Personalization of optimizing athlete training plans using gated recurrent unit (GRU)

Abstract

The training plan of athletes is crucial for their performance and health. However, due to the differences among athletes in many aspects, traditional training plans are difficult to adapt to the personalized needs of athletes. This article used gated recurrent unit (GRU) to conduct in-depth analysis and learning of athlete training data, and designed targeted training plans for athletes to achieve the goal of improving sports performance. Firstly, a set of athlete training data was collected, and wavelet transform and box plot methods were utilized to denoise and handle outliers in the data; then, the GRU model was used to analyze these data and extract the most valuable feature indicators, providing a basis for adjusting subsequent training schemes. Finally, taking 15 basketball players from A University as research subjects, a personalized training plan was designed to help them better train and improve training effectiveness. The experiment showed that 15 athletes were selected, and their average score on the Functional Movement Screen (FMS) test after the experiment was 14.73 points, which was 2.4 points higher than before the experiment. Before the experiment, only 4 people had an FMS score of 14 or above, accounting for 26.67%, which was very low. However, after the experiment, 12 people had an FMS score of 14 or above, accounting for as high as 80%. Based on the GRU model, personalized training plans were studied to improve athlete performance and health, providing new research ideas for athletes to develop personalized training plans. This can better adapt to the different needs of different types of athletes, thereby improving their training effectiveness and competitive level.