Related Experiment Video
Updated: May 27, 2025
04:24
A Novel Digital Platform for a Monitored Home-based Cardiac Rehabilitation Program
Published on: April 19, 2019
11.5K
Machine learning models for reinjury risk prediction using cardiopulmonary exercise testing (CPET) data: optimizing
Arezoo Abasi1,2, Ahmad Nazari3, Azar Moezy3
1Department of Health Information Management, School of Health Management and Information Sciences, Iran University of Medical Sciences, Tehran, Iran.
Biodata Mining
|February 18, 2025
Summary
Machine learning models accurately predict reinjury risk in elite soccer players using Cardiopulmonary Exercise Testing (CPET) data. CatBoost and SVM models show high performance, aiding athlete recovery and injury prevention.
Area of Science:
- Sports Medicine
- Exercise Physiology
- Data Science in Sports
Background:
- Cardiopulmonary Exercise Testing (CPET) offers insights into athlete health but traditional models struggle with reinjury prediction.
- Machine learning (ML) can uncover complex patterns in CPET data for improved injury risk assessment.
Purpose of the Study:
- Develop ML models to predict reinjury risk in elite soccer players using CPET data.
- Identify key physiological and performance variables linked to reinjury.
- Evaluate ML algorithm performance for accurate prediction.
Main Methods:
- Analyzed CPET data from 256 elite soccer players.
- Employed ML models: CatBoost, SVM, Random Forest, XGBoost.
- Assessed performance using accuracy, precision, recall, F1-score, AUC, and SHAP values.
Main Results:
- CatBoost and SVM demonstrated superior performance in predicting reinjury.
- CatBoost achieved 0.9138 accuracy and 0.9148 F1-score; SVM achieved 0.9725 AUC.
- Concussion history and lower heart rate metrics (HRmax, HR2) were significantly associated with reinjury risk.
Conclusions:
- ML models, especially CatBoost and SVM, offer precise, data-driven tools for predicting reinjury risk.
- These models enhance athlete recovery and risk management strategies.
- Future research should integrate external factors like training load and psychological readiness.
