UBC Theses and Dissertations
Using ontologies in the context of knowledge management systems Bera, Palash
Knowledge management systems (KMS) are information systems that combine and integrate functions for managing knowledge in organizations. Although substantial interest exists in KMS, a theoretically-based view of knowledge in the KMS context is not yet available. To clarify the notion of knowledge as managed by KMS, a conceptual framework is developed. The key concepts of this framework are derived by combining an action-based perspective with an artificial intelligence (AI) view of knowledge. The relationships among these concepts are identified, anchored to current literature, and represented graphically as conceptual models. Conceptual models are used to support the understanding of and communicating about application domains. The models contribute in proposing several theoretical and practical implications regarding KMS. To use KMS effectively, knowledge seekers need to be able to identify the knowledge required to perform their tasks. It is suggested that providing knowledge seekers with a visual representation of a formal ontology can facilitate performing knowledge identification. Formal sets of statements defining the relevant concepts and their relationships are called formal ontologies. Formal ontologies are often specified in ontological languages such as Web Ontology Language (OWL). The main requirements from such languages are that they have well-formalized syntax and that they will be computer-readable. However, not much attention has been paid to how they can be used to convey domain semantics. It is suggested that the use of philosophically-based ontological principles can help generate guidelines for developing conceptual models using OWL. Accordingly, a set of guidelines is proposed and it is demonstrated that application of such guidelines can provide clearer representation of domain phenomena such as interaction. Ontologies developed with these guidelines for modeling interaction are termed informed ontologies. From the developed conceptual models for KMS, it is identified that knowledge is intimately tied to the change of state of an entity. This change of state is facilitated by entities participating in interactions. Thus, it is proposed that the use of informed ontologies will lead to better knowledge identification than the use of uninformed ontologies. In a laboratory study, using business students as subjects, it was found that the use of informed ontologies for knowledge identification was advantageous.
Item Citations and Data