npcs: Neyman-Pearson Classification via Cost-Sensitive Learning

We connect the multi-class Neyman-Pearson classification (NP) problem to the cost-sensitive learning (CS) problem, and propose two algorithms (NPMC-CX and NPMC-ER) to solve the multi-class NP problem through cost-sensitive learning tools. Under certain conditions, the two algorithms are shown to satisfy multi-class NP properties. More details are available in the paper "Neyman-Pearson Multi-class Classification via Cost-sensitive Learning" (Ye Tian and Yang Feng, 2021), which will be posted on arXiv soon.

Version:	0.1.0
Depends:	R (≥ 3.5.0)
Imports:	dfoptim, nnet, randomForest, e1071, magrittr, MASS, smotefamily, rpart, foreach, naivebayes, caret, formatR
Suggests:	knitr, rmarkdown
Published:	2021-10-19
Author:	Ye Tian [aut, cre], Yang Feng [aut]
Maintainer:	Ye Tian <ye.t at columbia.edu>
License:	GPL-2
NeedsCompilation:	no
CRAN checks:	npcs results

Documentation:

Reference manual:	npcs.pdf
Vignettes:	npcs-demo

Downloads:

Package source:	npcs_0.1.0.tar.gz
Windows binaries:	r-devel: npcs_0.1.0.zip, r-release: npcs_0.1.0.zip, r-oldrel: npcs_0.1.0.zip
macOS binaries:	r-release (arm64): npcs_0.1.0.tgz, r-oldrel (arm64): npcs_0.1.0.tgz, r-release (x86_64): npcs_0.1.0.tgz, r-oldrel (x86_64): npcs_0.1.0.tgz

Linking:

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