Question about marginalisation

[Tom Varley]

Thanks for all the help you given me thus far.

I am now trying to alter the program and had questions about the
effects. In you paper, hep-ph/0602028, you say on page 6 :

"In particular, one does not need to carry out the
marginalization integral in
Eq. (2.6) explicitly. It is sufficient to ignore the coordinates of
the samples in parameter
space along the marginalized directions. This is one more major
advantage of the MCMC
method."

Is this true even in the case of non-flat priors?

Thanks again,

Tom

[Roberto Trotta]

Hi Tom

yes, the marginalization procedure works irrespective of the prior you
have taken. Of course it remains true that different priors might
affect the shape of the posterior.

Roberto