compute.BIC.init {varmixt}R Documentation

Computation of parameters of the mixture model given initial parameters

Description

This function computes parameters of the mixture model given initial parameters

Usage

compute.BIC.init(VAR,df,sd.init,pi.init,nmixt,stop.crit=stop.crit.2,display=TRUE,PPOST=FALSE,
                           niter.max=50000,VAR.OUT=FALSE,criterion=criterion)

Arguments

VAR Vector of estimated variance
df degrees of freedom of the estimated variance
sd.init vector of initial standard deviation for the mixture
pi.init vector of initial prior probability for the mixture
nmixt Integer : number of components
stop.crit Real : convergence criteria for the EM algo
display Boolean : Should the program write its results on the console ?
PPOST Boolean : Should the programm export the matrix of posterior probability of each gene ?
niter.max Integer : maximum number of iteration of the EM algorithm
VAR.OUT Boolean : Should the program export the vector of variance and the vector of corresponding components ?
criterion Criterion for the convergence of EM algo. Either "likelihood" for criterion based on loglikelihood or "parameter"

Author(s)

Paul Delmar


[Package varmixt version 0.2-4 Index]