Using Bayesian Inference and Neural Network Classification for Student Modeling in a Tutoring System

R. Nkambou and J.M.P. Tchétagni (Canada)

Keywords

knowledge representation, student modeling, hierarchical evaluation, diagnosis, fuzzy logic, Bayesian inference.

Abstract

This paper is about the management of student knowledge evaluation in an intelligent tutoring system in order to support diagnosis process. The evaluation consists in assessing the student’s state of knowledge at several levels of granularity. The domain knowledge is distributed among five levels. Each level is characterized by a particular paradigm of knowledge representation depending of the type of knowledge involved. The evaluation in upper levels is carried out by the propagation of the evaluation from lower level using Bayesian inferences and neural networks classifications.

Important Links:



Go Back