Abstract: | A truly Intelligent Tutoring System has to be capable of reasoning dynamically and effectively regarding the student's knowledge in order to provide individualised tutoring. Up to now, however, few intelligent tutors can modify a simulation in order to customise their curriculum to the student's needs. To achieve the latter objective, a student model is required to perform the task of recording student attitude and action and, based on these data, to outline the student's learning and cognitive profile. The student modelling methods currently being used, have proved to be simplistic and static. Therefore, researchers have attempted other approaches such as numerical methods for uncertainty management and machine learning techniques. This dissertation focuses on a particular numerical method, Fuzzy Logic, and tests whether a fuzzy logic approach can enhance the role of student modelling by handling effectively uncertain or imprecise information to make reliable inferences about student's current state of understanding. For this purpose, we have built an alternative student model and named it IRIS, in an attempt to grasp a clearer view of the student modelling field.
|