Table 3.
Bayesian optimisation search space for the machine learning model parameters
| Model | Parameter | Search space | |||
|---|---|---|---|---|---|
| Lower limit | Upper limit | Type | Scale | ||
| Logistic regression | C (regularisation strength) | 0·02 | 1 | Decimal | Logarithmic |
| Random forest | Number of estimators | 100 | 5000 | Integer | Linear |
| Maximum depth | 1 | 3 | Integer | Linear | |
| Minimum samples leaf | 200 | 500 | Integer | Linear | |
| Maximum features | 2 | 5 | Integer | Linear | |
| Gradient-boosting tree | Number of estimators | 300 | 800 | Integer | Linear |
| Learning rate | 0·001 | 0·01 | Float | Logarithmic | |
| Maximum features | 2 | 5 | Decimal | Linear | |
| Maximum depth | 1 | 4 | Decimal | Linear | |