Authoring of probabilistic sequencing in adaptive hypermedia with Bayesian networks

Abstract

One of the difficulties that self-directed learners face on their learning process is choosing the right learning resources. One of the goals of adaptive educational systems is helping students in finding the best set of learning resources for them. Adaptive systems try to infer the students' characteristics and store them in a user model whose information is used to drive the adaptation. However, the information that can be acquired is always limited and partial. In this paper, the use of Bayesian networks is proposed as a possible solution to adapt the sequence of activities to students. There are two research questions that are answered in this paper: whether Bayesian networks can be used to adaptively sequence learning material, and whether such an approach permits the reuse of learning units created for other systems. A positive answer to both question is complemented with a case study that illustrates the details of the process. © J.UCS.