pmrf.fitting.BayesianResults

class pmrf.fitting.BayesianResults(measured=None, initial_model=None, fitted_model=None, solver_results=None, settings=None)

Bases: FitResults

Abstract base class for results obtained from a Bayesian fitting process.

This class extends FitResults to include methods specific to posterior sampling and distribution training.

Parameters:

• measured (Network | NetworkCollection | None)

• initial_model (Model | None)

• fitted_model (Model | None)

• solver_results (Model)

• settings (FitSettings | None)

__init__(measured=None, initial_model=None, fitted_model=None, solver_results=None, settings=None)

Parameters:

• measured (Network | NetworkCollection | None)

• initial_model (Model | None)

• fitted_model (Model | None)

• solver_results (Model | None)

• settings (FitSettings | None)

Return type:

None

Methods

__delattr__(name, /)	Implement delattr(self, name).
__dir__()	Default dir() implementation.
__eq__(other)	Return self==value.
__format__(format_spec, /)	Default object formatter.
__ge__(value, /)	Return self>=value.
__getattribute__(name, /)	Return getattr(self, name).
__gt__(value, /)	Return self>value.
__init__([measured, initial_model, ...])
__init_subclass__	This method is called when a class is subclassed.
__le__(value, /)	Return self<=value.
__lt__(value, /)	Return self<value.
__ne__(value, /)	Return self!=value.
__new__(**kwargs)
__reduce__()	Helper for pickle.
__reduce_ex__(protocol, /)	Helper for pickle.
__repr__()	Return repr(self).
__setattr__(name, value, /)	Implement setattr(self, name, value).
__sizeof__()	Size of object in memory, in bytes.
__str__()	Return str(self).
__subclasshook__	Abstract classes can override this to customize issubclass().
_decode_recursive(group)
_encode_recursive(obj, group)
_group_to_dict(group)
_read_from_group(group)	Internal driver to load full object state.
_read_network(group)
_read_settings(group)
_save_dict_to_group(d, group)
_write_network(group, ntwk)
_write_settings(group)
_write_to_group(group[, metadata])	Internal driver to save full object state.
decode_solver_results(group)	Decode solver results from an HDF5 group.
encode_solver_results(group)	Encode solver results into an HDF5 group.
fit_posterior([train_dist, equal_weights, ...])	Fit a trainable distribution to the posterior samples.
load_hdf(path)	Load fit results from an HDF5 file.
plot_s_db([use_initial_model])	Plots the S-parameters (Magnitude in dB) of the Measured vs Fitted data.
posterior_samples([equal_weights])	Retrieve samples drawn from the posterior distribution.
prior_samples([equal_weights])	Retrieve samples drawn from the prior distribution.
save_hdf(path[, metadata])	Save the fit results to an HDF5 file.
weights()	Retrieve the weights associated with the posterior samples.

Attributes

__annotations__
__dataclass_fields__
__dataclass_params__
__dict__
__doc__
__hash__
__match_args__
__module__
__weakref__	list of weak references to the object (if defined)
fitted_model
initial_model
measured
settings
solver_results

abstract prior_samples(equal_weights=False)

Retrieve samples drawn from the prior distribution.

Parameters:

equal_weights (bool, optional, default=False) – If True, returns unweighted (resampled) samples.

Returns:

The array of prior samples.

Return type:

jnp.ndarray

abstract posterior_samples(equal_weights=False)

Retrieve samples drawn from the posterior distribution.

Parameters:

equal_weights (bool, optional, default=False) – If True, returns unweighted (resampled) samples.

Returns:

The array of posterior samples.

Return type:

jnp.ndarray

abstract weights()

Retrieve the weights associated with the posterior samples.

Returns:

Array of sample weights.

Return type:

jnp.ndarray

fit_posterior(train_dist=None, equal_weights=False, drift_sigma=0.0, boost_method=None, boost_samples=10000, **train_kwargs)

Fit a trainable distribution to the posterior samples.

Parameters:

• train_dist (TrainableDistributionT or None, optional) – The distribution class to train. If None, defaults to MargarineMAFDistribution.

• equal_weights (bool, optional, default=False) – If True, uses equal weights for training; otherwise uses sample weights.

• drift_sigma (float, optional, default=0.0) – Standard deviation for drift augmentation to broaden the posterior support.

• boost_method (str or None, optional) – Method to boost sample count (‘kde’ or None).

• boost_samples (int, optional, default=10000) – Number of samples to generate if boosting is enabled.

• **train_kwargs – Additional keyword arguments passed to the distribution’s training method.