Model | Hyperparameters | Range | Optimal values |
---|---|---|---|
Logistic regression | penalty | [l1, l2] | l2 |
Cs | [0.001, 0.1, 1, 100, 1000] | 1 | |
Extra trees | min_samples_leaf | [5, 8, 10] | 5 |
criterion | [gini, entropy, log_loss] | entropy | |
max_features | [sqrt, log2, none] | sqrt | |
Random forest | max_depth | [3, 5, 10] | 10 |
min_samples_split | [2, 5, 10] | 5 | |
GBDT | learning_rate | [0.01, 0.1, 0.2] | 0.2 |
max_depth | [3, 5, 8] | 8 | |
n_estimators | [10, 20] | 20 | |
XGBoost | gamma | [0.5, 1, 5] | 0.5 |
colsample_bytree | [0.6, 0.8, 1.0] | 1.0 | |
max_depth | [3, 4, 5] | 5 | |
LGBM | n_estimators | [8, 16, 24] | 24 |
num_leaves | [6, 12, 16] | 16 | |
max_bin | [255, 510] | 510 |