Construction of a "speed-technical" model for elite female race walkers based on random forest.

Abstract

Objective: The rapid development of race-walking techniques and the growing global competition have heightened the importance of optimizing the relationship between technique and speed. This study aims to investigate the interaction between race-walking speed and technique and to establish the optimal combination of technical factors that contribute to peak performance. Methods: A "Speed-Technical" model was developed using random forest algorithms, with the SHAPley (Shap) method applied to evaluate the influence of various technical indicators on race-walking speed. Results: The analysis revealed the following hierarchy of factors impacting speed: Step frequency > Step length > Thigh angle > Flight distance > Upper-forearm angle > Head undulation distance > Landing angle > Rear pedal angle > Backpedal distance > Back swing distance > Arm swing angle > Front support distance > Front swing distance > Flight time. The optimal technical ranges for maximizing race-walking speed were found to be: Step frequency (>230 steps/min), Step length (>1.12 m), Thigh angle (50°-65°), Flight distance (0.26 m), Upper-forearm angle (77°), Head undulation distance (0.6-0.8 m), Landing angle (25°-30°), Rear pedal angle (32°-39°), Backpedal distance (0.37-0.43 m), Back swing distance (0.43-0.47 m), Arm swing angle (57°-62°), Front support distance (0.19-0.25 m), Front swing distance (0.25-0.30 m), and Flight time (<0.042 s). Conclusion: The study identifies the key technical factors that most significantly impact race-walking speed, offering novel insights that complement previous findings while highlighting differences in optimal ranges compared to traditional models. These results enhance our understanding of the intricate relationship between technique and speed, providing valuable implications for training and performance optimization.