Table 15 A-Test scores comparing EHST with best baseline models across datasets. Values close to 0 or 1 imply large effect sizes.

From: Explainable attention-based deep learning for classification and interpretation of heart murmurs using phonocardiograms

Dataset

Accuracy

Macro F1

AUC

HeartWave

0.12

0.14

0.15

CirCor DigiScope

0.18

0.20

0.16

PhysioNet CinC

0.21

0.19

0.17

Pascal (A+B)

0.20

0.22

0.18

GitHub Valvular

0.13

0.15

0.14

Shenzhen (HSS)

0.16

0.18

0.17

Back to article page