VBmix (version 0.3.2)

vbconstr:

Description

estimates the variational posterior distribution of a GMM that aggregates a constrained collection of GMM. A lower bound is calculated and monitored at each iteration. This posterior can be used for various purposes (e.g. MC proposal distribution). It can be transformed using extractSimpleModel, outputing a GMM.

Usage

vbconstr(models, ncomp, thres = 0.1, maxit = NULL)

Arguments

models

GMM made with the weighted sum of the collection of GMM to aggregate. a is used to model constraints between components in this GMM.

ncomp

number of components in the posterior.

thres

threshold for lower bound variations between 2 iterations. Convergence is decided if this variation is below thres.

maxit

if NULL, the stopping criterion is related to thres. If not NULL, maxit iterations are performed.

Value

estimated posterior with ncomp components.

References

Bruneau, P., Gelgon, M., and Picarougne, F. (2010) _Parsimonious reduction of Gaussian mixture models with a variational-Bayes approach_, Pattern Recognition, Volume 43, Pages 850-858.

See Also

vbcomp extractSimpleModel

Examples

temp1 <- newGmm()
for(i in 1:10) temp1 <- appendToGmm(temp1, gmmpen[[i]])
temp2 <- vbconstr(temp1, 50)