penalized: L1 (lasso and fused lasso) and L2 (ridge) penalized estimation in GLMs and in the Cox model

A package for fitting possibly high dimensional penalized regression models. The penalty structure can be any combination of an L1 penalty (lasso and fused lasso), an L2 penalty (ridge) and a positivity constraint on the regression coefficients. The supported regression models are linear, logistic and Poisson regression and the Cox Proportional Hazards model. Cross-validation routines allow optimization of the tuning parameters.

Version: 0.9-45
Depends: R (≥ 2.10.0), survival, methods
Suggests: globaltest
Published: 2014-12-19
Author: Jelle Goeman, Rosa Meijer, Nimisha Chaturvedi
Maintainer: Jelle Goeman <jelle.goeman at>
License: GPL-2 | GPL-3 [expanded from: GPL (≥ 2)]
NeedsCompilation: no
Materials: ChangeLog
In views: MachineLearning, Survival
CRAN checks: penalized results


Reference manual: penalized.pdf
Vignettes: Penalized user guide
Package source: penalized_0.9-45.tar.gz
Windows binaries: r-devel:, r-release:, r-oldrel:
OS X Snow Leopard binaries: r-release: penalized_0.9-45.tgz, r-oldrel: penalized_0.9-45.tgz
OS X Mavericks binaries: r-release: penalized_0.9-45.tgz
Old sources: penalized archive

Reverse dependencies:

Reverse depends: DIFlasso, DIFtree, lmmlasso, multiPIM, ROC632, subtype, uplift
Reverse imports: apricom, DIFboost, hdnom, pensim
Reverse suggests: catdata, fscaret, Grace, lda, mlr, ordinalNet, peperr