Package jdistlib.inference

Class AdjustedMicrocanonicalLangevin

java.lang.Object

jdistlib.inference.AdjustedMicrocanonicalLangevin

All Implemented Interfaces:

Sampler

public final class AdjustedMicrocanonicalLangevin
extends Object
implements Sampler

Metropolis-adjusted microcanonical Langevin sampler (MHMCHMC).

This is the adjusted, asymptotically exact counterpart of MCLMC. It uses the reversible isokinetic McLachlan splitting and a unit-norm momentum. The method requires at least two dimensions.

Constructor Summary

Constructors

Constructor	Description
AdjustedMicrocanonicalLangevin()

Method Summary

Modifier and Type	Method	Description
ChainResult	sample(LogDensity target, double[] initialState, SamplingOptions options, RandomEngine random)

Methods inherited from class java.lang.Object

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

Constructor Details

AdjustedMicrocanonicalLangevin

public AdjustedMicrocanonicalLangevin()

Method Details

sample

public ChainResult sample(LogDensity target,
 double[] initialState,
 SamplingOptions options,
 RandomEngine random)

Specified by:

sample in interface Sampler