Keywords: Ontology Engineering, Lexical Semantics, DOGMA, Object Role Modeling, ORM, Ontology, Formal ontology engineering, Lexical semantics, Gloss, WordNet, Ontologies and WordNet, DogmaModeler, Semantic Interoperability, Emergent Semantics, Semantics in Distributed Systems, Ontology, Object Role Modeling,