Optimizing Physical Education Schedules for Long-Term Health Benefits

Liang Tan¹Qin Chen²Jianwei Wu¹Mingbang Li³Tianyu Liu^1*

• ¹Chengdu University of Traditional Chinese Medicine, Chengdu, China
• ²Hunan University of Humanities, Science and Technology, Loudi, Hunan Province, China
• ³College of Physical Education and Health Sciences, Zhejiang Normal University, Jinhua, Zhejiang, China

Physical education (PE) is critical to promoting long-term health and wellness. This study provides an efficient method for optimizing PE schedules using deep learning (DL) models. The proposed DL model predicts the improvement in the fitness score based on demographic and activity-related variables. The proposed model combines CNN and LSTM layers to extract spatial and temporal features. Thereafter, a fusion layer combines the obtained spatial and temporal features. Finally, a customized loss is designed to predict the fitness score. Extensive experimental results reveal that the proposed model outperforms competitive models, with considerable gains in Mean Squared Error (MSE), R 2 , and Mean Absolute Error (MAE) of 1.35%, 1.18%, and 1.22%, respectively. The findings demonstrate that the proposed model can optimize PE schedules to increase fitness levels and provide long-term health benefits. These findings provide a robust framework that educational institutions and policymakers can use to design effective PE programs tailored to diverse populations.