ONAM: Fitting Interpretable Neural Additive Models Using Orthogonalization

An algorithm for fitting interpretable additive neural networks for identifiable and visualizable feature effects using post hoc orthogonalization. Fit custom neural networks intuitively using established 'R' 'formula' notation, including interaction effects of arbitrary order while preserving identifiability to enable a functional decomposition of the prediction function. For more details see Koehler et al. (2025) <doi:10.1038/s44387-025-00033-7>.

Version: 1.0.0
Depends: keras3, reticulate
Imports: dplyr, scales, rlang, ggplot2, pROC
Suggests: akima, RColorBrewer, testthat (≥ 3.0.0)
Published: 2025-11-11
DOI: 10.32614/CRAN.package.ONAM (may not be active yet)
Author: David Köhler ORCID iD [aut, cre]
Maintainer: David Köhler <koehler at imbie.uni-bonn.de>
BugReports: https://github.com/Koehlibert/ONAM_R/issues
License: MIT + file LICENSE
NeedsCompilation: no
Materials: README
CRAN checks: ONAM results

Documentation:

Reference manual: ONAM.html , ONAM.pdf

Downloads:

Package source: ONAM_1.0.0.tar.gz
Windows binaries: r-devel: not available, r-release: not available, r-oldrel: ONAM_1.0.0.zip
macOS binaries: r-release (arm64): ONAM_1.0.0.tgz, r-oldrel (arm64): ONAM_1.0.0.tgz, r-release (x86_64): ONAM_1.0.0.tgz, r-oldrel (x86_64): ONAM_1.0.0.tgz

Linking:

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