matflow.param_classes.surrogate.Surrogate

class matflow.param_classes.surrogate.Surrogate(X, y, parameter_names=None, model=None, scale=True)

Bases: ParameterValue

Methods

build_model
dump_element_group_to_HDF5_group	Write a list (from an element group) of parameter values to an HDF5 group.
dump_to_HDF5_group	Write this parameter value to an HDF5 group.
fit
generate_sobol	Wrapper function for scipy.stats.sobol_indices
get_unscaled_data
make_prediction
make_prediction_sobol
perform_inference
prepare_JSON_dump	Prepare this parameter value for serialisation as JSON.
save_from_HDF5_group	Extract a parameter value from an HDF5 group.
save_from_JSON	Extract a parameter value from JSON data.
scale_data
to_dict	Convert to a serialisable dictionary.

Attributes

pytensor_config_cxx

Path to the c++ compiler that Pytensor should use.

build_model(kernel='matern', cross_validate=False, cross_validator=5, scoring='neg_mean_absolute_error', **kwargs)

classmethod dump_element_group_to_HDF5_group(objs, group)

Write a list (from an element group) of parameter values to an HDF5 group.

Parameters:

• objs (list[Self]) –

• group (HDF5Group) –

dump_to_HDF5_group(group)

Write this parameter value to an HDF5 group.

Parameters:

group (HDF5Group) –

fit(Y_actual, Y_error, use_std=True, **kwargs)

generate_sobol(use_fit=False, **kwargs)

Wrapper function for scipy.stats.sobol_indices

Parameters:

• n (int) – number of samples

• **kwargs – additonal parameters

Returns:

SobolResult

Return type:

sobol

get_unscaled_data(unscale_X=True, unscale_y=True)

make_prediction(X, return_std=False, scalar_output=False)

make_prediction_sobol(X)

perform_inference(Y_actual, Y_error, initval=None, use_std=True, **kwargs)

prepare_JSON_dump()

Prepare this parameter value for serialisation as JSON.

Return type:

dict[str, Any]

pytensor_config_cxx: ClassVar[str | None] = None

Path to the c++ compiler that Pytensor should use. If setting, this should be done before an instance is created.

classmethod save_from_HDF5_group(group, param_id, workflow)

Extract a parameter value from an HDF5 group.

Parameters:

• group (HDF5Group) –

• param_id (int) –

• workflow (Workflow) –

classmethod save_from_JSON(data, param_id, workflow)

Extract a parameter value from JSON data.

Parameters:

• param_id (int | list[int]) –

• workflow (Workflow) –

scale_data(scale_X=True, scale_y=True)

to_dict()

Convert to a serialisable dictionary.