Predicting students’ dropout at university using Artificial Neural Networks

This study is part of an ongoing project investigating the first stage of the process of student transition to university. This paper aims to contribute to the continuing debates on the possibilities how to reduce the student failure and improve educational processes with the help of data mining techniques, in particular of the artificial neural networks. The population consists of 810 students enrolled for the first time in a health care professions degree course at the University of Genoa in the academic year 2008-09. The research is based on the analysis of data and information originating from primary sources: administrative data related to the careers of students; statistical data collected during the research through an ad hoc survey; data derived from telephone interviews with students who had not completed the enrolment in the subsequent years. The neural network correctly predicted 84 percent of the cases of group 1,81 percent of the cases of group 2, and 76 percent of the cases belonging to group of dropouts. The application of the artificial neural network model can offer a valid tool to design educational interventions to deliver to those who score high in the level of risk.