Modelo de predicción del rendimiento académico para el curso de nivelación de la Escuela Politécnica Nacional a partir de un modelo de aprendizaje supervisado automatizado en R

Abstract

The purpose of this project is to describe a statistical methodology based on decision trees and logistic regression where the learning problem is formulated in terms of the error function minimization using the gradient descent method, turning the learning problem into an optimization problem. The considered variables for the methodology are socioeconomic, demographic, home-related, institutional and of academic performance in the admission exam and in the preparatory course, to predict the student’s probability to approve the nivelation course in a time window prior to the end of the semester. Additionally, the algorithm is implemented in the statistical software R, which is responsible for automatically performing each of the steps of the described methodology.