Table 9 Hyperparameter set-up for the best found ELM models for each classification dataset
From: Extreme learning machine versus classical feedforward network
Dataset | Hidden layer size | Activation function | Regularisation coefficient |
|---|---|---|---|
Breast cancer (Wisconsin) | 3000 | ReLU | 10 |
Breast cancer (Ljubljana) | 300 | Threshold | 10 |
Wine | 100 | Sigmoid | 0.1 |
Urban land cover | 10,000 | Threshold | 0.1 |
Sonar | 1000 | Threshold | 0.1 |
Glass | 1000 | Tanh | 0.01 |
Landsat | 10,000 | ReLU | 0.3 |
CNAE-9 | 10,000 | Tanh | 10 |
Forest cover type | 3000 | ReLU | 1 |