engleski

Discovering patterns of student behaviour in e- learning environment

The benefits of e-learning have been widely recognized in today’s education. However, behaviour of students attending a course through an e-learning platform and its connection to students’ satisfaction with a course is still not investigated enough. This paper analyses a course log data in e-learning environment in addition to some students’ descriptive variables at University of Osijek, and aims to discover patterns in students’ behaviour that could enable to create profiles of satisfied and unsatisfied students. The final purpose is to reveal knowledge about student behaviour that will assist academic teachers in increasing the level of their students’ satisfaction. The methodology used in the research includes several data mining methods, such as statistical tests of dependence, and support vector machines. The results show that satisfied students put more effort in frequent viewing of all course materials, they are more active in uploading assignments, and have more previously earned ECTS points than unsatisfied students. The extracted characteristics could be used to improve student satisfaction with the course by stimulating those activities. In order to generalize results, the research is to be extended to include more e-courses on different levels of academic education.

student behaviour, e-learning, data mining, classification trees, support vector machines

Indeksirano u bazi podataka ERIC.