Modelo de pronóstico en la educación superior en pruebas específicas de ingeniería, licenciatura, y pensamiento científico matemático en Saber Pro aplicando Machine Learning

Abstract

By means of obtaining the Icfes Databases with regard to the Saber Pro exams and the specific exams Saber Pro of Engineering, teaching and Mathematical Scientific thought from the years 2017 to 2019, to make a forecasting model with Machine Learning, using the algorithms K Nearest Neighbors, K-Means, Naive Bayes and Neural Network TensorFlow Keras, of which 2 of 4 algorithms obtained optimal results (Naive Bayes and K Nearest Neighbors), to verify what has been the behavior of the students when they take the Saber Pro exam and finished university, so that it can be evidenced that the variables: family members, mother's studies, place of residence, domestic equipment, reading at home, number of siblings and scores, affect educational achievement in the quality of university education.