Skip to main content
. 2024 Nov 10;14(22):2516. doi: 10.3390/diagnostics14222516

Table 2.

Sport and injury metrics addressed by AI techniques.

Sport AI Method Author (Year) Application Sample Size Data Source
Athletics ML Rahlf A.L.
(2022) [34]		 Identification of risk factors
for running-related injuries
using ML techniques.		 120 runners Private, collected via surveys and biomechanical measurements
Connor M.
(2022) [35]		 Adaptive training plan
generation using control
systems and AI.		 Simulation
(N = 1800)		 Simulated data, no real data collected
Alpine Skiing ML Hermann A.
(2021) [36] 		Knee injury prevention in
alpine skiing using
mechatronic ski bindings.		 Not specified Research-based, design, and
biomechanical
assessment
Basketball
Handball		 ML (ANN, SVM, Markov 1) Claudino J.G
(2019) [37]		 Evaluation of performance
and injury risk in basketball, handball, and volleyball.		 6456 athletes
(various sports)		 Public, data from academic research
Baseball CNN Koseler K.
(2017) [38]		 Statistical analysis and injury
risk prediction in baseball. 		Not specified Public, systematic literature review
Volleyball
Beach Volleyball		 Deep CNN Kautz T.
(2017) [39]		 Activity recognition and
performance analysis in beach volleyball.		 30 participants
(11 women and
19 men)		 Private, collected via wearable
sensors attached to the wrist
Futsal RF, XGBoost 2 Ruiz-Pérez I.
(2021) [40]		 Predicting lower-extremity
soft tissue injuries in elite
futsal players.		 139 players Private, preseason screening and monitoring
Football ML Rico-González M.
(2023) [41]		 Focused on a
common injury type.		 Not specified Public, systematic
literature review
ML (XGBoost) Rommers N.
(2020) [42]		 Systematic review of ML
applications in football for
injury risk prediction.		 734 players Private, data from football academy monitoring
ML Kolodziej M.
(2021) [43]		 Injury risk assessment in elite youth football players. 62 players Private, collected from neuromuscular performance tests
Nassis GP.
(2023) [5]		 Identification of neuromuscular performance parameters as risk factors for non-contact injuries. Not specified Public, literature
Review
Hecksteden A.
(2023) [44] 		Review of ML applications in football with an emphasis on
injury risk.		 84 players Private, screening and monitoring with ML methods
Rossi A.
(2018) [6] 		Combining screening and
monitoring data with ML for
injury prediction.		 Not specified Private, GPS data from training
sessions and games
Windsor J.
(2022) [45]		 Effective injury forecasting in football using GPS training
data and ML techniques.		 77 players Private, biomechanical assessments and medical history
Ayala F.
(2019) [46]		 Analysis of foot biomechanics and injury history in varsity
football athletes.		 96 players Private, preseason screening
KNN, ANN Calderón-Díaz M.
(2023) [47]		 Development of preventive
models for hamstring injuries
in professional soccer.		 Not specified Private, biomechanical analysis with sensors
Hockey RF, Neural Networks Schickendantz M.S.
(2020) [48]		 Injury risk prediction and
management in NHL hockey players.		 2322 players Public, data from publicly reported NHL injury databases

1 A Markov model is a type of statistical model that predicts future states based only on the current state, without requiring knowledge of the sequence of events that preceded it. 2 Extreme Gradient Boosting is a powerful ML algorithm based on decision trees. It builds multiple trees sequentially, with each new tree correcting errors from the previous ones.