A logistic regression modeling study of college students’ mental health status in predicting suicidal behavior

The problem of college student suicide has been widely discussed, and the prediction of students’ suicide risk through their mental health status is one of the keys to preventing college student suicide. To this end, this paper proposes a suicidal behavior prediction model based on a logistic regression model. It solves the sample imbalance problem by sampling method and adjusts the model parameters using parameters and learning curves. The experiments show that the AUC values of the training and test sets are 0.922 and 0.934, respectively, and the model has a good predictive ability. Logistic regression analysis showed that emotional health problems were significant factors affecting suicidal behavior, in which the correlation coefficients of anxiety and depression with suicidal thoughts were 0.198 and 0.138, respectively, and the results of multifactorial regression analysis of these two emotions showed that economic status, inter-parental violence, and family dysfunction were related to depression and anxiety.