Skip to main content

Table 3 The performance of the prediction models with important factors based on various indices for two ratios

From: Machine Learning Algorithms for understanding the determinants of under-five Mortality

Train/test ratios

Measures

Decision tree

Random forest

Naive Bayes

K-Nearest neighbour

Logistic regression

SVM regression

Neural network

Ridge regression

70/30

Sensitivity

83.90

58.40

21.83

57.78

70.59

70.23

81.03

71.43

Specificity

95.53

96.05

93.73

94.04

95.55

95.55

95.75

94.33

Precision

33.21

42.16

5.27

9.56

33.82

33.82

36.64

14.09

Accuracy

95.23

94.28

92.60

93.65

94.77

94.75

95.31

94.04

F1 Score

47.59

48.97

8.49

16.40

45.73

45.66

50.46

23.54

Negative Predictive value

99.56

97.91

98.68

99.51

99.02

99.00

99.40

99.61

Cohen’s kappa values

0.46

0.46

0.06

0.15

0.44

0.44

0.48

0.22

80/20

Sensitivity

84.3

58.40

21.21

60.00

71.79

71.64

71.51

71.68

Specificity

95.60

96.05

93.70

94.01

95.67

95.62

96.25

94.36

Precision

34.43

42.16

5.13

9.34

35.90

35.16

45.05

14.84

Accuracy

95.23

94.28

92.55

93.67

94.89

94.85

95.29

94.05

F1 Score

48.89

48.97

8.26

16.16

47.86

47.17

55.28

24.58

Negative Predictive Value

99.55

97.91

98.67

99.56

99.01

99.03

98.74

99.59

Cohen’s kappa values

0.47

0.46

0.06

0.15

0.46

0.45

0.53

0.23