Learning Style Classification by Using Bayesian Networks Based on the Index of Learning Style

Abstract

In this paper, we propose a classification model constructed with an algorithm based on Object-Oriented Bayesian networks (OOBN) to determine the learning style of a student. For this, the Felder-Silverman Learning Style Model (FSLSM) is used, which is based on the Index of Learning Style (ILS) questionnaire. The idea is to use the answers to the questionnaire as provided by a student - as the input of the OOBN model to classify the learning style of this student. The classifications made by the OOBN model are validated with the full questionnaire with 44 questions as well as a short version with only 20 questions. The results of the OOBN classification with 44 individually answered questions represent the ground truth to compare the classifier performance in case a reduced set of questions is used. The OOBN with 20 questions suggest that the approach is classifying the students into the correct learning style dimensions in most cases. This indicates a possible to use BN within an Adaptive Learning System (ALS) like HASKI.