A multimodal approach to psychological resilience prediction in football players: Integrating biomechanical analysis, physiological feedback, and machine learning

Football is a high-intensity sport that demands not only technical and physical excellence but also strong psychological resilience. This study investigates the relationship between biomechanics, physiological feedback, and psychological health in football players, employing a hybrid predictive model that integrates autoregressive analysis and XGBoost. A multimodal dataset comprising biomechanical indicators (postural stability, muscle activation, reaction time), physiological markers (heart rate variability [HRV], electrodermal activity [EDA]HRV, EDA, respiratory rate), and behavioral responses (decision volatility, self-reported stress levels) was collected from professional and semi-professional football players over a six-month period. The results demonstrate that neuromuscular stability and cognitive efficiency significantly influence psychological resilience, with postural control and reaction time emerging as key predictors of anxiety levels. The hybrid ARIMA-XGBoost model achieved superior predictive accuracy (R2 = 0.89, RMSE = 0.61), outperforming traditional machine learning models. These findings highlight the practical value of integrating biomechanical monitoring with psychological assessments for personalized stress management and performance optimization in competitive sports.