Smartlearn : predicting learning performance and discovering smart learning strategies in flipped classroom

Abstract

In flipped classroom, students are expected to learn new contents in online learning system before attending offline classes to reinforce their knowledge. This online and offline blended education model has become more and more popular. However, spending more time to actively engage in online learning does not result in better learning performance, so that how to wisely arrange online learning plan is a big challenge. In this paper, we build a LASSO model to accurately predict students' performance in course projects and their final grade by online learning behaviour data in flipped classroom. The LASSO selected features show that learning online between first and second flipped classes after midnight, and during the second flipped class would benefit students' project performance but studying one day before the examination and studying at night is counterproductive. Our results provide novel insight into guiding students to learn wisely and perform better in flipped classroom.