#include <np_mcmc.h>

Public Member Functions
	MCMC (std::default_random_engine &generator, InitClusters &init_clusters, UpdateClusters &update_clusters, UpdateClusterPopulation &update_cluster_population, int subset_count, distribution_t &likelihood)

void	run (dataset_t &dataset, int T)

const membertrix &	getMembershipMatrix () const

const membertrix &	getMaxLikelihoodMatrix () const

Protected Member Functions
void	considerMaxLikelihood ()

Detailed Description

The MCMC class can be equiped with InitCluster, UpdateCluster, and UpdateClusterPopulation objects that each will be used in a Monte-Carlo Monte Chain algorithm. This algorithm is run for T steps and returns a membership matrix in the form of a membertrix object.

Constructor & Destructor Documentation

◆ MCMC()

MCMC::MCMC	(	std::default_random_engine &	generator,
		InitClusters &	init_clusters,
		UpdateClusters &	update_clusters,
		UpdateClusterPopulation &	update_cluster_population,
		int	subset_count,
		distribution_t &	likelihood
	)

Constructor for MCMC.

Parameters

[in]	generator	A random number generator
[in]	init_clusters	An object that initializes clusters
[in]	update_clusters	An object that updates the cluster parameters
[in]	update_cluster_population	An object that generates and removes clusters

Member Function Documentation

◆ considerMaxLikelihood()

void MCMC::considerMaxLikelihood ( )

protected

Calculate maximum likelihood of current assignments

[out] _max_likelihood_membertrix The max likelihood membership matrix up to now [out] _max_likelihood The max likelihood up to now

◆ getMaxLikelihoodMatrix()

const membertrix & MCMC::getMaxLikelihoodMatrix ( ) const

◆ getMembershipMatrix()

const membertrix & MCMC::getMembershipMatrix ( ) const

The results of the MCMC algorithm is a membership matrix.

Returns: The membership matrix (const)

◆ run()

void MCMC::run	(	dataset_t &	dataset,
		int	T
	)

Run the MCMC method (can be one or multiple chains) for T steps.