Mining Students' Interaction Data from a System that Support Learning by Reflection

In this paper we utilising some popular educational data mining (EDM) methods to explore and mine educational data resulted from a system that supports reflection for learning called Reflect. Our objective is to study whether using the Reflect helps students learn better in the subject. First, we explore data by means of statistical analyses. Then, we used clustering and classification techniques to identify learning behaviours associated with positive or negative outcomes. The results suggest that there is a positive correlation between reflection and student performance characterised by the level of activities carried out in the Reflect and the exam mark. We claim that our approach has resulted in the identification of some behaviour that leads to the increase in the students’ performances.