transforms#

Transformation functions to normalize output responses

Classes

`Identity_Scaler`()	Dummy scaler class which simply returns the input
`Logit_Scaler`([range_response, loc, ...])	Scaler which normalizes based on a logit transform Can be fit to select an appropriate range for the logit
`Standard_Scaler`()	Scaler which normalizes based on zero mean and unit st-dev
`Target_Transform`()	Base class for obsidian Target transforms

class obsidian.parameters.transforms.Identity_Scaler[source]#

Bases: Target_Transform

Dummy scaler class which simply returns the input

forward(X: Tensor, fit: bool = False)[source]#: Evaluate the forward transformation on input data X

inverse(X: Tensor)[source]#: Inverse transform the transformed data X_t

class obsidian.parameters.transforms.Logit_Scaler(range_response: int | float = 1, loc: int | float = 0, override_fit: bool = False, standardize: bool = True)[source]#

Bases: Target_Transform

Scaler which normalizes based on a logit transform Can be fit to select an appropriate range for the logit

forward(X: Tensor, fit: bool = False)[source]#: Evaluate the forward transformation on input data X

inverse(X: Tensor)[source]#: Inverse transform the transformed data X_t

class obsidian.parameters.transforms.Standard_Scaler[source]#

Bases: Target_Transform

Scaler which normalizes based on zero mean and unit st-dev

forward(X: Tensor, fit: bool = False)[source]#: Evaluate the forward transformation on input data X

inverse(X)[source]#: Inverse transform the transformed data X_t

class obsidian.parameters.transforms.Target_Transform[source]#

Bases: ABC

Base class for obsidian Target transforms

abstract forward(X: Tensor, fit: bool = False)[source]#: Evaluate the forward transformation on input data X

abstract inverse(X: Tensor)[source]#: Inverse transform the transformed data X_t