Mining of E-learning Behavior using SOM Clustering

Learning Management System, such as Moodle, has been utilized extensively as part of e-learning implementation for higher institutions. The flexibility of LMS to convey the learning materials in many ways and approaches enable the instructor to implement blended learning. The student's interaction and activities while learning are captured by Moodle in the log data file and are useful to identify the way student learn in e-learning. This study investigates Kohonen SOM clustering performance in order to analyse the students' e-learning usage and to identify the cluster of student's learning characteristics. The data being analysed is captured from Moodle log file of students taking Data Structure and Algorithm subject at Faculty of Computing, Universiti Teknologi Malaysia. The experiment shows that SOM is able to produce good clustering group compared to other clustering techniques. Two clusters of students were identified which are low browsing and high browsing groups of students.