autoprognosis.plugins.ensemble.risk_estimation module

class RiskEnsemble(models: ~typing.List, weights: ~numpy.ndarray, time_horizons: ~typing.List, explainer_plugins: ~typing.List = [], explanations_model: ~typing.Dict | None = None, explanations_nepoch: int = 10000, hooks: ~autoprognosis.hooks.base.Hooks = <autoprognosis.hooks.default.DefaultHooks object>)

Bases: object

Weighted risk ensemble.

Parameters:

  • models – List [N] List of base models.

  • weights – List (time_horizons|, N) list of weights for each model and each time horizon.

  • time_horizons – List List of time horizons used for evaluation.

  • explainer_plugins – List List of explainers attached to the ensemble.

enable_explainer(explainer_plugins: list = [], explanations_nepoch: int = 10000) → None
explain(X: DataFrame, *args: Any, **kwargs: Any) → DataFrame
fit(X: DataFrame, T: DataFrame, Y: DataFrame) → RiskEnsemble
is_fitted() → bool
name(short: bool = False) → str
predict(X_: DataFrame, eval_time_horizons: DataFrame | None = None) → DataFrame
class RiskEnsembleCV(time_horizons: ~typing.List, ensemble: ~autoprognosis.plugins.ensemble.risk_estimation.RiskEnsemble | None = None, models: ~typing.List | None = None, weights: ~typing.List[float] | None = None, explainer_plugins: ~typing.List = [], explanations_model: ~typing.Dict | None = None, explanations_nepoch: int = 10000, n_folds: int = 3, hooks: ~autoprognosis.hooks.base.Hooks = <autoprognosis.hooks.default.DefaultHooks object>)

Bases: RiskEnsemble

Cross-validated Weighted ensemble, with uncertainity prediction support

Parameters:

  • ensemble – List Base ensembles. Excludes models/weights

  • models – List [N] List of base models.

  • weights – List (time_horizons, N) list of weights for each model and each time horizon.

  • time_horizons – List List of time horizons used for evaluation.

  • explainer_plugins – List List of explainers attached to the ensemble.

enable_explainer(explainer_plugins: list = [], explanations_nepoch: int = 10000) → None
explain(X: DataFrame, *args: Any, **kwargs: Any) → DataFrame
fit(X: DataFrame, T: DataFrame, Y: DataFrame) → RiskEnsemble
is_fitted() → bool
name(short: bool = False) → str
predict(X_: DataFrame, eval_time_horizons: DataFrame | None = None) → DataFrame
predict_with_uncertainty(X_: DataFrame, eval_time_horizons: DataFrame | None = None) → DataFrame