Original title: Machine Learning-Based Estimation of Sports Performance in Elite Athletes
The researchers used athlete records to predict two kinds of sports performance: competition achievements and endurance capacity. They tested machine-learning models, which learn patterns from past examples, and used a method called SHAP to identify which factors mattered most to each prediction. Training years, weekly training hours, and age mattered most for achievements. Muscle mass, weight, and peak heart rate mattered most for endurance.
Imagine sorting a pile of sports certificates and trying to guess who has the longest-lasting stamina. The certificates tell you about past results, but they do not reveal the same things as a person’s body size or heart rate during intense effort. In this paper, the certificates stand for competition achievements, while body traits and peak heart rate help estimate endurance; the researchers check which clues best match each outcome.
Coaches and sports clinicians may be measuring two different things when they assess how well an athlete performs. The results suggest that competition achievements relate more to training and experience, while an absolute measure of endurance relates more to body composition and peak heart rate.
In the test set, the Achievement Score model accounted for less variation in results than the Endurance Score model, and its confidence interval extended from 0.00 to 0.48. The abstract does not establish that these predictions apply beyond the athletes and screening exams analyzed, or that the identified factors cause either type of performance.
Ask anything about this paper. Answers come from its abstract, and say so when the abstract doesn't cover your question.
Machine Learning-Based Estimation of Sports Performance in Elite Athletes — Titanilla Takács, G Y Bohus, Márton Tokodi et al. — Sports — doi:10.3390/sports14100431
Abstract licence: CC BY
Explanation generated by Synthesis from the abstract — it is not the original paper.