Issues in Structured Knowledge Representation A Definitional Approach with Application to Case-Based Reasoning and Medical Informatics

Doktorsavhandling, 2003

Several issues concerned with structured knowledge representation based on definitional structures are discussed: the realisation of knowledge-based systems using declarative programming, similarity assessment in knowledge representation and case-based reasoning, and the importance of human-computer interaction and information visualisation in knowledge-based systems. To illustrate the basic ideas, real-world applications from the area of oral medicine are used. We present MedView, a project aiming at developing models, methods and tools to support clinicians within oral medicine in their everyday work and research. The definitional model used in MedView constitutes the main governing principle in the project, not only in the formalisation of clinical terms and concepts, but in visualisation models and in the design and implementation of individual tools and the system as a whole as well. In the context of case-based reasoning, a novel similarity framework based on definitional structures is presented. In this framework, a uniform definitional model is used for both case representation, similarity assessment and adaptation. The similarity model is general enough to capture many different types of similarity measures, e.g., ordinal, cardinal and asymmetric measures. We also describe how a uniform declarative model can be used as the basis for an interaction model and two information visualisation tools in the area of oral medicine. We describe how known visualisation techniques to a large extent can be modeled using the conceptual model of the underlying data, enabling close interaction with the user and tight coupling between two different visualisation tools.