Estimators

Emukit

Contains the Scikit-Learn wrapper for Emukit

The defined Scikit-Learn wrapper for the Emukit library.

class calisim.estimators.emukit_estimator.EmukitEstimator(method_kwargs: dict | None = None)[source]
fit(X: ndarray, y: ndarray | None = None) → EmukitEstimator[source]

Fit the estimator.

Parameters:

  • X (np.ndarray) – The simulation inputs.

  • y (np.ndarray | None, optional) – The simulation outputs. Defaults to None.

Returns:

The estimator.

Return type:

EmukitEstimator

predict(X: ndarray, return_std: bool = False) → ndarray | tuple[source]

Make a prediction.

Parameters:

  • X (np.ndarray) – The simulation inputs.

  • return_std (bool, optional) – Whether to return the standard deviation. Defaults to False.

Returns:

The model predictions.

Return type:

np.ndarray | tuple

score(X: ndarray, y: ndarray, sample_weight: ndarray | None = None) → float[source]

Assess the estimator.

Parameters:

  • X (np.ndarray) – The simulation inputs.

  • y (np.ndarray) – The simulation outputs.

  • sample_weight (np.ndarray | None, optional) – Weighting factor for the samples. Defaults to None.

Returns:

The assessment score.

Return type:

float

set_predict_request(*, return_std: bool | None | str = '$UNCHANGED$') → EmukitEstimator

Request metadata passed to the predict method.

Note that this method is only relevant if enable_metadata_routing=True (see sklearn.set_config()). Please see User Guide on how the routing mechanism works.

The options for each parameter are:

  • True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

  • False: metadata is not requested and the meta-estimator will not pass it to predict.

  • None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

  • str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Note

This method is only relevant if this estimator is used as a sub-estimator of a meta-estimator, e.g. used inside a Pipeline. Otherwise it has no effect.

Parameters:

return_std (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for return_std parameter in predict.

Returns:

self – The updated object.

Return type:

object

set_score_request(*, sample_weight: bool | None | str = '$UNCHANGED$') → EmukitEstimator

Request metadata passed to the score method.

Note that this method is only relevant if enable_metadata_routing=True (see sklearn.set_config()). Please see User Guide on how the routing mechanism works.

The options for each parameter are:

  • True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

  • False: metadata is not requested and the meta-estimator will not pass it to score.

  • None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

  • str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Note

This method is only relevant if this estimator is used as a sub-estimator of a meta-estimator, e.g. used inside a Pipeline. Otherwise it has no effect.

Parameters:

sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

Returns:

self – The updated object.

Return type:

object

GPyTorch

Contains the Scikit-Learn wrapper for GPyTorch

The defined Scikit-Learn wrapper for the GPyTorch library.

class calisim.estimators.gpytorch_estimator.SingleTaskGPRegressionModel(likelihood: Likelihood, noise_init: float | None = None)[source]

The single task exact Gaussian process.

Parameters:

gpytorch (ExactGP) – The GPyTorch module.

forward(X: Tensor) → MultivariateNormal[source]

The forward pass.

Parameters:

X (torch.Tensor) – The input tensor.

Returns:

The predicted

mean and covariance.

Return type:

gpytorch.distributions.MultivariateNormal

calisim.estimators.gpytorch_estimator.get_single_task_exact_gp(lr: float, max_epochs: int, device: str = 'cpu') → ExactGPRegressor[source]

Get an instance of a single task Gaussian process.

Parameters:

  • lr (float) – The learning rate.

  • max_epochs (int) – The maximum number of epochs.

  • device (str, optional) – The device to train the model. Defaults to “cpu”.

Returns:

The Gaussian process.

Return type:

ExactGPRegressor

OpenTurns

Contains the Scikit-Learn wrappers for OpenTurns

The defined Scikit-Learn wrappers for the OpenTurns library.

class calisim.estimators.openturns_estimator.FunctionalChaosEstimator(parameters: JointDistribution, total_degree: int = 2)[source]
fit(X: ndarray, y: ndarray | None = None) → FunctionalChaosEstimator[source]

Fit the estimator.

Parameters:

  • X (np.ndarray) – The simulation inputs.

  • y (np.ndarray | None, optional) – The simulation outputs. Defaults to None.

Returns:

The estimator.

Return type:

FunctionalChaosEstimator

predict(X: ndarray) → ndarray[source]

Make a prediction.

Parameters:

X (np.ndarray) – The simulation inputs.

Returns:

The model predictions.

Return type:

np.ndarray

score(X: ndarray, y: ndarray, sample_weight: ndarray | None = None) → float[source]

Assess the estimator.

Parameters:

  • X (np.ndarray) – The simulation inputs.

  • y (np.ndarray) – The simulation outputs.

  • sample_weight (np.ndarray | None, optional) – Weighting factor for the samples. Defaults to None.

Returns:

The assessment score.

Return type:

float

set_score_request(*, sample_weight: bool | None | str = '$UNCHANGED$') → FunctionalChaosEstimator

Request metadata passed to the score method.

Note that this method is only relevant if enable_metadata_routing=True (see sklearn.set_config()). Please see User Guide on how the routing mechanism works.

The options for each parameter are:

  • True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

  • False: metadata is not requested and the meta-estimator will not pass it to score.

  • None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

  • str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Note

This method is only relevant if this estimator is used as a sub-estimator of a meta-estimator, e.g. used inside a Pipeline. Otherwise it has no effect.

Parameters:

sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

Returns:

self – The updated object.

Return type:

object

class calisim.estimators.openturns_estimator.KrigingEstimator(parameters: JointDistribution, basis: str = 'constant', covariance: str = 'SquaredExponential', covariance_scale: float = 1.0, covariance_amplitude: float = 1.0, n_out: int = 1)[source]
fit(X: ndarray, y: ndarray | None = None) → KrigingEstimator[source]

Fit the estimator.

Parameters:

  • X (np.ndarray) – The simulation inputs.

  • y (np.ndarray | None, optional) – The simulation outputs. Defaults to None.

Returns:

The estimator.

Return type:

KrigingEstimator

predict(X: ndarray) → ndarray | tuple[source]

Make a prediction.

Parameters:

X (np.ndarray) – The simulation inputs.

Returns:

The model predictions.

Return type:

np.ndarray | tuple

score(X: ndarray, y: ndarray, sample_weight: ndarray | None = None) → float[source]

Assess the estimator.

Parameters:

  • X (np.ndarray) – The simulation inputs.

  • y (np.ndarray) – The simulation outputs.

  • sample_weight (np.ndarray | None, optional) – Weighting factor for the samples. Defaults to None.

Returns:

The assessment score.

Return type:

float

set_score_request(*, sample_weight: bool | None | str = '$UNCHANGED$') → KrigingEstimator

Request metadata passed to the score method.

Note that this method is only relevant if enable_metadata_routing=True (see sklearn.set_config()). Please see User Guide on how the routing mechanism works.

The options for each parameter are:

  • True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

  • False: metadata is not requested and the meta-estimator will not pass it to score.

  • None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

  • str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Note

This method is only relevant if this estimator is used as a sub-estimator of a meta-estimator, e.g. used inside a Pipeline. Otherwise it has no effect.

Parameters:

sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

Returns:

self – The updated object.

Return type:

object