Table 2 Optimal hyperparameters for each model, along with their investigated range and cross-validation scores.
From: Ensemble learning for prediction of inorganic scale formation: A case study in Oman
Model | Hyperparameter | Optimal value | Range | Cross-validation score |
|---|---|---|---|---|
LR | C | 2 | 1–100 | 76.6% |
Fit intercept | True | True, False | ||
Solver | Saga | lbfgs, Sag, Newton | ||
Penalty norm | L2 | L1, L2 | ||
Max iterations | 200 | 100, 200, 300, 400 | ||
RF | Criterion | Entropy | Entropy, Gini | 88.3% |
Max_depth | 5 | 2–10 | ||
n_estimators | 100 | 50–500 | ||
Min_samples_split | 3 | 2–20 | ||
Min_samples_leaf | 1 | 1–5 | ||
Max_features | Auto | Sqrt, log2, Auto | ||
oob_score | True | True, False | ||
DT | Criterion | Entropy | Entropy, Gini | 81.4% |
Min_samples_split | 100 | 2–150 | ||
Min_samples_leaf | 1 | 1–5 | ||
Max_depth | 6 | 2–10 | ||
SVM | Kernel | Rbf | Linear, Poly, Rbf | 79.4% |
Probability | True | True, False | ||
C | 1 | 0.001, 0.01, 0.1, 1, 10 | ||
KNN | n_neighbors | 5 | 1–11 | 84.8% |
Metric | Minkowski | Minkowski, Euclidean, l1, l2 | ||
P | 2 | 1, 2 | ||
NB | Fit_prior | True | True, False | 74.2% |
Class_prior | None | None, True |
