Gradient boosting¶
| Field | Type | Orange's equivalent | Python's default | Orange's default |
|---|---|---|---|---|
| - | - | Method. | - | GBLearner |
n_estimators |
list(int) | Number of trees. | 100 |
100 |
learning_rate |
list(float) | Learning rate. | 0.1 |
0.1 |
random_state |
list(int | null) | Replicable training. | None |
0 |
max_depth |
list(int) | Limit depth of individual trees. | 3 |
3 |
min_samples_split |
list(int) | Do not split subsets smaller than. | 2 |
2 |
subsample |
list(float) | Fraction of training instances. | 1.0 |
1 |
NOTE: This table describes the standard
GBLearner. Other methods, such asXGBLearner,XGBRFLearner, andCatGBLearner, are selected through different learner classes and may have different default parameters.
For reference: Orange/widgets/model/owgradientboosting.py.