Predictive model for physical performance in athletics: Correlation between anthropometric data and cardiorespiratory capacity in students from a private school

• Autores: Ovalle, Christian, Sánchez Puche, Everardo, Ortiz Gomez, Genesis Andrea, Cornejo Vega, Jairo Samir
• Formato: artículo
• Fecha de Publicación: 2024
• Institución: Universidad Tecnológica del Perú
• Repositorio: UTP-Institucional
• Lenguaje: inglés
• OAI Identifier: oai:repositorio.utp.edu.pe:20.500.12867/14643
• Enlace del recurso: https://hdl.handle.net/20.500.12867/14643; https://doi.org/10.3991/ijoe.v20i15.52857
• Nivel de acceso: acceso abierto
• Materia: Predictive model; Cardiorespiratory capacity; Machine learning; Anthropometric data; https://purl.org/pe-repo/ocde/ford#2.02.04

In the current educational context, physical education and student sports development face challenges marked by continuous technological evolution. This study proposes a predictive model supported by machine learning and artificial intelligence (AI), establishing a connection between cardiorespiratory capacity (VO2max) and student anthropometric data. With a sample of 179 students aged 13 to 18, the model-building process included preparing and partitioning a dataset, training, and evaluation under the CRISP-DM methodology. A multiple linear regression model was applied, incorporating weight, age, height, sex, and body mass index (BMI) to analyze their relationship with the dependent variable (VO2max). Performance metrics revealed a significant correlation between anthropometric measurements and cardiorespiratory fitness (CRF), with a 24% improvement in training, although test accuracy was -0.8%. Including additional variables, such as sex and age, they have improved the predictive equations. However, the ability of the model to predict VO2max was limited, suggesting the complexity of the relationship between these factors. In a comprehensive evaluation, five linear regression models achieved a correlation accuracy of 22% with the complete data set.