Statistical Learning for the Prediction of School Dropouts

Abstract

Students leaving school without a basic qualification face numerous disadvantages compared to their graduated peers. Therefore, Educational systems in the Netherlands strive to give all students at least a basic qualification. They try to achieve this, among other things, by dropout prevention programs. Although the number of dropouts in the Netherlands is reduced, there are still a large number of dropouts. A significant number of them drop out unexpectedly and, therefore, without intervention of the dropout prevention programs. In this research project, we explore the use of data mining techniques to help identifying the students at risk of dropping out. We will show that data mining has great potential to help schools in this task.