Table 14 Results of classifiers (in %) with Relief.

From: An extensive experimental analysis for heart disease prediction using artificial intelligence techniques

Classifier

Training accuracy

Testing Accuracy

Precision

Sensitivity

Specificity

F1 score

AUC

LR

82.9

85.3

87

86

84

87

90

DT

85.8

86.6

86

90

82

88

90

RF

86.2

88.2

89

89

87

89

94

KNN

86.9

85.2

84

86

79

85

91

SVM

82.8

84.9

87

86

84

86

90

GNB

81.6

84.5

86

86

82

86

90

XGBoost

98

90.3

92

90

91

91

94

AdaBoost

84.8

85.7

85

90

80

87

91

SGD

83.1

84.5

87

85

84

86

90

GB

91.1

89.9

90

92

88

91

95

ETC

84.2

86.1

87

89

83

88

92

CatBoost

85.4

81.9

95

71

95

81

94

LightGBM

91.3

84.9

95

76

95

85

94

MLP

86.9

84.9

83

92

77

87

92

RNN

87.5

87

94

69

94

79

82

LSTM

85.4

85.7

91

71

92

80

81

GRU

86.1

89.1

95

74

95

83

85

Bi-LSTM

85.8

88.2

93

76

93

84

84

Bi-GRU

86.9

88.7

94

74

94

83

84

CNN

86.3

87.8

94

73

94

82

85

Hybrid Model

86.3

88.2

95

73

95

82

92

  1. Significant values are in [bold].