NEWS | R Documentation |
No improvements; compatibility with dplyr 1.0.0
lmer
tidier gets ddf.method
(applies only to
lmerTest fits)
glmmTMB
gets exponentiate
options
experimental GLMMadaptive
tidiers
fixes for updates in tibble package
gls
tidier gets confint (GH #49)
redundant estimate.method
in MCMC tidiers goes
away; use robust
to compute point estimates/uncertainty via
median and MAD rather than mean and SE
misc fixes: lme4
tidiers (confint for ran_vals,
profile conf intervals fixed), R2jags
, gamlss
...
ran_vals works for glmmTMB
don't ignore conf.level in tidy.(merMod|glmmTMB) (GH #30,31: @strengejacke)
levels correct in tidy.brmsfit (GH #36: @strengejacke)
component argument works for random effects in glmmTMB (GH #33: @strengejacke)
brmsfit
and rstanarm
methods allow
conf.method="HPDinterval"
tidy.brmsfit gets component column (GH #35: @strengejacke), response column for multi-response models (GH #34: @strengejacke)
component tags are stripped from tidied brmsfit objects
"Intercept" terms in brms fits are re-coded as
"(Intercept)" by default, for dotwhisker/cross-model
compatibility; for previous behaviour, specify fix.intercept=FALSE
modify examples, for CRAN compliance
reduced size of stored fits for examples, for CRAN compliance
more consistent term names in brmsfit
, rstanreg
tidiers
improved tidy.MCMCglmm
all methods return tibbles (tbl_df
) rather than data
frames
the value of the group
variable for fixed-effect
parameters has changed from "fixed"
to NA
brmsfit
and rstanarm
tidiers are more
consistent with other tidiers (e.g. conf.level
rather than
prob
)
Sorted out some of the confusion over random effect naming: "ran_vals" extracts conditional modes/BLUPs/varying parameters (deviations from population-level estimates), while "ran_coefs" extracts group-level estimates
improved nlme
tidiers
improved glmmTMB
tidiers (can handle some
zero-inflation parameters)
lme4
tidiers now optionally take a pre-computed
profile
argument when using conf.method="profile"
The default behaviour of most mixed-model tidiers has changed/will gradually be changed to the following (description modified from TJ Mahr at https://github.com/tidyverse/broom/issues/96):
Random effect variances and covariances can now be extracted.
effects = "ran_pars"
returns the standard deviations/correlations
of random effects (if scales="sdcor"
[default])
or their variances
and covariances (if scales = "varcov"
)
Random effects estimates are now extracted with
effects = "ran_coefs"
for the group-level
estimates (previously these effects were
extracted with tidy(model, "random")
) or
effects = "ran_vals"
for the conditional modes
(deviations of the group-level parameters from the
population-level estimates)
effects
can take a vector of values (those listed
above, plus "fixed" for fixed effects). The default value
is effects = c("ran_pars", "fixed")
which extracts random
effect variances/covariances and fixed effect estimates.
term
names for random-effect parameters no longer
contain a (redundant) group specifier (at least for lme4
models); use something like
tidyr::unite(term,term,group,sep=".")
to collapse the two columns