Abstract

BAUTISTA CANON, Elmer; QUIRAMA SALAMANCA, Jenny E.  and  BAUTISTA CANON, Edilfonso. Predictive model of the learning progress of uniminuto students applying machine learning techniques. Conrado [online]. 2021, vol.17, n.83, pp. 305-310.  Epub Dec 10, 2021. ISSN 1990-8644.

The purpose of this research is to validate the hypothesis where it is proposed to explain the relationship of the learning progress of UNIMINUTO students in the exit tests, with respect to the variables of the entrance tests to higher education, based on the development of a supervised learning predictive model. For the development of the project, a methodology with 4 phases was established, descriptive analysis of the data for 10,505 instances with 69 variables, predictive analysis, prescriptive analysis and applications of the model. As a result, three predictive models were developed with logistic regression algorithms, support vector machines and neural networks, which showed an efficiency close to 75% and the evaluation of the precision metrics, recall and f1, with similar percentages of efficiency. In conclusion, it was possible to develop 3 predictive models of the learning progress of UNIMINUTO students, based on the information provided by ICFES, which relates the input variables with the output variables of higher education in Colombia.

Keywords : Supervised learning; predictive model; logistic regression; support vector machines; neural networks; value added education.

        · abstract in Spanish     · text in Spanish     · Spanish ( pdf )