Table 15 Results of classifiers (in %) with Lasso.

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

83.4

85.3

87

86

84

87

91

DT

85.7

86.1

86

90

81

88

92

RF

86.9

88.2

86

94

81

90

94

KNN

87.2

84.5

85

87

81

86

90

SVM

83.6

85.7

87

87

84

87

90

GNB

82.8

86.1

87

88

89

87

91

XGBoost

88

87.4

88

90

85

89

93

AdaBoost

85.1

84

85

86

81

86

92

SGD 8

3.7

86.1

88

87

85

87

91

GB

91.8

85.3

88

85

86

86

94

ETC

85.2

85.3

87

86

84

87

93

CatBoost

85.9

80.7

96

68

96

80

94

LightGBM

88.3

82.4

92

74

93

82

94

MLP

90.2

83.6

88

82

86

85

91

RNN

87.7

85.3

96

67

96

79

82

LSTM

86.7

88.2

93

70

93

80

81

GRU

87.1

86.6

94

70

94

80

82

Bi-LSTM

86.2

88.7

94

70

94

80

82

Bi-GRU

87.3

88.2

95

66

95

78

80

CNN

87.3

87.2

94

70

94

80

83

Hybrid Model

88.6

86.1

93

70

94

80

93

  1. Significant values are in [bold].