Resumen

RICO PAEZ, Andrés. Progressive Predictive Models of the Academic Performance of University Students. RIDE. Rev. Iberoam. Investig. Desarro. Educ [online]. 2022, vol.12, n.24, e044.  Epub 25-Jul-2022. ISSN 2007-7467.  https://doi.org/10.23913/ride.v12i24.1196.

The objective of this research was to develop progressive predictive models of the academic performance of university students in Mexico and evaluate them for different machine learning techniques. In this study, grades of academic activities of 260 university students were collected to create prediction models of academic results using machine learning techniques. The models were built at different stages throughout the course and were evaluated using the accuracy of the predictions by applying it to the prediction of 112 students in a subsequent course. An accuracy of up to 70.5 % was observed in a time of 21 % of the total duration of the course. This type of methodology can be replicated for different types of courses because the recording of grades is common in almost all courses. In addition, this methodology is flexible in terms of choosing the time stage in which to make the predictions, maintaining a compromise between the accuracy of the predictions and that they be made at the earliest possible stage to detect problems with academic performance, avoiding, in as far as possible, the failure and desertion of students.

Palabras llave : machine learning; mathematical model; prevention; academic performance.

        · resumen en Español | Portugués     · texto en Español     · Español ( pdf )