Contributions to Adaptive Educational Hypermedia Systems via on-line Learning Style Estimation

In order to establish an online diagnostic system for Learning Style Estimation that contributes to the adaptation of learning objects, we propose an easily applicable expert system founded on Bayesian Networks. The proposed system makes use of Learning Style theories and associated diagnostic techniques, simultaneously avoiding certain error sources. It takes under consideration learners’ answers to a specific questionnaire, as well as a classification of learners who have been previously examined. System application based on Kolb’s Learning Style Inventory is also presented, in order to study the changes to the final estimation made by the system. The questionnaire in use is the David Kolb’s Learning Style Inventory revised, an on-line form which avoids logical fault implications. With this aim, a case study has been carried out involving 885 university students, from which conclusions are drawn. As a result of the expert system use, the influence of factors, such as reflection of cross-cultural environment differences and/or of slippery or lucky answers’ influence on learning style estimation is reduced.