A learner model based on multi-entity Bayesian networks and artificial intelligence in adaptive hypermedia educational systems

The aim of this paper is to present a probabilistic and dynamic learner model in adaptive hypermedia educational systems based on multi-entity Bayesian networks (MEBN) and artificial intelligence. There are several methods and models for modelling the learner in adaptive hypermedia educational systems, but they're based on the initial profile of the learner created in his entry into the learning situation. They do not handle the uncertainty in the dynamic modelling of the learner based on the actions of the learner. The main hypothesis of this paper is the management of the learner model based on MEBN and artificial intelligence, taking into accounts the different action that the learner could take during his/her whole learning path. In this paper, the use of the notion of fragments and MEBN theory (MTheory) to lead to a Bayesian multi-entity network has been proposed. The use of this Bayesian method can handle the whole course of a learner as well as all of its shares in an adaptive educational hypermedia. The approach that we followed during this paper is marked initially by modelling the learner model in three levels: we started with the conceptual level of modelling with the unified modelling language, followed by the model based on Bayesian networks to be able to achieve probabilistic modelling in the three phases of learner modelling.