Fig. 3

A Results of Bayesian hyperparameter optimization. The final hyperparameters were determined by the Optuna hyperparameter tuning framework. The Optuna optimizer maximized the out-of-sample area under the receiver operating characteristic (AUCROC). DT, Decision Tree; Enet, Elastic Networks; KNN, K Nearest Neighbors; Lightgbm, Light Gradient boosting machine; RF, Random Forest; XGBoost, eXtreme Gradient Boosting; SVM, support vector machines; MLP, multilayer perceptron; HPO, hyperparameter optimization. B Model performance presentation. (a) The ROC curve of each model in the training group (b) and the testing group; the X-axis represents "1-specificity," while the Y-axis represents sensitivity. (c) The precision-recall of each model in the training group (d) and the testing group, the X-axis represents precision, the Y-axis represents recall. C Model performance illustration. (a). The calibration curve of each model in the training group (b) and the testing group (c). The DCA curve of each model in the training group. (d). The DCA curve of each model in the testing group. BS, Best calibration. DCA, Decision curve analysis