Table 7. Results of activation functions in training set.
| Activation functions | Accuracy | Loss |
|---|---|---|
| Sigmoid | 0.6541 | 0.8521 |
| tanh | 0.7551 | 0.6170 |
| Leaky-ReLU | 0.7615 | 0.5952 |
| ELU | 0.7414 | 0.6412 |
| SELU | 0.7299 | 0.6698 |
| SoftPlus | 0.7078 | 0.7197 |
| ReLU | 0.7630 | 0.5904 |
Note:
Values in bold represent the optimum values for each group.