Table 11 Hyperparameter set-up for the best found MLP models for each regression dataset
From: Extreme learning machine versus classical feedforward network
Dataset | Hidden layer size | # of layers | Batch size | Early stopping patience | Activ. func. | Regul. coeff. |
|---|---|---|---|---|---|---|
Servo | 256 | 4 | 32 | 8 | ReLU | 0.001 |
Machine CPU | 256 | 8 | 32 | 8 | ReLU | 10 |
Auto MPG | 32 | 4 | 16 | 8 | ReLU | 0.1 |
Delta Ailer. | 256 | 1 | 32 | 4 | ReLU | 0.001 |
Bank | 256 | 1 | 16 | 8 | ReLU | 0.01 |
California H. | 128 | 4 | 32 | 8 | ReLU | 0.01 |