Using physical literacy to predict physical activity among university students: a machine learning logistic regression model

Physical inactivity has become an increasingly challenging concern for health systems and policymakers globally. The purpose of this study was to examine whether physical literacy could be used to predict the physical activity behaviors of Chinese university students in order to provide an efficient tool for encouraging public health. 1,591 Chinese university students (females = 813, males = 778) with a mean age of 19.63 years (SD = 1.36) participated in this study. Findings indicated a practical machine learning logistic regression model (LRM) for predicting PA behaviors among university students. The accuracy on the confusion matrix was .715, and the Receiver operating characteristic curve and Concordance test indicated that the LRM is a good model. The researchers conclude that the present LRM can be a useful tool for higher education institutions to monitor and foster students' PA behaviors to promote a healthy lifestyle and life-long participation in physical activities.