| Machine Learning–Based Risk Assessment for Hamstring Strain Injuries in Field Sport Athletes |
| Paper ID : 1757-SPORTCONGRESS |
| Authors |
|
Sina Lagzi * 1 Student Research Committee, Ahvaz Jundishapur University of Medical Sciences, Ahvaz, Iran
2 Rehabilitation Research Center, Ahvaz Jundishapur University of Medical Sciences, Ahvaz, Iran |
| Abstract |
| Introduction: Hamstring strain injuries are among the most frequent non-contact injuries in field sports such as soccer, and track events. They often lead to significant time loss, reduced performance, and a high risk of recurrence. Therefore, early identification and prevention of hamstring injury risk are essential for maintaining athlete performance and team success. Recent advances in artificial intelligence (AI) and wearable technology have enabled the analysis of large datasets from athlete movement and workload to detect early indicators of injury risk more accurately than traditional screening methods. Methods: This literature review focused on studies applying machine learning (ML) to predict hamstring injury risk in athletes. Included research utilized wearable sensors, GPS-based metrics, or video-based gait data to monitor sprint mechanics, acceleration, deceleration, workload, and stride symmetry. ML models were trained to identify movement patterns and workload profiles associated with elevated injury risk. Results: Evidence shows that combining GPS-based workload measures with video-derived kinematic data improves predictive accuracy compared to single-source metrics. ML algorithms can detect subtle deviations in sprint mechanics, stride asymmetry, or accumulated workload that precede hamstring injury. These models enable individualized load monitoring and preventive interventions, allowing coaches and practitioners to manage training intensity and implement targeted exercises before injuries occur. Non-invasive, field-based monitoring also supports large-scale, real-time evaluation without disrupting athlete routines. Conclusion: AI-driven machine learning models represent a promising approach to proactive hamstring injury prevention in field sport athletes. By integrating kinematic and workload data, these systems enhance risk prediction, support personalized training adjustments, and may reduce injury incidence and time loss. Future research should focus on model validation across diverse sports, standardization of datasets, and the development of interpretable algorithms for practical implementation in athletic environments. |
| Keywords |
| Hamstring Strain Injury, Injury Prevention, Machine Learning, Wearable Technology |
| Status: Abstract Accepted (Poster Presentation) |
