An application of the model logistic regression in the prediction of the student performance

Abstract

This work presents a methodology to carry out predictions of the success in the first semester of the students of the Common Plan Engineering career, in a student cohort of first year of the University of Antofagasta.

To carry out the analyses were considered three criteria of demand for classify as successful to a student in the first semester of their career. Applying a model of logistic regression, the results were compared with the of the discriminant analysis method, analyzing besides their agreement and index of predactibility.