Fig. 3

Performance comparison of eight models on the cohorts. A. ROC curves for each model in the training cohort; B. ROC curves for each model in the validation set; C. The changes in Kappa values of eight machine learning algorithms in the training set and the validation set; D. LightGBM based on the SHAP algorithm; E. Randomforest model based on the SHAP algorithm; F. Venn diagram of cross-variables among multiple algorithms