Variables influencing university dropout: A machine learning-based study

Abstract

University dropout is a growing problem having considerable academic, social and economic consequences. It may depend on several factors, such as for example the knowledge area. In previous works we studied dropout and transfer paths using machine learning, obtaining several key factors that are predictive for analyzing drop out and transfer paths patterns. In this work we delve into this topic, making a more exhaustive study using again machine learning. Results show that Polynomial SVM is the method that obtains the highest performance for predicting university dropout. On the other hand, it is possible to identify the key factors affecting university dropout, showing in addition different factors depending on the field of study.