Deep learning approach for predicting university dropout: A case study at roma tre university

Based on current trends in graduation rates, 39% of today young adults on average across OECD countries are expected to complete tertiary-type A (university level) education during their lifetime. In 2017, an average of 10.6% of young people (aged 1824) in the EU-28 were early leavers from education and training. Therefore the level of dropout in the scenery of European education is one of the major issue to be faced in a near future. The main aim of the research is to predict, as early as possible, which student will dropout in the Higher Education (HE) context. The accurate knowledge of this information would allow one to effectively carry out targeted actions in order to limit the incidence of the phenomenon. The recent breakthrough on Neural Networks with the use of Convolutional Neural Networks (CNN) architectures has become disruptive in AI. By stacking together tens or hundreds of convolutional neural layers, a “deep” network structure is obtained, which has been proved very effective in producing high accuracy models. In this research the administrative data of about 6000 students enrolled from 2009 in the Department of Education at Roma Tre University had been used to train a Convolutional Neural Network based. Then, the trained network provides a predictive model that predicts whether the student will dropout. Furthermore, we compared the results obtained using deep learning models to the ones using Bayesian networks. The accuracy of the obtained deep learning models ranged from 67.1% for the first-year students up to 94.3% for the third-year students.