Utilizing Machine Learning Techniques for Learning Analytics: A Case Study of Moodle LMS Activity Log Analysis
Abstract
Learning analytics collects data, analyzes, and interprets the learning process that has taken place. The output of this method can be used to improve the quality of teaching or learning. Moodle is a popular learning management system (LMS) used for online learning. Various learning activities carried out by students are recorded in the activity log. This paper shows the potential of using machine learning methods to analyze activity logs taken from Moodle LMS. The sample used in this study refers to implementing the Digital Society course, which students from different fields of science attend. This paper describes using supervised and unsupervised learning on activity log data taken from the Moodle LMS. The variables used as datasets include the frequency of activity reading pdf material, scores, videos, forums, quizzes, and graduation status. The supervised learning model that was built succeeded in obtaining an accuracy of 100% in the application of logistic regression and Naïve Bayes Classification. Unsupervised learning clustered all the data and showed the cluster related to the frequency of online learning activities and students' assessment success status.