Identifying factors affecting sports achievement and endurance using artificial intelligence

Abstract Introduction Sports achievements and performance are not solely based on outstanding endurance, especially in tactical and technical sports. Artificial intelligence (AI) can help us to identify correlations that can contribute to the successful preparation of athletes. Purpose Our aim was t...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:European heart journal 2024-10, Vol.45 (Supplement_1)
Hauptverfasser: Sydo, N, Csulak, E, Kiss, A R, Petrov, I, Takacs, T, Bohus, G Y, Staub, L, Toser, Z, Balla, D, Vago, H, Kovacs, A, Merkely, B
Format: Artikel
Sprache:eng
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Abstract Introduction Sports achievements and performance are not solely based on outstanding endurance, especially in tactical and technical sports. Artificial intelligence (AI) can help us to identify correlations that can contribute to the successful preparation of athletes. Purpose Our aim was to further analyze and evaluate our previously initiated AI-based sports performance studies, identifying the parameters that determine athletes’ achievement and endurance. Methods First, sports cardiology screening was performed in all athletes containing the following examinations: patient’s history, ECG, laboratory test, body composition analysis, echocardiography and cardiopulmonary exercise testing. A database from the sports cardiology screening results was established. Then, we created two scoring systems based on the best ever result (Achievement Score) and the performance on the cardiopulmonary exercise test (Endurance Score). We identified the most important influencing factors using a neural network and characterized the strength of the variables using Shappley Additive Explanation (SHAP) values. Results We examined 1932 tests of 891 athletes and the AI analysis included 917 tests of 546 athletes (20.2±6.2 years, males: 397, 72.7%). Majority of the examined athletes were swimmers (27.4%), followed by basketball players (21.8%), water polo players (15.2%), handball players (13.3%) and football players (13.1%). The most important Achievement Score determinants were the weekly training hours (SHAP=0.27), training years (SHAP=0.26) and the age (SHAP=0.15). Meanwhile, Endurance Score was mainly affected by skeletal muscle mass (SHAP=0.85), weight (SHAP=0.6) and peak heart rate during exercise (SHAP=0.32). Our results were validated on a test population (N=102) and in the Achievement Score the mean absolute error (MAE) was 0.64, the determination coefficient (R2) was 0.56, in Endurance Score the MAE was 0.84 and the R2 was 0.71. Conclusion Based on our results, while endurance is mainly determined by the athlete’s physical characteristics, in athletic achievement experience is more important. With our scoring systems the athlete’s achievement and endurance can be predicted well.
ISSN:0195-668X
1522-9645
DOI:10.1093/eurheartj/ehae666.2989