The clever combination of these indicators (which are generated by a given user and automatically recorded by Moodle) seems to be able to predict if a student is going to pass the exam or not; they
developed a 3-layer ANN, an SVM with linear kernel (with best predictive index) and an SVM with RBF kernel.
Importance of the indicator for the students’ learning methodology:
Interaction patterns are meant both with the learning material and also between users (with number of viewed discussion forums), under the assumptions that a high degree of interaction between users is expected to give a higher mark, and the more resources they saw, the more likely they were to pass the exam. The distribution of the accesses (with total number of events and number of course access per week/month) relies in the fact that a student following a constant and steady path in learning with regularly distributed accesses will be more likely to have a positive outcome for the exam rather than a student who connects two days before the exam.
G. Biondi, V. Franzoni, A. Mancinelli, A. Milani. Student behaviour models for a university LMS. In O. Gervasi, B. Murgante, S. Misra, A. Rocha, C. Garau (eds.), Computational Science and Its Applications--Proceedings of the ICCSA 2022 Workshops, LNCS 13379, ISBN 9783031105449 (July 2022), pp. 33-43. http://link.springer.com/chapter/10.1007/978-3-031-10545-6_3