Table 10 Results of classifiers (in %) with FDA.

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

78.2

82.8

86

82

83

84

87

DT

86.5

83.2

82

89

77

85

91

RF

87.4

86.6

87

89

83

88

93

KNN

81.9

84.9

87

85

85

86

89

SVM

78.8

83.6

85

85

81

85

89

GNB

78

82.4

86

82

83

84

88

XGBoost

98.9

89.5

91

90

89

90

94

AdaBoost

84.5

84

86

85

83

85

90

SGD

79.3

82.8

86

82

83

84

87

GB

98.3

87.1

90

90

88

90

95

ETC

83.8

84.1

87

86

84

86

92

CatBoost

85.9

81.9

96

70

96

81

94

LightGBM

93.9

81.5

92

73

93

81

94

MLP

80.5

81.1

84

82

80

83

89

RNN

84.8

86.6

94

69

94

79

82

LSTM

84

85.3

93

79

93

85

86

GRU

89.4

85.3

95

74

95

83

85

Bi-LSTM

89.4

85.7

94

76

94

84

85

Bi-GRU

90

85.7

93

73

94

82

83

CNN

86.4

86

94

73

94

82

86

Hybrid Model

80.7

85.3

95

67

95

79

93

  1. Significant values are in [bold].