Abstract
Mental health issues among students are becoming increasingly serious, and physical exercise is closely linked to both physical and mental health. Therefore, there is a correlation between physical education and students' physical and mental well-being. When examining the relationship between physical education mechanisms in higher education institutions and students' mental health, fluctuations such as individual differences often appear in the data. To address this issue, this paper proposes the "Sparrow Search Tree Network" (SSTNet), a model that employs the "Sparrow Search Algorithm" (SSA) to systematically analyze the relationship between the two. Additionally, SSTNet incorporates the Huber loss function to enhance the model's stability when dealing with outlier samples. The results indicate that SSTNet achieves higher predictive accuracy, model stability, and generalization capability than the control model. Applying SSTNet to analyze the modeled associations between university physical education mechanisms and students' mental health indicates that moderate exercise intensity coupled with a high-quality teaching and interactive environment is associated with lower DASS-21 total scores, that is, with lower psychological distress rather than with better scores on a distress scale.