Predicting noncontact injuries of professional football players using machine learning

Abstract

Noncontact injuries are prevalent among professional football players. Yet, most research on this topic is retrospective, focusing solely on statistical correlations between Global Positioning System (GPS) metrics and injury occurrence, overlooking the multifactorial nature of injuries. This study introduces an automated injury identification and prediction approach using machine learning, leveraging GPS data and player-specific parameters. A sample of 34 male professional players from a Portuguese first-division team was analyzed, combining GPS data from Catapult receivers with descriptive variables for machine learning models-Support Vector Machines (SVMs), Feedforward Neural Networks (FNNs), and Adaptive Boosting (AdaBoost)-to predict injuries. These models, particularly the SVMs with cost-sensitive learning, showed high accuracy in detecting injury events, achieving a sensitivity of 71.43%, specificity of 74.19%, and overall accuracy of 74.22%. Key predictive factors included the player's position, session type, player load, velocity and acceleration. The developed models are notable for their balanced sensitivity and specificity, efficiency without extensive manual data collection, and capability to predict injuries for short time frames. These advancements will aid coaching staff in identifying high-risk players, optimizing team performance, and reducing rehabilitation costs.

Fig 1. Machine learning pipeline used in the study for injury detection.

The process begins with data collection, followed by data preparation. Subsequent stages involve the development and validation of predictive models to ensure accuracy and reliability in injury detection.

Fig 2. Comparison of cost-sensitive and traditional learning approaches across 500 simulations.

This figure depicts the comparison of models’ average GMEAN values from 500 simulations, contrasting cost-sensitive learning with traditional learning methods for both train and test data splits. It also illustrates the impact of adding more features on the GMEAN scores of classifiers. The shaded areas represent the standard deviation, highlighting the variability within each model configuration.

Fig 3. Radar plots of model performance metrics across 500 simulations.

The radar plots presented here compare the quality and stability of various models and learning types over 500 simulations, based on the best combinations of model, learning type, and feature count as identified in Table 1. Stability is assessed through standard deviations depicted as lines on top of the bars, while quality is measured in terms of GMEAN values, sensitivity (Sen), specificity (Spe), and accuracy (Acc) for both training and testing data splits.