Process Modeling, Behavior Analytics and Group Performance Assessment of e-Learning Logs Via Fuzzy Miner Algorithm

The major objective of this study is to bring together a variety of topics such as process mining, e-learning, and educational data mining to discuss the opportunities of applying event modeling and process management techniques to e-learning systems. Accordingly, this work aims to identify and analyze behavioral patterns that affect and influence the quality of students' academic performance within/during an undergraduate computer programming course. Process mining Fuzzy Miner techniques enabled us to discover a set of patterns describing how the performance of groups following a particular e-learning program, differs from the performance of another group, as well as their time-related performance differences. The results showed that groups with high grades spent much more time (i.e., 47.4 days in average) to accomplish the e-learning activity, while groups with low grades spent less time to accomplish the same e-learning activity. Moreover, groups with high grades performed more actions creating more events throughout the course, almost double, compared with the groups with low grades. Accordingly, the students with high grades put much more effort by retaking and retrying the course quiz multiple times.