dsos: Dataset Shift with Outlier Scores

Test for no adverse shift in two-sample comparison when we have a training set, the reference distribution, and a test set. The approach is flexible (extensible) and relies on a robust and powerful test statistic, the weighted AUC. See Kamulete, V. M. (2021) <arXiv:2107.02990> for details. Outlier scores such as trust scores and prediction uncertainty can be used as the basis for comparison for example.

Version: 0.1.1
Imports: data.table (≥ 1.14.0), future.apply (≥ 1.9.0), ggplot2 (≥ 3.3.3), scales (≥ 1.1.1), simctest (≥ 2.6), stats (≥ 3.6.1)
Suggests: fdrtool (≥ 1.2.16), isotree (≥ 0.2.7), ranger (≥ 0.12.1), knitr (≥ 1.33), rmarkdown (≥ 2.7), testthat (≥ 3.0.2)
Published: 2022-08-16
Author: Vathy M. Kamulete ORCID iD [aut, cre], Royal Bank of Canada (RBC) [cph] (Research supported by RBC)
Maintainer: Vathy M. Kamulete <vathymut at gmail.com>
BugReports: https://github.com/vathymut/dsos/issues
License: GPL (≥ 3)
URL: https://github.com/vathymut/dsos
NeedsCompilation: no
Language: en-US
Materials: NEWS
CRAN checks: dsos results

Documentation:

Reference manual: dsos.pdf
Vignettes: Acknowledgements
Bring Your Own Scores
A 10-minute Crash Course

Downloads:

Package source: dsos_0.1.1.tar.gz
Windows binaries: r-devel: dsos_0.1.1.zip, r-release: dsos_0.1.1.zip, r-oldrel: dsos_0.1.1.zip
macOS binaries: r-release (arm64): dsos_0.1.1.tgz, r-oldrel (arm64): dsos_0.1.1.tgz, r-release (x86_64): dsos_0.1.1.tgz, r-oldrel (x86_64): dsos_0.1.1.tgz
Old sources: dsos archive

Linking:

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