Table 9 Comparative performance of different activation functions in the CNN-BLSTM model.
From: Hybrid CNN-BLSTM architecture for classification and detection of arrhythmia in ECG signals
Activation function | Accuracy (%) | Convergence stability | ||
|---|---|---|---|---|
ReLU | 98.71 | 0.971 | Moderate | Moderate |
Leaky ReLU | 98.85 | 0.974 | Good | Good |
ELU | 98.92 | 0.976 | Good | Good |
Swish | 99.06 | 0.98 | Very good | Stable |
Mish (proposed) | 99.52 | 0.985 | Excellent | Very stable |
