Skip to main content

Table 7 Experimental results on Statlog dataset with 13 features

From: A hybrid cost-sensitive ensemble for heart disease prediction

Mean ± SD

RF

LR

SVM

ELM

KNN

Proposed ensemble

E (%)

71.76 ± 7.44

77.16 ± 4.53

68.49 ± 6.06

77.31 ± 8.11

66.70 ± 4.26

86.36 ± 5.51\(^*\)

Precision (%)

65.19 ± 14.96

73.70 ± 7.77

68.15 ± 10.03

61.48 ± 29.28

59.26 ± 11.05

78.52 ± 7.37

Recall (%)

86.54 ± 10.48

83.13 ± 8.57

75.62 ± 6.28

82.45 ± 18.42

73.57 ± 13.65

92.56 ± 8.19

G-mean

82.18 ± 9.64

83.72 ± 14.18

76.29 ± 7.45

82.60 ± 14.51

76.35 ± 18.16

90.17 ± 8.08

MC (%)

75.12  ± 9.10

56.30 ± 7.77

62.69 ± 25.27

41.12 ± 33.75

85.19 ± 43.82

34.81 ± 24.58

Specificity (%)

78.05 ± 7.26

84.32 ± 8.97

76.96 ± 16.40

82.81 ± 8.72

79.23 ± 17.11

87.84 ± 5.73

AUC (%)

79.35 ± 11.28

83.16 ± 9.78

83.16 ± 9.82

81.27 ± 12.51

78.53 ± 6.94

87.99 ± 8.39

  1. \(*\) The average \(+-\) sd on 10-folds CV. The best result is bolded.