Filter samples obtained from PMC using a stricter (uniform) prior

[dvandyk]

When producing samples using PMC for a problem that has e.g. a unitarity bound, with absolute limits on the size parameters, Meril and I currently face similar problems: The PMC does not easily converge. We therefore filter samples from a PMC run w/o application of the bounds.

Maybe a useful interface would be to create an "indicator" function or a filter that uses a second set of priors. Care needs to be taken that the 2nd set of priors does not include any informative priors, and that its absolute ranges are more strict than the original ones for all parameters.