inversion_ideas.hyperparams.SensitivityWeights#

class inversion_ideas.hyperparams.SensitivityWeights(simulation, initial_model, *, data_weights=None, volumes=None, vmin=1e-12)#

Updatable sensitivity weights.

This class wraps a sensitivity weights array that can be updated by calling the update() method, passing a given model as argument.

Parameters:

simulationSimulation: Simulation used to get the jacobian matrix that will be used while updating the sensitivity weights.
initial_model(n_params) array: Initial model used to initialize the sensitivity weights.
data_weights(n_data,) array or None, optional: Array with data weights used to compute the sensitivty weights. Can use the inversion_ideas.DataMisfit.weights property.
volumes(n_params) array: Array with the volumes of the active cells. Sensitivity weights are divided by the volumes to account for sensitivity changes due to cell sizes.
vminfloat or None, optional: Minimum value used for clipping.

Attributes

`T`	Transpose.
`dtype`	The data-type for the wrapped array.
`initial_model`	Model used to compute the initial sensitivity weights.
`ndim`	Number of dimensions in the wrapped array.
`shape`	Shape of the wrapped array.
`size`	Total number of elements in the wrapped array.

Methods

`max`([axis, out])	Maximum value in the wrapped array along a given axis.
`min`([axis, out])	Minimum value in the wrapped array along a given axis.
`transpose`()	Transpose.
`update`(model, *args)	Update sensitivity weights for the given model.

copy

Attributes#

SensitivityWeights.T#

Transpose.

Returns:

SensitivityWeights.initial_model#: Model used to compute the initial sensitivity weights.

SensitivityWeights.ndim#

Number of dimensions in the wrapped array.

Returns:

SensitivityWeights.shape#

Shape of the wrapped array.

Returns:

SensitivityWeights.size#

Total number of elements in the wrapped array.

Returns:

Methods documentation

SensitivityWeights.max(axis=None, out=None, **kwargs)#

Maximum value in the wrapped array along a given axis.