Table 4 Performance of the machine-learning algorithms.

From: Machine learning for characterizing risk of type 2 diabetes mellitus in a rural Chinese population: the Henan Rural Cohort Study

Lab

Model

AUC

Accuracy(%)

Sensitivity(%)

Specificity(%)

PPV(%)

NPV(%)

AUPR

With lab

LR

0.841

(0.825–0.858)

75.23

78.49

74.91

23.37

97.28

0.493

 

CART

0.811

(0.793–0.829)

80.06

66.97

81.33

25.91

96.19

0.433

 

GBM

0.872

(0.858–0.886)

81.20

76.04

81.71

28.83

97.22

0.546

 

ANN

0.858

(0.842–0.873)

74.01

80.95

73.34

22.83

97.53

0.520

 

RF

0.868

(0.854–0.883)

85.90

79.57

78.14

26.19

97.52

0.538

 

SVM

0.835

(0.818–0.851)

76.42

74.65

76.59

23.71

96.88

0.490

No lab

LR

0.804

(0.787–0.821)

75.06

72.35

75.33

22.23

96.55

0.313

 

CART

0.767

(0.749–0.784)

62.79

79.26

61.18

16.60

96.80

0.235

 

GBM

0.817

(0.801–0.833)

70.28

78.96

69.43

20.11

97.13

0.345

 

ANN

0.808

(0.791–0.825)

70.52

78.03

69.79

20.11

97.02

0.328

 

RF

0.803

(0.786–0.820)

70.77

75.58

70.30

19.87

96.73

0.327

 

SVM

0.800

(0.783–0.818)

76.46

70.51

77.04

23.03

96.40

0.316

  1. Abbreviation: LR, logistic regression; CART, classification and regression tree; GBM, gradient boosting machine; ANN, artificial neural network; RF, Random forest; SVM, Support vector machine.
Back to article page