NNLM: Fast and Versatile Non-Negative Matrix Factorization

This is a package for Non-Negative Linear Models (NNLM). It implements fast sequential coordinate descent algorithms for non-negative linear regression and non-negative matrix factorization. It supports mean square error and Kullback-Leibler divergence loss. Many other features are also implemented, including missing value imputation, domain knowledge integration, designable W and H matrices and multiple forms of regularizations.

Version: 0.4.0
Depends: R (≥ 3.0.1)
Imports: Rcpp (≥ 0.11.0), stats, utils
LinkingTo: Rcpp, RcppArmadillo, RcppProgress
Suggests: testthat, knitr, rmarkdown, mice, missForest, ISOpureR
Published: 2015-12-26
Author: Xihui Lin [aut, cre], Paul C Boutros [aut]
Maintainer: Xihui Lin <xihuil.silence at gmail.com>
BugReports: https://github.com/linxihui/NNLM/issues
License: BSD_2_clause + file LICENSE
URL: https://github.com/linxihui/NNLM
NeedsCompilation: yes
Materials: NEWS
CRAN checks: NNLM results

Downloads:

Reference manual: NNLM.pdf
Vignettes: Fast and versatile NMF
Package source: NNLM_0.4.0.tar.gz
Windows binaries: r-devel: NNLM_0.4.0.zip, r-release: NNLM_0.4.0.zip, r-oldrel: NNLM_0.4.0.zip
OS X Snow Leopard binaries: r-release: NNLM_0.4.0.tgz, r-oldrel: not available
OS X Mavericks binaries: r-release: not available