Improving E-learning by integrating a metacognitive agent

Abstract

The major disadvantage of the current Learning Management Systems is the lack of learner assistance in their learning processes and, therefore, they can not replace the presence of the teacher who ensures the progress of learning. In fact, we proposed to integrate, for each learner, a metacognitive agent that supported the metacognitive assistance and extracts the defectsin the learning process and strategies. The goal is to invite the learner to correct himself and improve his learning method. Metacognitive questionnaires were distributed to a group of 100 students before, during and after a computer course. The goal is to evaluatethe metacognitive attributes and to determine their influence on the success of learning. Decision trees were used as data analysis tools to extract a set of rules and to discover the influence of these metacognitive attributes on the result obtained by the learners. The results indicate that there are relationships between the different metacognitive attributes and the learners' success. We note there is the influence of metacognitive incitement on learner outcomes, which reflects the degree of understanding of a learning pedagogical unit by the learner.