Table 3 Model Performance.

From: Development of risk models for early detection and prediction of chronic kidney disease in clinical settings

Female Model

Accuracy

Sensitivity

Specificity

Precision

F1 score

AUC

Random Forest

86.34%

92.26%

53.57%

91.66%

91.96%

0.90

Stratified K-fold Random Forest

86.02%

97.40%

31.25%

87.20%

92.02%

0.94

Neural Network

84.58%

88.44%

68.29%

92.16%

87.92%

0.89

Stratified K-fold Neural Network

94.32%

96.82%

83.01%

96.26%

96.54%

0.95

Male Model

Accuracy

Sensitivity

Specificity

Precision

F1 score

AUC

Random Forest

89.07%

95.48%

53.57%

91.92%

93.63%

0.88

Stratified K-fold Random Forest

85.71%

97.53%

54.84%

84.94%

90.80%

0.91

Neural Network

84.84%

90.52%

72.73%

87.61%

89.04%

0.88

Stratified K-fold Neural Network

92.82%

94.15%

89.49%

95.66%

94.90%

0.94