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

View source: R/PGMMmain.R

pgmmRJMCMCR Documentation

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.

Related to pgmmRJMCMC in bpgmm...