Navigation
| class NoiseScale( HyperParameter ) | Source |
|---|
Hyperparameter for the scale of a ScaledErrorDistribution
it is a measure of the noise.
Information about the scale of the noise is stored in his class. It is either in the form of a fixed number, when the noise scale is known or in the form of a Prior with limits. By default this prior is a JeffreysPrior..
The full use of priors is reserved for Bayesian calculations as in NestedSampler
Attributes
- scale : float
the value of the noiseScale. Default: 1.0 - stdev : float
the standard deviation of the noise scale. Default: None - prior : Prior
the prior for the noiseScale. Default: JeffreysPrior - fixed : boolean
keep the noise scale fixed at the value given by scale.
default: True - minimum : boolean
automatic noise scaling with a minimum. default: False
| NoiseScale( scale=1.0, isFixed=True, prior=None, limits=None, copy=None ) |
|---|
Constructor.
Parameters
- scale : float
float value of the noise scale - isFixed : bool
True: Consider the hyperparameter as fixed
False: Optimize the parameter too (when relevant)
It might need a prior and/or limits to be set
The default prior is JeffreysPrior - prior : None or Prior
None : no prior set
Prior : the prior probability on scale - limits : None or list of 2 floats
None : no limits set
[lo,hi] : limits to be passed to the Prior.
If limits are set, the default for Prior is JeffreysPrior - copy : NoiseScale
NoiseScale to copy
| copy( ) |
|---|
| minimumScale( scale=None ) |
|---|
Parameters
- scale : float
the value of the noise scale. Default: noiseScale.scale
| Methods inherited from HyperParameter |
|---|