choose.nmixt {varmixt}R Documentation

Fow to choose the number of components in the mixture model

Description

This function finds the number of components that minimizes the BIC criteria.

Usage

choose.nmixt(VAR,df,crit=c("AIC","BIC")[2],stop.crit=stop.crit.2,display=TRUE,
                       criterion=c("likelihood","parameter")[1])

Arguments

VAR Vector of estimated variance
df degrees of freedom of the estimated variance
crit method for choosing the number of component : "BIC" or "AIC"
stop.crit stop criterion for EM algorithm
display Boolean. Should the estimated parameters be displayed on the screen ?
criterion "likelihood" or "parameter". The convergence of EM algo will be decided with regard to the increase of likelihood or lack of noticeable in parameter value.

Author(s)

Paul Delmar and Julie Aubert

See Also

compute.dif,compute.dif.paired,compute.BIC


[Package varmixt version 0.2-4 Index]