io.improbable.keanu.algorithms.mcmc

Class Hamiltonian

java.lang.Object

io.improbable.keanu.algorithms.mcmc.Hamiltonian

All Implemented Interfaces:

PosteriorSamplingAlgorithm

public class Hamiltonian
extends java.lang.Object
implements PosteriorSamplingAlgorithm

Hamiltonian Monte Carlo is a method for obtaining samples from a probability distribution with the introduction of a momentum variable.

Algorithm 1: "Hamiltonian Monte Carlo". The No-U-Turn Sampler: Adaptively Setting Path Lengths in Hamiltonian Monte Carlo https://arxiv.org/pdf/1111.4246.pdf

Constructor Summary

Constructors

Constructor and Description
Hamiltonian()

Method Summary

Modifier and Type	Method and Description
NetworkSamples	getPosteriorSamples(BayesianNetwork bayesNet, java.util.List<? extends Vertex> fromVertices, int sampleCount) Sample from the posterior of a Bayesian Network using the Hamiltonian Monte Carlo algorithm
static Hamiltonian	withDefaultConfig()
static Hamiltonian	withDefaultConfig(KeanuRandom random)

Methods inherited from class java.lang.Object

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

Methods inherited from interface io.improbable.keanu.algorithms.PosteriorSamplingAlgorithm

getPosteriorSamples

Constructor Detail

Hamiltonian

public Hamiltonian()

Method Detail

withDefaultConfig

public static Hamiltonian withDefaultConfig()

withDefaultConfig

public static Hamiltonian withDefaultConfig(KeanuRandom random)

getPosteriorSamples

public NetworkSamples getPosteriorSamples(BayesianNetwork bayesNet,
                                          java.util.List<? extends Vertex> fromVertices,
                                          int sampleCount)

Sample from the posterior of a Bayesian Network using the Hamiltonian Monte Carlo algorithm

Specified by:

getPosteriorSamples in interface PosteriorSamplingAlgorithm

Parameters:

bayesNet - The bayesian network to sample from

fromVertices - the vertices to sample from

sampleCount - the number of samples to take

Returns:

Samples taken with Hamiltonian Monte Carlo