Current Proceedings on Technology

Yazarlar: Mouenis Anouar Tadlaoui, Mohamed Khaldi, Souhaib AAmmou

Konular:-

Anahtar Kelimeler:Bayesians networks,Learner modelling,Cognitive diagnosis.

Özet: Bayesian networks tend to be increasingly used for the management of uncertainty in modelling of the learner. They have been successfully used in many systems, with different objectives. However their use as part of the cognitive process modelling raises a number of problems. On the one hand the underlying semantics of arcs is often not clearly explained, and on the other hand the evolution of the knowledge structure is not taken into account. We will present here a detailed presentation of Bayesian networks and their use in modelling of the learner; in particular we focus our attention on the question of the orientation of the arcs. We will briefly present what could be a learner model consisting of several Bayesian networks. As a first step, we will return to the problem of management of uncertainty and Bayesian networks, and we will focus on some of the problems and difficulties encountered in the use of Bayesian network in the modelling of the learner. We will then demonstrate the need to consider multiple networks in a single model. Finally, we will present the formal structure which allows this consideration. The experiments presented in this work are arguments in favour of our hypothesis on the link between the level of expertise of the learner and the structure of Bayesian network that models.

---