Table 8 Hyperparameter tuning using Hyperopt
Regression method | Hyperparameter | Search space |
|---|---|---|
MLR | - | - |
Lasso | alpha | 10–3, 10–2, 10–1,…..,103 |
Ridge | alpha | 10–3, 10–2, 10–1,…..,103 |
Elastic Net | alpha | 10–3, 10–2, 10–1,…..,103 |
KNN | n_neighbors | 1, 2, 3,….., 11 |
DT | max_depth | 1, 2, 3, ….., 9 |
RF | max_depth n_estimators | 1, 2, 3, ….., 9 50, 55, 60, ….., 195 |
GBR | learning_rate max_depth | 10–5, 10–4, 10–3, …..,1 1, 2, 3, ….., 9 |
ADABoost | learning_rate | 10–5, 10–4, 10–3, …..,1 |
XGBoost | learning_rate max_depth | 10–5, 10–4, 10–3, …..,1 1, 2, 3, ….., 9 |
LightGBM | learning_rate max_depth | 10–5, 10–4, 10–3, …..,1 1, 2, 3, ….., 9 |
CatBoost | learning_rate | 10–5, 10–4, 10–3, …..,1 |
LSTM | no. of units learning_rate | 10, 20, 30, ….., 100 10–5, 10–4, 10–3, …..,1 |
GRU | no. of units learning_rate | 10, 20, 30, ….., 100 10–5, 10–4, 10–3, …..,1 |