Researchers developed machine learning models to predict weekly injury risk in competitive endurance runners by analyzing multiple types of risk factors simultaneously. They tracked 142 runners for a full year, collecting data on genetics, training history, strength, movement patterns, body composition, nutrition, and weekly training loads to see how well computer algorithms could forecast when injuries might occur.
Key Findings
- ✓Machine learning models achieved moderate accuracy in predicting weekly injury risk using multidisciplinary data
- ✓Random forest algorithms performed best among the computational approaches tested
- ✓Including broader risk factors improved prediction only for certain modeling methods
AI-generated • See paper for full context