Towards Intelligent Adaptative E-Learning Systems – Machine Learning for Learner Activity Classification

Abstract

As adaptivity in e-learning systems has become popular during the past years, new challenges and potentials have emerged in the field of adaptive systems. Adaptation, traditionally focused on the personalization of content, is now also required for learner communication and cooperation. With the increasing complexity of adaptation tasks, the need for automated processing of usage data, information extraction and pattern detection grows. We present learner activity mining and classification as a basis for adaptation in educational systems and discuss intelligent techniques in this context. Based on real usage data, we present the results of experiments comparing the behaviour and performance of different classification algorithms.