Table 5 Comparison of classification performance for six cross-domain tasks.

From: Coronary heart disease prediction method fusing domain-adaptive transfer learning with graph convolutional networks (GCN)

Source

Target

Accuracy (%)

AUC (%)

Precision (%)

Recall (%)

F1score (%)

MCC (%)

Balanced accuracy (%)

H

D

97.9

94.1

93.5

81.2

86.2

74.6

79.4

M

D

97.9

95.1

95.9

79.3

85.6

71.6

81.1

D

H

94.8

96.5

93.4

89.5

91.3

81.4

87.0

M

H

95.0

96.3

94.7

88.8

91.4

84.0

88.2

D

M

96.9

95.7

93.8

79.5

85.0

71.7

76.6

H

M

96.7

95.7

95.6

76.5

83.1

67.2

73.4

Back to article page