Knowledge modelling includes inserting and altering domain knowledge, but it also concerns changing the reasoning strategy. The reasoning strategy, already implemented in the system, is often a non-transparent and complex process, which makes it difficult to control and alter. Individual rules tend to be relatively simple and self-documented, but the logical interactions in large sets of rules may be opaque. To control and change the reasoning strategy, the relationships between different rules and facts need to be clearly presented through a conceptual model.
This paper suggests visualisation as a support with which to follow the reasoning strategy more easily and to change it in the knowledge-based system, or more particular rule-based systems. This visualisation includes both static and dynamic presentation of rules and facts in the knowledge base. The static presentation visualises the actual contents of the knowledge base without being affected by any external values. The dynamic presentation, in contrast, visualises the rules used and the facts relevant to a specific consultation, i.e., the presentation depends on the inputs inserted by the users. Since it concerns strategies, the presentation needs to be illustrated with a stepwise execution. A candidate for this modelling language is the visual modelling language Unified Modelling Language (UML), particularly since UML has become a standard for working with software-intensive systems. For static presentation, we use the UML’s sequence diagrams and for dynamic presentation, we use UML’s collaboration diagrams. To simplify the otherwise complex presentation of the rules, concepts are encountered to capture semantic notions at different levels of abstraction and are applied to the rules and facts in the different diagrams.