Domain Model
The domain model is a representation of the subject material to be taught. Generally, this model represents some process or system in a computationally interpretable form. This is a defining characteristic of “symbolic AI”; such models are variously described as qualitative, relational, and/or semantic. Clancey [7] presented evidence that all expert systems were “model-based.” In effect, a fundamental contribution of expert systems research has been to extend modeling formalisms beyond mathematical constructs in conventional science and engineering to include qualitative/relational models and associated modeling operations [10].
By design, an ITS domain model can be interpreted to solve problems presented to the student. This model constitutes the knowledge to be learned: facts, formalisms, causal processes, and reasoning processes (e.g., a diagnostic strategy). Typically, the program that applies the domain model to solve problems is called the inference engine(e.g., a rule interpreter), which applied to particular circumstances (e.g., a patient’s history and symptoms), produces a (partial) solution called a situation-specific model .
At first, most ITS researchers viewed the domain model as being isomorphic to stored structures and subconscious processes in the brain. This assumption motivated much productive research in which the ITS creates a model of the student’s knowledge and reasoning. In recent decades, models have been viewed more often as scientific toolsthat are constructed and applied to understand and manipulate processes and systems in the world [11, 21], exemplified by the student model in an ITS.