Abstract: | This article presents knowledge modelling in an intelligent tutoring system, in which the modelling approach allows the knowledge representation to somewhat depart from its actual or most obvious basis to fit specifically pedagogical needs. The knowledge domain is cost engineering. Although the formulae used in cost engineering problem-solving are usually considered as relations between variables, our approach makes use of these formulae to identify new concepts, the factors, of a more pedagogical nature. The introduction of these concepts should help the student to concentrate either on computations or on economic analysis, and the system to make more accurate and more useful tutoring interventions. Moreover, it leads to important qualitative aspects of the domain modelling and of the tutorial interactions that should also help the learning process. Finally, this concept creation brings to light a hierarchy of intermediate abstraction levels, which can then be used to derive an order of presentation of these concepts, an order of the corresponding prerequisites and an order of the exercise types. All these should facilitate the student's learning. |