Evaluation of behavioral changes for deduction the learning profiles

The e-learning systems have been of particular interest in recent years, research in this area is highly evolved to best support face to face learning systems. However, even if the experiments have demonstrated many advantages, limitations primarily related to significant dropout rates still persist. Indeed, this is due to several reasons including the lack of support and the feeling of isolation that the learner may have. Our paper proposes a solution to address this problem by providing appropriate support for student according to his learning style to increase their motivation and fight their feelings of isolation. Several solutions have been proposed to support learners in their learning process, ranging from suggestions on the association of working groups to analyzing facial expressions in order to deduce learners' emotions. In this paper, we suggest a support system allowing to provide learners with personalized assistance expressions in order to support them throughout their learning and that deduces their learning profiles by analyzing their interactions outcomes. This deduction is performed by adapting the algorithm classification ANTClust that will allow us (1) to deduce learners' learning profiles and (2) to track the evolution in their behavioral changes in order to infer their exact profiles.