Modeling students’ behavior using sequential patterns to predict their performance

Abstract

Online learning environments generate educational data that can be used to model students’ behavior and predict their performance. In online learning environments, in which students are free to choose their next activity, various factors such as time spent on individual tasks and the choice of next learning material may impact students’ performance. The main goal of this research is to enhance student learning by modeling students’ behavior and testing whether these behavioral patterns correlate with their performance. Using sequential pattern mining methods, we will identify the most frequent patterns in students’ online learning activities and test whether/which patterns correlate with higher or lower performance. By identifying which student behavioral patterns correlate with higher or lower performance, this study has the potential to inform redesign of online learning platforms and study guidelines that help students learn more and perform better.