Big Data as a Tool for Analyzing Academic Performance in Education

Abstract

Educational processes are constantly evolving and need upgrading according to the needs of the students. Every day an immense amount of data is generated that could be used to understand children’s behavior. This research proposes using three machine learning algorithms to evaluate academic performance. After debugging and organizing the information, the respective analysis is carried out. Data from eight academic cycles (2014–2021) of an elementary school are used to train the models. The algorithms used were Random Trees, Logistic Regression, and Support Vector Machines, with an accuracy of 93.48%, 96.86%, and 97.1%, respectively. This last algorithm was used to predict the grades of a new group of students, highlighting that most students will have acceptable grades and none with a grade lower than 7/10. Thus, it can be corroborated that the daily stored data of an elementary school is sufficient to predict the academic performance of its students using computational algorithms.