Fig 1.
Scheme prediction model development.
Table 1.
Optimal parameters for each machine learning model are selected through the grid search.
Table 2.
Characteristics of high-risk suicide (n = 7,443) and no high-risk suicide (n = 52,541).
Table 3.
Multivariate logistic regression analysis to identify factors associated with high risk of suicide.
Fig 2.
Receiver operating characteristic (ROC) curve.
Table 4.
Confusion matrix for prediction models (Test set).