Error with lme4 models with offset and re.form = NULL #335
Comments
|
OK, I assume that this is an lme4 problem, because if I check lme4 S3 functions for the model object
I don't see why simulate shouldn't work if predict works. I will probably have to send this to the lme4 crew. @bbolker - could you maybe have a look? For the user - as a quick fix, you could use predict and then simulate by hand using the appropriate distribution. |
|
As the error message shows, this is a problem with trying to evaluate the offset term in an environment where it can't be found (the bane of my lme4-developing existence):
PS it looks like |
|
Regarding the environment: OK, but I assume there is no fundamental reason for this to occur, so it's a programming bug on lme4 side, right? Sorry for adding to your pile / plight ;) About the offset included in the REs - my understanding / expectation would have been that regardless of the re.form setting, the offset always needs to be included, because it's part of the model, so re.form only affects whether the REs should be included, but offset + fixed effects will always be included? |
|
I think lme4/lme4@8c8ab5c should fix this ... although (a) I'm not sure it's 100% bulletproof (although I couldn't break it as easily as I thought I'd be able to) and (b) can't give a guarantee at the moment about when this will make it to CRAN (although I have my arm twisted; last CRAN submission was 7 May 2022, so should probably wait a couple of weeks at least) |
|
Great, thanks! I'll leave this open for a moment to check with the next CRAN version! |
|
This was fixed in glmmTMB 1.1.4 -> closed! |
CHANGES IN VERSION 1.1-33:
BUG FIXES:
* a boundary check could fail occasionally when large data
produced an NA value in a computed gradient; now warns
instead (GH #719, Mathias Ambuehl)
* 'allFit' now works better when 'optimx' and 'dfoptim'
packages are not installed (GH #724)
* 'refit' reset internal degrees of freedom component
incorrectly for REML fits (resulted in incorrect reported
REML criteria, but otherwise harmless: side effect of GH
#678)
NEW FEATURES:
* 'dotplot' and 'qqmath' methods gain a 'level' argument to
set the width of confidence intervals
* 'dotplot' method is now more flexible, using ".v" options
('lty.v', 'col.line.v', 'lwd.v') to set appearance of
vertical lines (Iago Gin<c3><a9> V<c3><a1>zquez)
* 'refit' gains a 'newweights' argument (GH #678)
CHANGES IN VERSION 1.1-32 (2023-03-14):
USER-VISIBLE CHANGES:
* 'formatVC()' gets a new optional argument 'corr' indicating
if correlations or covariances should be used for vector
random effects; this corresponds to 'print(<merMod>,
ranef.corr = ...)' . By default, it is FALSE for 'comp =
"Variance"', fixing (GH #707).
* 'qqmath.merMod' adds a (useless) 'data' argument for S3
compatibility. Going forward, the 'id' and 'idLabels'
arguments should always be specified by name. We have added
code to try to detect/warn when this is not done.
BUG FIXES:
* 'nobars' now retains the environment of its formula argument
(GH #713, Mikael Jagan)
CHANGES IN VERSION 1.1-31 (2022-11-01):
BUG FIXES:
* 'confint(fm, <single string>)' now works (after years of
being broken) again.
* simulating from binomial model with a factor response, when
the simulated response contains only a single factor level,
now works (Daniel Kennedy)
CHANGES IN VERSION 1.1-30 (2022-07-08):
USER-VISIBLE CHANGES:
* 'nl' (term names) component added to output list of
'mkReTrms (GH #679)'
* eliminate partial-matching of 'eta' (for 'etastart') (GH
#686: not actually "user-visible" unless
'getOption("warnPartialMatchDollar")' is 'TRUE')
* 'summary' method doesn't break for GLMMs other than
binomial/Poisson when 'merDeriv''s 'vcov.glmerMod' method is
attached (GH #688)
BUG FIXES:
* better handling of 'simulate(., re.form = NULL)' when model
frame contains derived components (e.g. 'offset()',
'log(x)')
(<florianhartig/DHARMa#335>)
* 'bootMer' works with 'glmmTMB' again (broken in 1.1-29)
* 'maxfun' argument to 'allFit' controls max function
evaluations for every optimizer type (GH#685)
CHANGES IN VERSION 1.1-29 (2022-04-07):
USER-VISIBLE CHANGES:
* prediction with new levels (when not allowed) returns a more
informative error message (displays a list of unobserved
levels)
BUG FIXES:
* 'glmer.nb' now works when 'lme4' is not loaded (GH #658,
@brgew)
* tests for singularity ('check.conv.singular') now run
independently of derivative computation (e.g., when
'calc.derivs=FALSE') (GH #660, @palday)
* 'influence.merMod' now works when data were originally
specified as a tibble
* fixed bug in 'cooks.distance' method for 'influence.merMod'
(i.e., objects created via 'influence(fitted_model)') (John
Fox) (GH #672)
* 'predict' works for formulas containing . when 'newdata' is
specified (GH #653)
* 'bootMer' now correctly inherits control settings from
original fit
CHANGES IN VERSION 1.1-28 (2022-02-04):
USER-VISIBLE CHANGES:
* construction of interacting factors (e.g. when 'f1:f2' or
'f1/f2' occur in random effects terms) is now more efficient
for partially crossed designs (doesn't try to create all
combinations of 'f1' and 'f2') (GH #635 and #636)
* 'mkNewReTrms' is exported
* singular-fit message now refers to 'help("isSingular")'
rather than '?isSingular'
TESTS:
* fix 'all.equal(p1,p2,p3)' and similar 'expect_equal()'
thinkos
* fix some tests only run when 'lme4:::testLevel() > 1'; adapt
tests for upcoming Matrix 1.4-1 which has
'names(diag(<sparse>))'
BUG FIXES:
* 'reOnly' preserves environment (GH #654, Mikael Jagan)
* backward-compatibility hooks changed to evaluate at run-time
(i.e., in '.onLoad()') rather than at build time (GH #649)
* 'lmList' no longer warns when 'data' is a tibble (GH #645)
Reported by a user. Reproducible example:
The text was updated successfully, but these errors were encountered: