require()
in function
.initmethod()
has been replaced with
requireNamespace()
in order to load functions from the
robustbase
package.fitsaemodel()
fit_model_b
(in
earlier versions, the class was called fitsaemodel
).fitsaemodel()
returns an object with the slots beta
(estimated
fixed-effects coefficients) and theta
(variance components)
equal to a vector of NA
’s.vcovbeta
)
is now computed in function summary()
and not
fitsaemodel()
. This goes unnoticed by the user.coef.fit_model_b()
only returns the
coefficients but does not print to the console anymore.fitsaemodel.control()
gains the additional
argument k_Inf
to specify the robustness tuning constant k
that represents infinity (used to define the maximum likelihood
estimator)digit
of the function
print.fit_model_b()
depends now on
getOption("digits")
.summary.fit_model_b()
returns now the model
summary.robpredict()
seed
to
specify the random seed used in the parametric bootstrap.robpredict()
is now
implemented with utils:txtProgressBar()
and can be switched
off using progress_bar = FALSE
.X
of an saemodel
instance, is now called (Intercept)
not
X.Intercept
.init.c
added vignette folder
dec
and decor
in
fitsaemodel.control()
to choose between different types of
decorrelation of the residuals and the type of matrix square roots
(Cholesky vs. SVD).fitsaemodel.huberm()
makedata()
fitsaemodel()
when different
robustness tuning are usedconvergence()
to learn more about the
convergence of the methodsareameans
in function
robpredict()
type
in plot method for class
meanssaemodel
removed argument full
in the summary method for objects
of class fitmodel
Initial release