Ravenports generated: 12 Jan 2021 01:22
[ravenports.git] / bucket_AC / R-glmnet
1 # Buildsheet autogenerated by ravenadm tool -- Do not edit.
2
3 NAMEBASE=               R-glmnet
4 VERSION=                4.1
5 KEYWORDS=               cran
6 VARIANTS=               standard
7 SDESC[standard]=        Generalized Linear Models for Lasso, etc
8 HOMEPAGE=               https://glmnet.stanford.edu
9 CONTACT=                CRAN_Automaton[cran@ironwolf.systems]
10
11 DOWNLOAD_GROUPS=        main
12 SITES[main]=            CRAN/src/contrib
13 DISTFILE[1]=            glmnet_4.1.tar.gz:main
14 DIST_SUBDIR=            CRAN
15 DF_INDEX=               1
16 SPKGS[standard]=        single
17
18 OPTIONS_AVAILABLE=      none
19 OPTIONS_STANDARD=       none
20
21 BUILDRUN_DEPENDS=       R-foreach:single:standard
22                         R-shape:single:standard
23
24 USES=                   cran gmake
25
26 DISTNAME=               glmnet
27
28 GENERATED=              yes
29
30 INSTALL_REQ_TOOLCHAIN=  yes
31
32 [FILE:805:descriptions/desc.single]
33 glmnet: Lasso and Elastic-Net Regularized Generalized Linear Models
34
35 Extremely efficient procedures for fitting the entire lasso or elastic-net
36 regularization path for linear regression, logistic and multinomial
37 regression models, Poisson regression, Cox model, multiple-response
38 Gaussian, and the grouped multinomial regression. There are two new and
39 important additions. The family argument can be a GLM family object, which
40 opens the door to any programmed family. This comes with a modest
41 computational cost, so when the built-in families suffice, they should be
42 used instead. The other novelty is the relax option, which refits each of
43 the active sets in the path unpenalized. The algorithm uses cyclical
44 coordinate descent in a path-wise fashion, as described in the papers
45 listed in the URL below.
46
47
48 [FILE:101:distinfo]
49 8f0af50919f488789ecf261f6e0907f367d89fca812baa2f814054fb2d0e40cb      2152465 CRAN/glmnet_4.1.tar.gz
50