A Hybrid Machine Learning Approach for the Prediction of Grades in Computer Engineering Students

The growing application of information and communication technologies (ICTs) in teaching and learning processes has generated an overload of valuable information for all those involved in education field. Historical information from students' academic records has become a valuable source of data that has been used for different purposes. Unfortunately, a high percentage of research has been developed from the perspective and the need of teachers and educational administrators. This perspective has left the student in the background. This paper proposes the application of a hybrid machine learning approach, with the aim of laying the groundwork for a future implementation of a recommendation system that allows students to make decisions related to their learning process. The work has been executed on the historical academic information of students of computer engineering degree. The results obtained in this article show the effectiveness of applying a hybrid machine learning approach. This architecture is composed of, on the one hand, techniques of supervised learning applied with the objective of classifying the data in clusters, and on the other hand, having this initial classification, unsupervised learning techniques applied with the objective of carrying out a predictive analysis of students' historical grade records.