Re: Bayes' Decision Rule

In article <1147339718.876787.185890@xxxxxxxxxxxxxxxxxxxxxxxxxxxx>,
Murat Aykut <murat_aykut@xxxxxxxxx> wrote:
I'm implementing Bayes Decision Rule. For this purpose I'm using
Multivariate Normal Distribution. But, in some cases Covariance matrix
becomes singular (there isn't inverse Covariance matrix). What is the
solution? What can I do? (I think that, i can add a very little to a
value which is in the same lines in the matrix. Or changing the
distribution. But, what is the true answer? I don't know).
Please help.

Please explain your problem more carefully.

If you are using a real Bayesian procedure, this should
not be a problem. On the other hand, if you are using
the so-called "non-informative" prior, the procedure is
only formally Bayes, and even if your sample size is
large enough, will not be admissible.

So, what is your problem, and what are your assumptions?



--
This address is for information only. I do not claim that these views
are those of the Statistics Department or of Purdue University.
Herman Rubin, Department of Statistics, Purdue University
hrubin@xxxxxxxxxxxxxxx Phone: (765)494-6054 FAX: (765)494-0558
.

Relevant Pages