what to do with `evppi()` when `eff` & `cost` have different lengths to input parameters `inputs`? about bcea HOT 6 CLOSED

@giabaio any ideas?..

from bcea.

I struggle to replicate this... Are you on the 2.4.4 version of BCEA? Is the object smoking_output available by default?

from bcea.

Hi,
You're right that I had taken out the jags output when sending to CRAN previously. I think I thought it was an issue but probably wasnt the actual problem.

I'd put it back in to the project recently to do the evppi testing, see:
n8thangreen@38ff5ff

I think this is the code taken wholesale from the BCEA book.

Smoking.Rmd is in the refactor-remove-evppi branch or I can send it to you separately?

from bcea.

One more on this, before we can perhaps close this: I've modified the smoking_output object under data/Smoking.RData to select only the relevant matrix of simulations. This actually makes sense, because to use this object for the createInputs() and then evppi() example, we don't need it to be an rjags object --- it can be (as it can be a bugs or stan object, but a simple matrix will just do...). With this change (and by updating the GitHub repo for plotrix --- BTW: I think it's not orphaned any more, so perhaps we don't even need to add that repo in the Remotes field of our DESCRIPTION file) it now passes all the checks, except for AppVeyor. @n8thangreen given that we have included the check under Windows with the normal flow, do we need to keep AppVeyor? Or could we just remove that too?...

from bcea.

Thanks for fixing this. After syncing the my fork its passing all the CHECKS in the GH Action. I'll remove the AppVeyor thing; I think thats just a legacy thing anyway

from bcea.

from bcea.