Automatic diagram generation based on ontological framework

A diagrammatic representation is an intuitive way to present domain knowledge and findings. Automated conversion of a domain description to a diagrammatic representation should be correct and expressive, hence a challenging research. However, the representations used for descriptive and diagrammatic formats are inconsistent with one another due to heterogeneity in their perspectives. Thus lack of semantic interoperability is a major hurdle in bridging the representational gap. This thesis proposes an ontological framework called OntoDiagram that converts domain descriptions to diagrams. OntoDiagram facilitates the interpretation of descriptive terms and the extraction of implicit domain knowledge. It then infers a diagrammatic representation that illustrates features in a consistent fashion. OntoDiagram is designed to be suitable for different domains and across multiple levels of abstraction. A pilot study of prototype system developed, for use in pediatric cardiology, shows that the approach is feasible and efficient in generating diagrams for domain descriptions.