Table 2.

Machine learning model features

Author	Age (years)	Death/S-AKI cases	Machine learning algorithme	Best Algorithm	Delete missing variables	Imputation method	Predictor selection	Validation	Calibration indicators
Lei Dong	NA	1769/7544	LR, Lasso, Rpart, RF, XGBoost, and ANN	XGBoost	>30%	KNN	LR, Lasso, and RF	Bootstrap and external validation	Calibration curve, Brier score, and kappa coefficient
Zhiyan Fan	NA	NA/2599	LR, RF, XGBoost, MLP, and SVC	XGBoost	>30%	MiceForest	Recursive feature elimination	Random sampling and external validation	NA
Tianyun Gao	67.0 ± 16.1	2352/12,196	KNN, XGBoost, NB, DT, SVM, linear/rbf), RF, and LR	RF	>25%	MiceForest	RF	10-fold cross-validation	Calibration curve
Xunliang Li	68.7(57.2,79.6)	1629/8129	LR, SVM, KNN, DT, RF, and XGBoost	XGBoost	NA	MiceForest	LASSO	Random sampling	Calibration curve
Xiaoqin Luo	69 (58,79)	NA/9537	XGBoost, RF, and SVM	XGBoost	NA	XGBoost	XGBoost	Random sampling	Calibration curve
Jie Tang	77 (71, 84)	1813/6613	LR, SVM, GBM, AdaBoost, XGBoost、CatBoost, NB, NN, MLP, KNN, and RF	CatBoost	>5%	KNN	Recursive feature elimination	Random sampling	Calibration plot
Jijun Yang	67(57,78)	1940/9158	LR, RF, GBM, and XGBoost	XGBoost	>20%	RF	Boruta algorithm	5-fold cross validation	Calibration curve
Hongshan Zhou	67.7 ± 15.2	16,154/3318	KNN, AdaBoost, MLP, SVM, LR, NB, GBDT, RF, LightGBM, and XGBoost	CatBoost	NA	KNN	Recursive feature elimination	Random sampling and external validation	NA