PUlasso: High-Dimensional Variable Selection with Presence-Only Data

Efficient algorithm for solving PU (Positive and Unlabelled) problem in low or high dimensional setting with lasso or group lasso penalty. The algorithm uses Maximization-Minorization and (block) coordinate descent. Sparse calculation and parallel computing are supported for the computational speed-up. See Hyebin Song, Garvesh Raskutti (2017) <arXiv:1711.08129>.

Version: 3.1.1
Depends: R (≥ 2.10)
Imports: Rcpp (≥ 0.12.8), methods, Matrix, doParallel, foreach, ggplot2
LinkingTo: Rcpp, RcppEigen, Matrix
Suggests: testthat, knitr, rmarkdown
Published: 2018-05-15
Author: Hyebin Song [aut, cre], Garvesh Raskutti [aut]
Maintainer: Hyebin Song <hsong56 at wisc.edu>
BugReports: https://github.com/hsong1/PUlasso/issues
License: GPL-2
URL: https://arxiv.org/abs/1711.08129
NeedsCompilation: yes
CRAN checks: PUlasso results


Reference manual: PUlasso.pdf
Vignettes: PUlasso-vignette
Package source: PUlasso_3.1.1.tar.gz
Windows binaries: r-devel: PUlasso_3.1.1.zip, r-release: PUlasso_3.1.1.zip, r-oldrel: PUlasso_3.1.1.zip
OS X binaries: r-release: PUlasso_3.1.1.tgz, r-oldrel: PUlasso_3.1.1.tgz
Old sources: PUlasso archive


Please use the canonical form https://CRAN.R-project.org/package=PUlasso to link to this page.