Modeling the learner within an ILE: Attribute scoring and unsupervised classification of the profile

The observation and the monitoring of the learner is a fundamental problem for interactive learning environments (ILE). The ILE needs to provide tools to replace traditional means and the analysis of the learner's interactions with his learning environment must fill this gap through digital traces. Thus, we propose in our research, an approach of identification, analysis and interpretation of the traces allowing all the actors of the ILE (including the learner) to improve the experience of the learner, throughout his learning cycle. This approach is based on the implementation of attributes associated with the learner to which we will assign scores, as well as iterative clustering algorithms for modeling the learner profile. In addition, it is based on the adaptation of the course according to this profile in order to provide course recommendations. We will use the data sources offered by our MOOC "Back to English" (Ibn Tofail University, Kenitra, Morocco) to illustrate our modeling. In this article, we will focus only on the learner modeling steps using scoring techniques applied to collected traces as well as the grouping of learners via unsupervised clustering algorithms.