XGBoostHyperParams#
- class openstef_models.models.forecasting.xgboost_forecaster.XGBoostHyperParams(**data: Any) None[source]
Bases:
HyperParamsXGBoost hyperparameters for gradient boosting tree models.
Configures tree-specific parameters for XGBoost gbtree booster. Provides control over model complexity, regularization, and training behavior for energy forecasting tasks.
These parameters control tree structure, learning rates, regularization, and sampling strategies. Proper tuning is essential for balancing model performance and overfitting prevention in time series forecasting.
Example
Creating custom hyperparameters for deep trees with regularization
>>> hyperparams = XGBoostHyperParams( ... n_estimators=200, ... max_depth=8, ... learning_rate=0.1, ... reg_alpha=0.1, ... reg_lambda=1.0, ... subsample=0.8, ... )
Note
These parameters are optimized for probabilistic forecasting with quantile regression. The default objective function is specialized for magnitude-weighted pinball loss.
- Parameters:
data (
Any)
-
n_estimators:
Annotated[int]
-
learning_rate:
Annotated[float]
-
max_depth:
Annotated[int]
-
min_child_weight:
Annotated[float]
-
gamma:
Annotated[float]
-
objective:
TypeAliasType
-
evaluation_metric:
TypeAliasType
-
reg_alpha:
Annotated[float]
-
reg_lambda:
Annotated[float]
-
max_delta_step:
float
-
max_leaves:
int
-
grow_policy:
Annotated[Literal['depthwise','lossguide']]
-
max_bin:
int
-
num_parallel_trees:
int
-
subsample:
Annotated[float]
-
colsample_bytree:
Annotated[float]
-
colsample_bylevel:
float
-
colsample_bynode:
float
-
tree_method:
Annotated[Literal['auto','exact','hist','approx','gpu_hist']]
-
use_target_scaling:
bool
- classmethod forecaster_class() type[XGBoostForecaster][source]
Get the forecaster class for these hyperparams.
- Return type:
type[XGBoostForecaster]- Returns:
Forecaster class associated with this configuration.
- model_config: ClassVar[ConfigDict] = {'arbitrary_types_allowed': False, 'extra': 'ignore', 'protected_namespaces': ()}
Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].