Table 4 Prediction performance of external validation in AUH

XGBoost

0.761 (0.742- 0.765)

57.64% (57.16–58.12)

81.43% (80.62–82.22)

93.13% (92.93–93.48)

30.29% (29.97–30.61)

CatBoost

0.748 (0.735–0.756)

59.09% (58.61–59-58)

83.25% (82.51–83.96)

93.21% (92.93–93.48)

34.33% (34.01–34.66)

Random Forest

0.741 (0.724–0.752)

58.89% (58.40.59.37)

81.84% (81.08–82.57)

92.49% (92.20–92.78)

34.38% (34.05–34.72)

Decision tree

0.710 (0.692–0.722)

56.14% (55.64–56.63)

80.11% (79.33–80.86)

91.28% (90.96–91.59)

32.99% (32.69–33.32)

logistic regression

0.699 (0.691–0.710)

51.30% (50.81–81.80)

84.72% (84.01–85.41)

92.73% (92.41–93.04)

31.41% (31.13–31.69)