Table 3.
Selected hyper-parameters for various ML classifiers
| Classifier | Hyper-parameter(s) | Value |
|---|---|---|
| SVM | Regularization Constant (C) | 1.0 |
| Decision function shape | One versus rest | |
| Gamma | Scale | |
| Kernel | Radial basis function | |
| Tolerance | 0.001 | |
| LogisticRegression | regularization constant (C) | 1.0 |
| Penalty | l2 | |
| Tolerance | 0.0001 | |
| KNeighborsClassifier | Metric | Minkowski Distance |
| # of Neighbors | 3 | |
| Weights | Uniform | |
| DecisionTreeClassifier | Criterion | Gini |
| Max Depth | 5 | |
| Min Samples Leaf | 1 | |
| Min Samples Split | 2 | |
| Splitter | Best | |
| RandomForestClassifier | Bootstrap | True |
| Criterion | Gini | |
| Max Depth | 20 | |
| Max Features | 10 | |
| N Estimators | 1000 | |
| XGBClassifier | Base Score | 0.97 |
| Booster | Gradient Boosting Tree | |
| Max Depth | 3 | |
| N Estimators | 1000 | |
| Objective | Logistic Regression |