New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
stanreg, brmsfit tidiers are inconsistent with tidyMCMC #291
Comments
I would love to use Will come back to this later after there's been time for people to discuss. |
This discussion should also happen in |
I think this can be closed. |
This issue has been automatically closed due to inactivity. |
This issue has been automatically locked. If you believe you have found a related problem, please file a new issue (with a reprex: https://reprex.tidyverse.org) and link to this issue. |
I see that @paul-buerkner wrote the method for
brmsfit
objects and I can anticipate some of the reasons he may have done this intentionally, but I'm not sure whether the current format makes sense.While some aspects (
par_type
,parameters
, etc.) are related to broader issues with mixed models (e.g., #96), that's not what I'm worried about since there are ongoing discussions about that. Instead, I'm concerned about the "confidence" intervals.tidyMCMC
is basically consistent with the general templates: we haveconf.int
andconf.level
arguments and it returns columns with namesconf.low
andconf.high
.tidy.brmsfit
andtidy.stanreg
, on the other hand, useintervals
andprob
arguments and return corresponding columns calledlower
andupper
.I would say that at minimum, the inconsistency between
tidyMCMC
andtidy.brmsfit
/tidy.stanreg
doesn't make sense. Now I suspecttidy.brmsfit
/tidy.stanreg
opted not use any reference to "conf" because the intervals are not really confidence intervals. My opinion is that it would be better to stick with thebroom
norms (using the "conf" prefix) since users should understand the distinction and it is better than breaking any function that relies on the generality oftidy
(like the one in my package that led me to realize this issue). I assume there are other model types supported by broom using "conf" that produce intervals that are also not truly confidence intervals, but I'm not certain.But if others disagree with me about the nomenclature, I think it would then be best to change
tidyMCMC
so that someone who is willing to deal with MCMC models in this specific way doesn't have to switch back and forth between naming schemes to support bothbrmsfit
/stanreg
andstanfit
/rjags
/whatever else objects.The text was updated successfully, but these errors were encountered: