TABLE 1. Hyperparameters of AECOPD Models.
| Model | Parameter | Value |
|---|---|---|
| Random forest | n_estimators | 300 |
| min_samples_split | 4 | |
| max_depth | 30 | |
| Decision tree | min_samples_leaf | 1 |
| min_samples_split | 2 | |
| Linear discriminant analysis | solver | Lsqr |
| shrinkage | auto | |
| AdaBoost | n_estimators | 45 |
| learning_rate | 1 | |
| Deep neural network | hidden layers | (45, 45, 45) |
| class_weights | 0:10, 1:50 | |
| Adam beta1, beta2 | 0.7, 0.8 |