The main new feature in this release is the addition ot varbvsmix
, which fits a linear regression with mixture-of-normals priors, using similar fully-factorized variational approximation methods. See help(varbvsmix)
in the R package for more details. It is also implemented in MATLAB as varbvsmix.m
.
Also includes a variety of miscellaneous bug fixes.