Effect of GAN-based data augmentation on the binary classification task
Precision | Recall | F1-score | ROC AUC | |
Original data (n=864) | 0.81±0.09 | 0.26±0.04 | 0.38±0.04 | 0.77±0.01 |
Augmented data (n=1728) | 0.86±0.04 | 0.46±0.04 | 0.60±0.05 | 0.81±0.00 |
Fold change | 1.06 | 1.77 | 1.58 | 1.05 |
P value | 0.614 | 0.008 | 0.006 | 0.002 |
AUC, area under the curve; GAN, generative adversarial network; ROC, receiver operator characteristic.