A Champion Model of Men's 100m Sprint Based on Functional Data Analysis

A data-driven quantitative champion model can provide valuable reference for sport training and can be integrated into AI systems of sports training. Inspired by the mechanism of energy metabolism in sprint, we propose a function with 3 interpretable parameters to simulate the speed rhythm for all the 3 phases of men's 100 meters sprint. We build a dataset containing 102 items of split time data from major 100 meters games. The MSE results of regression show that our model is better that the literature. We further build a champion model on this dataset and integrate this model in the intelligent track-and-field ground system on our campus, and therefore collect sport big data now.