pgmmRJMCMC: bpgmm Model-Based Clustering Using Baysian PGMM Carries out...

In bpgmm: Bayesian Model Selection Approach for Parsimonious Gaussian Mixture Models

bpgmm Model-Based Clustering Using Baysian PGMM Carries out model-based clustering using parsimonious Gaussian mixture models. MCMC are used for parameter estimation. The RJMCMC is used for model selection.

Description

bpgmm Model-Based Clustering Using Baysian PGMM Carries out model-based clustering using parsimonious Gaussian mixture models. MCMC are used for parameter estimation. The RJMCMC is used for model selection.

Usage

pgmmRJMCMC(X, mInit, mVec, qnew, delta = 2, ggamma = 2, burn = 20,
  niter = 1000, constraint = C(0, 0, 0), dVec = c(1, 1, 1),
  sVec = c(1, 1, 1), Mstep = 0, Vstep = 0, SCind = 0)

Arguments

X	the observation matrix with size p * m
mInit	the number of initial clusters
mVec	the range of the number of clusters
qnew	the number of factor for a new cluster
delta	scaler hyperparameters
ggamma	scaler hyperparameters
burn	the number of burn in iterations
niter	the number of iterations
constraint	the pgmm initial constraint, a vector of length three with binary entry. For example, c(1,1,1) means the fully constraint model
dVec	a vector of hyperparameters with length three, shape parameters for alpha1, alpha2 and bbeta respectively
sVec	sVec a vector of hyperparameters with length three, rate parameters for alpha1, alpha2 and bbeta respectively
Mstep	the indicator of whether do model selection on the number of clusters
Vstep	the indicator of whether do model selection on variance structures
SCind	the indicator of whether use split/combine step in Mstep

Examples

library("fabMix")
library("mclust")
library("pgmm")
library("mvtnorm")
library("mcmcse")
library("MASS")
library("gtools")
n <- 500
p <- 10
q <- 4
K <- 10
nsim <- 10
burn <- 20
qnew <- 4
Mstep <- 1
Vstep <- 1
constraint <- c(0, 0, 0)
mInit <- 20
mVec <- c(1, 20)
X <- t(simData(
  sameLambda = TRUE,
  sameSigma = TRUE,
  K.true = K, n = n, q = q, p = p, sINV_values = 1 / ((1:p))
)$data)

pgmmRJMCMC(X,
  mInit, mVec, qnew,
  niter = nsim, burn = burn,
  constraint = constraint, Mstep = Mstep, Vstep = Vstep
)

bpgmm documentation built on June 2, 2022, 1:10 a.m.