Package: spikeslab
Version: 1.1.6
BUILD: bld20220426

---------------------------------------------------------------------------------
CHANGES TO RELEASE 1.1.6

o CRAN compliance update.

---------------------------------------------------------------------------------
CHANGES TO RELEASE 1.1.4

o This build corrects a long-standing numerical issue with the BMA
that can occur under certain atypical settings"

---------------------------------------------------------------------------------
CHANGES TO RELEASE 1.1.3

o improved interconnectivity between user functions.  Enhancements to cv.spikeslab.
o compression of data files.

---------------------------------------------------------------------------------
CHANGES TO 1.1.2

o minor fixes to R-side wrapper, adjustments to packaging, and documentation.

---------------------------------------------------------------------------------
CHANGES TO RELEASE 1.1.0

RELEASE 1.1.0 is a recommended upgrade of the product.

o cv.spikeslab() now takes advantage of the CRAN package "snow".  It
allows users to create a socket cluser on the local machine, enabling
parallel execution of this function.  It scales with the number of CPU
cores on the local machine.

---------------------------------------------------------------------------------
CHANGES TO RELEASE 1.0.3

o minor fixes to R-side wrapper, adjustments to packaging, and documentation.

---------------------------------------------------------------------------------
CHANGES TO RELEASE 1.0.2

o minor adjustments to packaging

---------------------------------------------------------------------------------
CHANGES TO RELEASE 1.0.1

o minor adjustments to the R-side wrappers and documentation

---------------------------------------------------------------------------------
CHANGES TO RELEASE 1.0.0

RELEASE 1.0.0 is the first and initial release of this package.

Fits a rescaled spike and slab model using a continuous bimodal
prior. Can be used for prediction and variable selection in low and
high-dimensional linear regression models.

Key features include:

o Option for ultra-fast handling of high-dimensional data.
o Variable selection using the generalized elastic net (gnet).
o Grouping of variables with unique regularization (no limit on the number).
o Predict wrapper for predicting on test data.
o Sparse PC approach for multiclass analysis of gene expression data.
o Depends on the randomForest and lars R-packages.


---------------------------------------------------------------------------------