nessai.stopping_criteria#

Stopping criteria for nested sampling.

This module defines various stopping criteria for nested sampling algorithms.

Classes#

`StoppingCriterionRegistry`	Registry for stopping criteria.
`StoppingCriterion`	Class for defining stopping criteria.
`ESS`	Effective sample size (ESS) stopping criterion.
`LogEvidenceRatio`	Log ratio of evidence between live points and all samples.
`LogEvidenceRatioNestedSamples`	Log ratio of evidence between live point and nested samples.
`EvidenceError`	Evidence error stopping criterion.
`DifferenceLogEvidence`	Difference in log evidence stopping criterion.
`FractionalError`	Fractional error stopping criterion.

Module Contents#

class nessai.stopping_criteria.StoppingCriterionRegistry#

Registry for stopping criteria.

This class is used to register and retrieve stopping criteria classes. It allows for easy extension and management of different stopping criteria.

class nessai.stopping_criteria.StoppingCriterion#

Class for defining stopping criteria.

Parameters:

namestr

Name of the stopping criterion.

tolerancefloat

Tolerance value for the stopping criterion.

comparisonstr

Comparison operator for the stopping criterion.: Valid options are: ‘<’, ‘>’, ‘<=’, ‘>=’, ‘==’, ‘!=’.

is_met(value: float | int) → bool#

Check if the stopping criterion is met.

Parameters:

valuefloat or int: The value to compare against the stopping criterion.

Returns:

bool: True if the stopping criterion is met, False otherwise.

class nessai.stopping_criteria.ESS(tolerance: float = 5000.0)#

Bases: StoppingCriterion

Effective sample size (ESS) stopping criterion.

Parameters:

tolerancefloat: Tolerance value for the stopping criterion.

class nessai.stopping_criteria.LogEvidenceRatio(tolerance: float = 0.0)#

Bases: StoppingCriterion

Log ratio of evidence between live points and all samples.

Parameters:

tolerancefloat: Tolerance value for the stopping criterion.

class nessai.stopping_criteria.LogEvidenceRatioNestedSamples(tolerance: float = 0.0)#

Bases: StoppingCriterion

Log ratio of evidence between live point and nested samples.

Parameters:

tolerancefloat: Tolerance value for the stopping criterion.

class nessai.stopping_criteria.EvidenceError(tolerance: float = 0.1)#

Bases: StoppingCriterion

Evidence error stopping criterion.

Parameters:

namestr: Name of the stopping criterion.
tolerancefloat: Tolerance value for the stopping criterion.
comparisonstr: Comparison operator for the stopping criterion.

class nessai.stopping_criteria.DifferenceLogEvidence(tolerance: float = 0.1)#

Bases: StoppingCriterion

Difference in log evidence stopping criterion.

This is the standard nested sampling stopping criterion.

Parameters:

tolerancefloat: Tolerance value for the stopping criterion.

class nessai.stopping_criteria.FractionalError(tolerance: float = 0.1)#

Bases: StoppingCriterion

Fractional error stopping criterion.

Parameters:

tolerancefloat: Tolerance value for the stopping criterion.