Efficient algorithms for fitting the regularization path of linear regression, GLM, and Cox regression models with grouped penalties. This includes group selection methods such as group lasso, group MCP, and group SCAD as well as bi-level selection methods such as the group exponential lasso, the composite MCP, and the group bridge.
Version: | 3.2-1 |
Depends: | R (≥ 3.1.0) |
Imports: | Matrix |
Suggests: | grpregOverlap, knitr, survival |
Published: | 2019-02-26 |
Author: | Patrick Breheny [aut, cre] (<https://orcid.org/000-0002-0650-1119>), Yaohui Zeng [ctb] |
Maintainer: | Patrick Breheny <patrick-breheny at uiowa.edu> |
BugReports: | http://github.com/pbreheny/grpreg/issues |
License: | GPL-3 |
URL: | http://pbreheny.github.io/grpreg, https://github.com/pbreheny/grpreg |
NeedsCompilation: | yes |
Citation: | grpreg citation info |
Materials: | README NEWS |
In views: | MachineLearning |
CRAN checks: | grpreg results |
Reference manual: | grpreg.pdf |
Vignettes: |
Getting started |
Package source: | grpreg_3.2-1.tar.gz |
Windows binaries: | r-devel: grpreg_3.2-1.zip, r-devel-gcc8: grpreg_3.2-1.zip, r-release: grpreg_3.2-1.zip, r-oldrel: grpreg_3.2-1.zip |
OS X binaries: | r-release: grpreg_3.2-1.tgz, r-oldrel: grpreg_3.2-1.tgz |
Old sources: | grpreg archive |
Reverse depends: | grpregOverlap |
Reverse imports: | bestglm, DMRnet, geoGAM, grpss, naivereg, refund |
Please use the canonical form https://CRAN.R-project.org/package=grpreg to link to this page.