Fig 1.
Flowchart of patient inclusion.
Fig 2.
Workflow diagram.
Table 1.
Baseline characteristics of the patients.
Table 2.
Comparison of ROC performance before and after optimization.
Table 3.
Optimized model prediction performance.
Fig 3.
Feature selection using the Boruta algorithm.
Fig 4.
Receiver operating characteristic curves of 16 models for in-hospital mortality in patients with coronary heart disease with diabetes mellitus.
Fig 5.
Permutation importance (a, c) and SHAP summary plots (b, d) for the Gradient Boosting Classifier and Random Forest Classifier, showing variable associations with in-hospital mortality.
Fig 6.
The SHapley Additive exPlanations (SHAP) waterfall.