Added function peptest for hypothesis
testing
Added function estimation.pep which can be used for
estimation under Bayesian model averaging
Added function posteriorpredictive.pep which
simulates values from the posterior predictive distribution under
Bayesian model averaging
Added function comparepriors.lm which returns
selected models under different choices of prior on the model parameters
and the model space
The argument xnew of S3 method predict
has to be a data frame instead of a matrix.
The older functions full_enumeration_pep and
mc3_pep were unified and replaced by the new function
pep.lm. The main differences are the following: a)
pep.lm takes as arguments a formula and a data frame
instead of a matrix (input data matrix) and a vector (response vector)
and b) with the new argument
algorithmic.choice="automatic", the choice of the model
selection algorithm (full enumeration or MC3) can also be done
automatically and does not necessarily have to be chosen by the
user.
Finally, the new function returns few additional list elements:
fullmodel representing the full model that was used,
mapp containing the mapping between Xi’s and the original
explanatory variable names and allvisitedmodsM containing
all
visited' models after the burnin period (relevant for MC3). The older functionsfull_enumeration_pepandmc3_pep`
return a warning message that are deprecated.