Some Approaches in Learning Style Diagnosis

Adaptive Educational Hypermedia Systems among the emerging technologies serves Knowledge and Learning Systems. Aiming to improve pervasiveness and efficiency of asynchronous e-learning we give emphasis to the development of diagnostic tools capable to recognize certain learner’s characteristics to the purpose of providing learning material tailored to the learner’s specific needs. Learning Style diagnosis can be approached either by the use of probabilistic expert systems or by the use of fuzzy expert systems. In order to establish a probabilistic Learning Style’s diagnostic expert system we propose a method based on Bayesian Networks. Our system analyzes information supplied by the system’s descendant users, as well as providing a stochastic analysis of the problem. It is a system that utilizes what appear to be reasoning capabilities so as to reach Learning Style’s estimations. In order to classify learners in a predefined set of classes, the proposed system takes under consideration learners ’ answers to a certain questionnaire, as well as a classification of learners who have been previously examined. As it was expected, our method takes advantage of previously accumulated knowledge and proves to be more accurate than Learning Style’s direct estimation, i.e. an estimation based only on user responses. Further to this, we propose a Fuzzy Cognitive Maps technique, which modifies the Learning Style into appropriate quantitative characteristics. The importance of Fuzzy Cognitive Map’s technique becomes even greater since this method takes into consideration the previously gained experience on learners ’ style diagnoses