Usmile: Threshold-Free Class-Specific Comparison of Binary Classifiers

Implements the U-smile methodology for threshold-free, class-specific comparison of probabilistic binary classifiers. The package quantifies prediction improvement and worsening separately for non-events and events using the Brier alteration (BA), relative Brier (RB), improvement proportion (I) coefficients, and relative likelihood ratio (rLR) coefficients, and provides U-smile, prediction improvement-worsening, receiver operating characteristic, and precision-recall plots. The original U-smile framework is described in Kubiak et al. (2024) <doi:10.1371/journal.pone.0303276>, its three-level extension for imbalanced binary classification in Wieckowska et al. (2025) <doi:10.1371/journal.pone.0321661>, and the likelihood-based extension in Wieckowska and Guzik (2026) <doi:10.1038/s41598-026-40545-z>.

Version:	0.2.0
Depends:	R (≥ 3.5.0)
Imports:	ggplot2, grid, pROC
Suggests:	e1071, naivebayes, nnet, parsnip, randomForest, ranger, testthat (≥ 3.0.0), xgboost
Published:	2026-08-08
DOI:	10.32614/CRAN.package.Usmile
Author:	Barbara Więckowska [aut, cre]
Maintainer:	Barbara Więckowska <bbwieckowska at gmail.com>
BugReports:	https://github.com/bbwieckowska/Usmile/issues
License:	MIT + file LICENSE
URL:	https://github.com/bbwieckowska/Usmile
NeedsCompilation:	no
Citation:	Usmile citation info
Materials:	README, NEWS
CRAN checks:	Usmile results

Documentation:

Reference manual:

Usmile.html , Usmile.pdf

Downloads:

Package source:	Usmile_0.2.0.tar.gz
Windows binaries:	r-devel: not available, r-release: Usmile_0.2.0.zip, r-oldrel: Usmile_0.2.0.zip
macOS binaries:	r-release (arm64): Usmile_0.2.0.tgz, r-oldrel (arm64): Usmile_0.2.0.tgz, r-release (x86_64): Usmile_0.2.0.tgz, r-oldrel (x86_64): Usmile_0.2.0.tgz

Linking:

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