Development of an Ontology for an Industrial Domain

Introduction

The vast amount of information on the World Wide Web has made it increasingly difficult to access and retrieve the required information or data. In response to the problem, the World Wide Web Consortium (W3C) formally proposed the Semantic Web to be the next evolutionary step for the Web in 2001. The Semantic Web aims to attach semantic information to Web resources and would provide the semantic structure or scaffolding that would enable autonomous software agents to traverse the Web in search of or process information on behalf of users or other systems.

Within this context, ontologies are important from at least two perspectives. First, they provide a key component of the Semantic Web because an ontology can function as a repository of vocabulary of unambiguous domain-related concepts and their meanings anchored in consensus domain knowledge (Jasper et al. 1999). This semantic structure would enable autonomous agents to access and process information on the Web. An example is the use of ontologies to support negotiation in e-commerce (Tamma et al. 2005). Second, ontologies provide sharable knowledge models on particular problem domains for construction of knowledge-based systems in a distributed and open environment such as the Internet. Therefore, from the knowledge engineering perspective, an ontology constitutes a crucial building block for developing knowledge-based systems.

Knowledge engineering is the process of eliciting expertise, organizing it into a computational structure, and representing it in a knowledge-based system. The process of knowledge engineering can be viewed from the cognitive informatics perspective (Wang et al, 2002/06; Patel et al., 2003; Chan et al., 2004; Kinsner et al., 2005; Yao et al., 2006; Wang and Kinsner, 2006; Wang, 2002/03/06/07), with a focus on the problem solving expertise in cognition (Chan 2002). The effort spent in engineering knowledge is often substantial due to the tacit nature of expertise; and the process of acquiring knowledge for building the knowledge base is widely recognized as a major bottleneck in the development process. If several knowledge-based systems on the same problem domain are to be constructed, the effort required to build the knowledge bases for the different systems is often duplicated. A possible solution for the problem is to share any knowledge on a given problem domain that has been acquired among systems. Four different approaches for sharing knowledge have been adopted within the Knowledge Sharing Effort sponsored by Air Force Office of Scientific Research, Defense Advanced Research Projects Agency, the Corporation for National Research Initiatives, and the National Science Foundation (Neches et al. 1998). Similar to their objectives, the work presented here aims to construct ontologies which can overcome the barriers to sharing that arise due to lack of consensus across knowledge bases on vocabulary and semantic interpretations in domain models. A critical step in the process of developing an ontology is performing a detailed analysis of the domain. In this chapter, we present a method for knowledge acquisition and ontology construction to support the development of a knowledge-based system, and we demonstrate application of the method to the domain of natural gas pipeline operations. The proposed method involves first eliciting and organizing knowledge using the Inferential Modelling Technique (IMT), and then—based on the initial classification of knowledge elements in the problem domain— constructing an application ontology using an automated knowledge modeling tool called the Knowledge Modeling System. The knowledge represented in the application ontology provides the basis for implementing the advisory system.