Predicting Academic Performance Path Using Classification Algorithms

Abstract

The objective of this research is to determine the academic performance route of students entering the Systems Engineering program. The academic performance route is defined by three courses, which develop sequentially in the first semesters, where students show difficulty to be approved. The population is represented by 827 students, the research was approached from a quantitative approach, the research design is non-experimental and the scope or level of research is correlational. The methodology implemented is CRISP-DM (Cross Industry Standard Process for Data Mining) using machine learning algorithms, through binary classification models using logistic regression algorithms, random forests and XGboost. The results have allowed predicting whether a student would pass or fail in each of the courses, determining their academic performance path. The classification models have been able to achieve an accuracy between 87% and 93%.