[R] prior covariance in Mclust

From: Dieter Vanderelst <dieter.vanderelst_at_gmail.com>
Date: Thu, 28 Jun 2007 12:55:34 +0200


I'm trying to use Mclust to fit a Gaussian Mixture Model to a mulitdimensional data set.

Because of the specific source of my data, I know that all components have the same variance and that the covariance between dimensions is zero (modelname=VII).
Furthermore, I have a reliable estimate of the variance of the components.

I want to to use this estimate as a prior in mclust, hoping that exploiting this knowledge will yield better estimates of the number of components and their means (which are the unknowns).

First I have a general question: Is this a sensible thing to do? As far as I can see (which might be not too far), this will indeed lead to more robust estimates. But is this true?

Another question concerns the practical side of things. How can I do this in mclust? I've read through the manual but this leaves me uncertain about the exact implementation (and earlier posts about this problem seem not to have been answered by the mailing list).

If specifying a prior for the covariance matrix is possible (and sensible) in mclust, could someone provide an example of how to specify a prior on the covariance matrix?

Thank you,

R-help_at_stat.math.ethz.ch mailing list
https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. Received on Thu 28 Jun 2007 - 11:31:07 GMT

Archive maintained by Robert King, hosted by the discipline of statistics at the University of Newcastle, Australia.
Archive generated by hypermail 2.2.0, at Thu 28 Jun 2007 - 12:32:39 GMT.

Mailing list information is available at https://stat.ethz.ch/mailman/listinfo/r-help. Please read the posting guide before posting to the list.