Application of machine learning to the learning analytics of the Moodle platform to create heterogeneous groups in on-line courses

In university courses to promote collaborative activities among students, on-line learn-ing environments such as e-learning platforms are used. Effective collaborative activitiesinvolve the creation of heterogeneous groups of 4 or 5 students. In the university contextthe formation of groups is difficult due to the high number of students. Groups are oftenunbalanced and not very functional if chosen randomly. Some e-learning platforms, suchas Moodle, lack an intelligent mechanism that allows the automatic creation of heterogeneous groups of students. We applied clustering algorithms on Moodle learning analytics (LA) that allowed to build groupings that identify the different characteristics ofstudents based on their behaviors kept on the platform. Therefore we have developedan intelligent numerical tool which, using clusters obtained from Machine Learning onthe LA, generates heterogeneous groups. These groups are made available on the platform for the teacher. The project will conclude with the development of a Moodle pluginto automate the exchange of data and information between the Machine Learning algorithm and the Moodle platform.