Table 5 Ablation study results for cnn-bigru-attention model layers.
From: Deep learning approaches for assessing pediatric sleep apnea severity through SpO2 signals
Model | Average RMSE | Average R2 | Average Four-class kappa | Average Four-class accuracy (%) |
|---|---|---|---|---|
CNN-BiGRU-Attention | 2.31 | 0.79 | 0.63 | 75.95 |
CNN-BiGRU | 2.66 | 0.74 | 0.56 | 70.26 |
BiGRU-Attention | 6.31 | 0.21 | 0.23 | 46.26 |
CNN -Attention | 4.47 | 0.32 | 0.32 | 54.14 |
CNN | 3.69 | 0.44 | 0.43 | 64.4 |
