GWForge.population_fisher.fit

Maximum-likelihood fits of a population model to a set of samples.

Attributes

SIMPLEX_PARAMETERS
FIT_OPTIONS
OUT_OF_SUPPORT_PENALTY
DEFAULT_BOUNDS

Functions

fit_model(model, samples[, free_parameters, start, ...])	Fit a population model to samples by maximum likelihood.
compare_to(fitted, reference, names)	A plain-text table of fitted values against the values used to generate.

Module Contents

GWForge.population_fisher.fit.SIMPLEX_PARAMETERS = ('lam_0', 'lam_1')

GWForge.population_fisher.fit.FIT_OPTIONS

GWForge.population_fisher.fit.OUT_OF_SUPPORT_PENALTY = 100000000.0

GWForge.population_fisher.fit.DEFAULT_BOUNDS

GWForge.population_fisher.fit.fit_model(model, samples, free_parameters=None, start=None, bounds=None, fixed_parameters=None)

Fit a population model to samples by maximum likelihood.

Parameters:

• model (GWForge.population_fisher.base.PopulationModel)

• samples (dict) – Event parameters to fit to. For a fiducial, these should be all injections, not just the detected ones: the population model describes the astrophysical population, and fitting to the detected subset would absorb the selection function into the hyper-parameters.

• free_parameters (sequence of str or None) – Defaults to every hyper-parameter not in fixed_parameters.

• start (dict or None) – Starting point. Defaults to model.fiducial.

• bounds (dict or None) – Name to (low, high). Merged onto DEFAULT_BOUNDS.

• fixed_parameters (dict or None)

Returns:

The fitted values, merged onto the starting point, with the keys "_success", "_message" and "_mean_log_prob" describing the optimisation.

Return type:

dict

GWForge.population_fisher.fit.compare_to(fitted, reference, names)

A plain-text table of fitted values against the values used to generate.

Parameters:

• fitted (dict)

• reference (dict)

• names (sequence of str)

Return type:

str