Árbol de clasificación para la identificación de síntomas asociados a la depresión en estudiantes de una universidad pública

• Published in: Retos (Madrid) 2024 (52), p.104-114
• Main Authors: Gonzalez Bustamante, David Alejandro, Ortiz González, Kevin, Sandoval, Karol Michelle, Amú Ruiz, Francisco Antonio
• Format: Article
• Language: Spanish
• ISSN: 1579-1726; 1988-2041; 1988-2041
• DOI: 10.47197/RETOS.V52.100138

Introduction: Depression, a common disorder affecting 5% of the global population, has seen an increase due to the aftermath of the COVID-19 pandemic. University students, especially impacted by pandemic-related confinement, experienced disruptions in their emotional well-being, academic performance, and social interactions. Since classification trees have proven effective in predicting these events in various population groups, the objective of this study was to create a tool for identifying depressive symptoms based on classification trees through machine learning. Methodology: 680 university students were assessed before and after participating in a technology-mediated physical exercise program during the second semester of 2020. The Beck's Depression Inventory (BDI-II) and FANTÁSTICO healthy lifestyle questionnaires were implemented, as well as the Functional Movement Screening (FMS) tool. Program adherence was measured through recorded attendance. A classification tree model was developed to predict depression four months after the initial assessment using the statistical software package R. Results: The classification tree model identified that the initial Beck-II test score, low mood, engagement in exercise, and the initial Fantástico checklist score are the most successful predictive variables. The model exhibited sensitivity of 0.83, specificity of 0.72, and accuracy of 0.72. Conclusions: This model can be valuable for anticipating depressive behavior in university students or for monitoring depressive symptomatology. Introducción: la depresión, un trastorno común que afecta al 5% de la población mundial, ha experimentado un aumento debido a los efectos posteriores a la pandemia de COVID-19. Los estudiantes universitarios, especialmente afectados por el confinamiento durante la pandemia, experimentaron afectaciones en su emocionalidad, rendimiento académico e interacción social. Ya que los árboles de clasificación se han destacado en la predicción de estos eventos en diversos grupos poblacionales, el objetivo de este estudio fue crear una herramienta de identificación de síntomas depresivos basada en árboles de clasificación mediante machine learning. Metodología: se evaluaron 680 estudiantes universitarios antes y después de participar en un programa de ejercicio físico mediado por tecnología durante el segundo semestre de 2020. Se utilizaron los cuestionarios de Depresión de Beck-II y FANTÁSTICO para estilos de vida saludables, así como la herra