Domain Specific Structural Analysis for Intelligent Tutoring Systems: Automatable Representation of Declarative, Procedural and Model-Based Knowledge with Relationships to Software Engineering
Joseph M. Scandura
ITS development has been idiosyncratic largely due to an inability to represent expert knowledge (content) as internally consistent hierarchies. This article shows how to construct consistent behavioral (specification) and design (knowledge) hierarchies that are correct with respect to their specifications. Any given concept (or operation) can be represented hierarchically in a manner that insures each level of representation is behaviorally consistent with every other level. Each level is represented explicitly in terms of exactly one of three kinds of refinement: component, category and operation. The introduction of abstract (equivalence classes of) values provides necessary precision while avoiding intractable mathematical problems. Combinations of these refinement types are adequate for representing procedural as well as declarative knowledge and both domain specific and domain independent knowledge. Interacting systems (models), for example, are shown to involve both declarative and procedural knowledge. In short, the process of structural analysis (SA) is postulated to provide a sufficient basis for reducing any given concept or process, no matter how complex, to arbitrarily simple elements lending themselves to automation.