Combining Internal- and External-Training-Loads to Predict Non-Contact Injuries in Soccer

Abstract

The large amount of features recorded from GPS and inertial sensors (external load) and well-being questionnaires (internal load) can be used together in a multi-dimensional non-linear machine learning based model for a better prediction of non-contact injuries. In this study we put forward the main hypothesis that the use of such models would be able to inform better about injury risks by considering the evolution of both internal and external loads over two horizons (one week and one month). Predictive models were trained with data collected by both GPS and subjective questionnaires and injury data from 40 elite male soccer players over one season. Various classification machine-learning algorithms that performed best on external and internal loads features were compared using standard performance metrics such as accuracy, precision, recall and the area under the receiver operator characteristic curve. In particular, tree-based algorithms based on non-linear models with an important interpretation aspect were privileged as they can help to understand internal and external load features impact on injury risk. For 1-week injury prediction, internal load features data were more accurate than external load features while for 1-month injury prediction, the best performances of classifiers were reached by combining internal and external load features.

Who Will Score? A Machine Learning Approach to Supporting Football Team Building and Transfers

“…It contained the data that were collected from 46 professional rugby league players throughout the 2015 NRL season, including some data collected by GPS (Global Positioning System). Similar approaches were presented in [ 13 , 14 ]. In these articles, the authors proposed their systems for predicting the possibility of the injury of the football players.…”

A Data Mining Approach to Predict Non-Contact Injuries in Young Soccer Players

“…Although other algorithms such as k-nearest neighbors (KNN) or extreme gradient boosting (XGBoost) showed a good predictive ability in previous studies (Rommers et al, 2020;Vallance, Sutton-Charani, Imoussaten, Montmain, & Perrey, 2020), they are often considered as blackbox models. The reason of this fact is linked to their difficulties of interpretation and their inability to provide information regarding the interaction between the different features (Cortez & Embrechts, 2013).…”

“…The predicted maturity offset, defined as the years before or after PHV (Mirwald et al, 2002), is a useful non-invasive somatic indicator that predicts the time during which the athletes will experience their adolescent growth spurt (Malina, Bouchard, & Bar-Or, 2004). Particularly, the 6 months before and after PHV (maturity offset ranging from -0.5 to +0.5) have been identified as a critical period for the onset of injuries in young soccer players (Bult, Barendrecht, & Tak, 2018;van der Sluis et al, 2014). During this time, known also as the period of 'adolescent awkwardness' (Philippaerts et al, 2006), young athletes experience a decline in performance and motor control.…”

Analysis of Running Performance in the Offensive and Defensive Phases of the Game: Is It Associated with the Team Achievement in the UEFA Champions League?

“…In soccer, running performance (RP) has been extensively studied over the last two decades [1,2]. Detailed knowledge about this performance in match play is essential for the design and implementation of specific fitness training [3][4][5][6]. Previous studies have demonstrated that players regularly transit between brief bouts of high-intensity running and longer periods of low-intensity running [1,7].…”