Compare fitted models

[Pascal Pettigrew]

Hi,

I fitted both models ("presence" and "absence" of trap response) in SPACECAP and I wondered if it have a way to select the "best" model? Can I use Bayesian p-value ? 

Thank you!

Pascal

[Jeffrey Royle]

hi Pascal,

 the Bayesian p-value will give you an assessment of model fit but it is not a model selection criterion.  For that you would use a posterior model probability, or Bayes factor, or one of several other approaches in the literature. These are not currently implemented in SPACECAP and so you would have to either use JAGS/WinBUGS or some other implementation.

regards

andy

--
You received this message because you are subscribed to the Google Groups "SPACECAPforum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to spacecapforum+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.