Adaptive Exploration of Assessment Results under Uncertainty

In the context of Intelligent Learning Environments (ILE), adaptivity plays a key role. In order to achieve adaptive behavior an ILE should have a rich representation of the learning context, which is defined, among others, by the learner’s characteristics, the type of the educational material, the advisory history, etc. Actually, the user model used by the system, and especially the representation and maintenance of user’s knowledge can be considered as one of the critical factors that affect the system’s effectiveness, in terms of it’s capability to adapt to the individual learner’s needs. In general, the evaluation of user knowledge derives from tests and tasks that the system proposes to the user to accomplish. This paper describes an approach that refines assessment results through user knowledge exploration, incorporating probabilities. We argue that the proposed approach leads to a better mapping of the assessment results to user knowledge in terms of its adaptivity to the response style of each individual learner.