Predictive Model for Estimating Internal Transfer of Informatics Engineering Students

Student dropout is a significant problem affecting higher education institutions. This phenomenon is the result of multiple causes and has different forms the aim of this work was to predict which students of the Departamento de Ingenieria Informatica, Universidad de Santiago de Chile, will migrate to a different programme. For this purpose, we considered logistic regression models and support vector machines (SVM) with linear and radial kernels. Results showed that radial kernel SVM can satisfactorily predict this phenomenon, with an accuracy of 88.0% for the 6-year programme and of 66.67% for the 4-year programme. © 2020 IEEE.