Predicting college students' exercise dependence: a machine learning approach.

Exercise dependence behavior among college students is a critical issue in sports psychology that deserve closer examination, and artificial intelligence offer a useful ways to explore its mechanisms and predicting associated risks. In this study, data were collected from 2,745 college students using three standardized questionnaires, covering (i) exercise dependence behavior, (ii) psychological characteristics (e.g., exercise identity, weight biases), and (iii) basic demographic information. We used four widely used machine learning algorithms: logistic regression, random forest, extreme gradient boosting (XGBoost), and multilayer perceptron, and their outputs were further integrated through an ensemble learning techniques to further enhance the robustness and predictive power of the models. The stacking ensemble model achieved a mean AUC of 0.96 in identifying exercise dependence risk among college students, demonstrating that integrating multiple machine learning approaches can yield robust and highly accurate risk prediction in this setting. Among the variables, the most influential predictors of exercise dependence behavior included prolonging exercise to obtain the desired effect, allocating most leisure time on exercise, experiencing difficulty in reducing exercise frequency, and actual exercise time longer than originally planned. These findings uncovers the key psychological and behavioral mechanisms underlying in exercise dependence among college students and show that artificial intelligence methods can be effectively applied to support risk monitoring in sport and psychological health contexts.