JoSAE
, rsae
and sae
to the Suggests-Field in file DESCRIPTION.auxiliaryWeights
to class saeObj
, allowing for weighting
the auxiliary information for incomplete spatial support. This is inspired by
package forestinventory
.This lead to refactoring the functions used in method predict
.
I have nearly halved the lines of code in these functions, they are now much
easier to read, understand and maintain.
The return value predict() has changed, it is giving the (pseudo) small area estimator and the (pseudo) synthetic estimator as well but not the attributes hinting to Mandallaz' publications
If you don't care about weights and (pseudo) small area estimator and the (pseudo) synthetic
estimators, you might want to stick with the old predict method from version 1.0.0.
You can use
predict(..., version = "1.0.0)
or set
options(maSAE_version = "1.0.0")
to do so.
The predictions for the extended (pseudo) synthetic estimator and its variance
are identical for version 1.0.0 and 2.0.0. This is ensured by tests in
runit_tests/runit-v1.R.
bind_data
to coerce different sampling phase data into a
suitable data.frame
.maASE
and forestinventory
.NEWS.md
file to track changes to the package.