Academic Failure: identifying “high risk” students by data mining methods

Carrion, Fabian [FUCAM] Meskens, Nadine [UCL] Vandamme, Jean-Philippe [FUCAM]

Our study is mainly aiming to classify students within two groups: the 'low risk' students (LRS), who have a high probability of succeeding, and the ‘risk’ students (RS), who will probably have to face problems and may succeed thanks to the measures taken by the university. This ‘two groups’ classification has even less interest as the academic year advances, it is not helpful to wait until April or even February to guide students who have a real need for support. Via questionnaires, we collected for every student, in a number of Belgian and French universities and during several years, a range of characteristics such as their age, their parents’ level of education, their perception of the university environment, etc. Having this collected, with sufficient and wide-ranging information’s, our objective will be to establish a statistical model which will make it possible to predict academic success as early as possible in the academic year. A final aim is thus to provide a process allowing an optimal distribution of teaching resources in order to curb academic failure.