Currently, the mental health problems of students in colleges and universities are becoming more and more prominent, this study constructed a Bayesian network-based prediction model of college students’ mental health status and designed corresponding interventions. In terms of methodology, firstly, multi-dimensional student behavioral features were represented and extracted, including consumption features (dietary regularity, diligence, number of meals shared) and Internet access features (length of time on the Internet, downtime, traffic use); secondly, the Jenks Natural Breaks algorithm was used to label the featured data, and the Apriori algorithm was used to mine the behavioral association rules of the psychologically healthy and psychologically abnormal students; then a Bayesian network model was built to predict the mental health status of students and design interventions accordingly. Then a Bayesian network model is established to predict students’ mental health status, and the results are compared with those of decision tree, support vector machine, and boosting algorithm. The results show that the Bayesian network prediction model has the best performance, with an accuracy of 0.9415, a recall of 0.9387, and an F1 value of 0.9389, which are higher than those of the other three algorithmic models in the anxiety binary classification experiment; and in the anxiety multiclassification prediction experiment, the Bayesian network model has an Fmacro value of 0.8549, and an Fmicro value of 0.8814, which are also better than the other models. The study also found that the group of psychologically abnormal students is usually characterized by less regular diet, less diligence, fewer number of people sharing meals, longer time on the Internet and more traffic use, and later time off the Internet on weekdays. The Bayesian network prediction model constructed in this study has high accuracy in predicting the mental health status of college students, which can provide technical support for mental health monitoring and precise intervention in colleges and universities.
