Class StableSampler.BaseStableSampler

java.lang.Object

org.apache.commons.rng.sampling.distribution.StableSampler

org.apache.commons.rng.sampling.distribution.StableSampler.BaseStableSampler

All Implemented Interfaces:

ContinuousSampler, SharedStateContinuousSampler, SharedStateSampler<SharedStateContinuousSampler>

Direct Known Subclasses:

StableSampler.Alpha1CMSStableSampler, StableSampler.Beta0WeronStableSampler, StableSampler.CauchyStableSampler, StableSampler.WeronStableSampler

Enclosing class:

StableSampler

private abstract static class StableSampler.BaseStableSampler
extends StableSampler

Base class for implementations of a stable distribution that requires an exponential random deviate.

Since:

1.4

Nested Class Summary

Nested classes/interfaces inherited from class StableSampler

StableSampler.Alpha1CMSStableSampler, StableSampler.Beta0CMSStableSampler, StableSampler.Beta0WeronStableSampler, StableSampler.CMSStableSampler, StableSampler.SpecialMath, StableSampler.WeronStableSampler

Field Summary

Fields

Modifier and Type	Field	Description
private final ContinuousSampler	expSampler	The exponential sampler.
private static final double	NEG_PI_2	-pi / 2.
private static final double	NEG_PI_4	-pi / 4.
private static final double	PI_2_SCALED	pi/2 scaled by 2^-53.
private static final double	PI_4_SCALED	pi/4 scaled by 2^-53.

Constructor Summary

Constructors

Constructor	Description
BaseStableSampler(UniformRandomProvider rng)

Method Summary

Modifier and Type	Method	Description
(package private) double	getOmega()	Gets a random value for the omega parameter (w).
(package private) double	getPhi()	Gets a random value for the phi parameter.
(package private) double	getPhiBy2()	Gets a random value for the phi parameter divided by 2.

Methods inherited from class StableSampler

nextLong, of, of, sample, toString, withUniformRandomProvider

Methods inherited from class Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, wait, wait, wait

Methods inherited from interface ContinuousSampler

samples, samples

Field Details

PI_2_SCALED

private static final double PI_2_SCALED

pi/2 scaled by 2^-53.

See Also:

• Constant Field Values

PI_4_SCALED

private static final double PI_4_SCALED

pi/4 scaled by 2^-53.

See Also:

• Constant Field Values

NEG_PI_2

private static final double NEG_PI_2

-pi / 2.

See Also:

• Constant Field Values

NEG_PI_4

private static final double NEG_PI_4

-pi / 4.

See Also:

• Constant Field Values

expSampler

private final ContinuousSampler expSampler

The exponential sampler.

Constructor Details

BaseStableSampler

BaseStableSampler(UniformRandomProvider rng)

Parameters:

rng - Underlying source of randomness

Method Details

getOmega

double getOmega()

Gets a random value for the omega parameter (w). This is an exponential random variable with mean 1.

Warning: For simplicity this does not check the variate is not 0. The calling CMS algorithm should detect and handle incorrect samples as a result of this unlikely edge case.

Returns:

omega

getPhi

double getPhi()

Gets a random value for the phi parameter. This is a uniform random variable in (-pi/2, pi/2).

Returns:

phi

getPhiBy2

double getPhiBy2()

Gets a random value for the phi parameter divided by 2. This is a uniform random variable in (-pi/4, pi/4).

Note: Ideally this should not have a value of -pi/4 or pi/4 as the CMS algorithm can generate infinite values when the phi/2 uniform deviate is +/-pi/4. This can result in NaN output. Under certain parameterizations phi/2 close to the limits also create NaN output. Thus output should be checked regardless. Avoiding the extreme values simplifies the number of checks that are required.

Returns:

phi / 2