mcmcoutput-class: Finmix 'mcmcoutput' class

In simonsays1980/finmix: An R package for Bayesian estimation of finite mixture distributions

Description

The mcmcoutput class stores all MCMC samples and corresponding information.

Details

Calling mixturemcmc() on appropriate input arguments performs MCMC sampling and returns an mcmcoutput object that stores all samples and corresponding information like hyper-parameters, the finite mixture model specified in a model object and the prior that specifies the prior distribution. All slots are listed below. Note that not all slots must be available in a object of class mcmcoutput. Some slots get only occupied, if a hierarchical prior had been used in MCMC sampling, or if posterior samples should be stored. Furthermore, the slots also look different, if MCMC sampling had been performed for a model with fixed indicators (see for subclasses for example mcmcoutputfix, mcmcoutputbase, mcmcoutputhier or mcmcoutputpost).

The class mcmcoutput is a class union and includes all classes that define objects to store MCMC samples and is used to dispatch methods for mcmcoutput objects. For the user this detail is not important, especially as this class has no exported constructor. Objects are solely constructed internally within the function mixturemcmc().

Class methods

This class comes along with a couple of methods that should give the user some comfort in handling the MCMC sampling results. There are no setters for this class as the slots are only set internally.

Show

• show() shows a short summary of the object's slots.

Getters

• getM() returns the M slot.

• getBurnin() returns the burnin slot.

• getRanperm() returns the ranperm slot.

• getPar() returns the par slot.

• gteWeight() returns the weight slot, if available.

• getLog() returns the log slot.

• getEntropy() returns the entropy slot, if available.

• getHyper() returns the hyper slot, if available.

• getPost() returns the post slot, if available.

• getST() returns the ST slot, if available.

• getS() returns the S slot, if available.

• getNK() returns the NK slot, if available.

• getClust() returns the clust slot, if available.

• getModel() returns the model slot.

• getPrior() returns the prior slot.

Plotting

Plotting functionality for the mcmcoutput helps the user to inspect MCMC results.

• plotTraces() plots traces of MCMC samples. See plotTraces() for further information.

• plotHist() plots histograms of parameters and weights. See plotHist() for further information.

• plotDens() plots densities of parameters and weights. See plotDens() for further information.

• plotPointProc() plots the point process of component parameters. See plotPointProc for further information.

• plotSampRep() plots the sampling representation of component parameters. See plotSampRep() for further information.

• plotPostDens() plots the posterior density of component parameters. Note that this function can only be applied for mixtures of two components. See plotPostDens() for further information.

Slots

• M An integer defining the number of iterations in MCMC sampling.

• burnin An integer defining the number of iterations in the burn-in phase of MCMC sampling. These number of sampling steps are not stored in the output.

• ranperm A logical indicating, if MCMC sampling has been performed with random permutations of components.

• par A named list containing the sampled component parameters.

• weight An array of dimension M x K containing the sampled weights.

• log A named list containing the values of the mixture log-likelihood, mixture prior log-likelihood, and the complete data posterior log-likelihood.

• hyper A list storing the sampled parameters from the hierarchical prior.

• post A named list containing a list par that contains the posterior parameters as named arrays.

• entropy An array of dimension M x 1 containing the entropy for each MCMC draw.

• ST An array of dimension M x 1 containing all MCMC states, for the last observation in slot y of the fdata object passed in to mixturemcmc() where a state is defined for non-Markov models as the last indicator of this observation.

• S An array of dimension N x storeS containing the last storeS indicators sampled. storeS is defined in the slot storeS of the mcmc object passed into mixturemcmc().

• NK An array of dimension M x K containing the number of observations assigned to each component for each MCMC draw.

• clust An array of dimension N x 1 containing the recent indicators defining the last "clustering" of observations into the mixture components.

• model The model object that specifies the finite mixture model for which MCMC sampling has been performed.

• prior The prior object defining the prior distributions for the component parameters that has been used in MCMC sampling.

See Also

• mcmcoutputperm for the corresponding class defined for relabeled MCMC samples

• mcmcoutputfix for the mcmcoutput sub-class for models with fixed indicators

• mcmcoutputbase for the mcmcoutput sub-class for models with unknown indicators

• mcmcoutputhier for the mcmcoutput sub-class for MCMC samples with hierarchical priors

• mcmcoutputpost for the mcmcoutput sub-class for MCMC samples with stored posterior density parameters

simonsays1980/finmix documentation built on Dec. 23, 2021, 2:25 a.m.