NEWS | R Documentation |
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