rxmc.transforms.scale#

rxmc.transforms.scale(parameter: Parameter | None = None, log: bool = True, name=None) → Transform[source]#

A latent multiplicative normalisation \(\rho\, y\).

The Kennedy & O’Hagan forward-model scale: it changes the mean, not the covariance, so it belongs on the model (PhysicalModel(params, transform=scale())).

Parameters:

parameter (Parameter, optional) – The scale parameter. Defaults to log_rho (or rho when log=False).
log (bool, optional) – If True (default) the sampled value is \(\log\rho\) and the prediction is scaled by exp(value); otherwise by value.
name (str, optional) – Name for the default parameter.