Analysis of the Exercise Intention-Behavior Gap among College Students Using Explainable Machine Learning

Cui Cui^*Jixin Yin

• Huanghe Jiaotong University, Jiaozuo, China

The physical fitness of college students has become a significant global public health concern, with one major challenge being the intention-behavior gap in exercise engagement. This study aims to identify the key factors contributing to this gap and provide targeted recommendations for intervention. To this end, we first collected survey data from 1866 college students who are active TikTok users in China, including variables such as gender, academic grade, health belief perception, and planned behavior perception. Then, multiple machine learning models were developed to predict the presence of the gap, and the best-performing model was interpreted using Shapley Additive Explanations to assess feature importance. Finally, the results indicate that perceived barriers are the most influential factor affecting the gap. Notably, male students in higher academic years have lower perceived barriers. Moreover, stronger subjective norms regarding physical inactivity are less likely to exhibit the gap. These findings suggest that universities should focus on reducing perceived barriers, cultivating a supportive exercise culture, and optimizing the allocation of physical education resources to promote the effective translation of exercise intentions into actions.