Smart Higher Education: Implementation of a Predictive Analytics Model for the Early Identification of Student Dropout Risk at a Technological Higher Education Institution.

Abstract

Student dropout represents one of the major challenges for higher education institutions specializing in technology. The objective of this research was to design and evaluate a predictive analytics model capable of timely identification of students at risk of dropping out, using academic variables recorded on institutional platforms. A quantitative, correlational study was conducted using a simulated dataset of 185 students.

Decision Tree, Random Forest, and Logistic Regression algorithms were compared to estimate dropout risk. The simulated results showed that Random Forest achieved the best predictive performance, with an accuracy of 93.4%, enabling the early identification of students with a higher probability of dropping out. It is concluded that incorporating predictive models into university digital ecosystems can strengthen institutional decision-making and support preventive academic support strategies.