Adaptive Adjustment of Training Load in Track and Field Sports Based on Machine Learning

Based on machine learning methods, this study proposes a model for adaptive adjustment of training load in track and field sports, aiming to improve the training effect of athletes and avoid overtraining. By constructing a deep neural network (DNN) model and combining key factors such as the athlete's fatigue index, training load and cumulative training load, the model can dynamically predict and adjust the training load. The experimental results show that DNN is superior to support vector machine (SVM) and random forest (RF) in terms of training load prediction accuracy and error control, and has strong practical application value. The study found that with the increase of training load, the athlete's fatigue index gradually increased, while the model-based load adjustment effectively prevented overtraining and ensured the maximization of training effect. This study provides new ideas for personalized training load adjustment and provides scientific support for the long-term health and competitive performance of athletes.