permimp: Conditional Permutation Importance

An add-on to the 'party' package, with a faster implementation of the partial-conditional permutation importance for random forests. The standard permutation importance is implemented exactly the same as in the 'party' package. The conditional permutation importance can be computed faster, with an option to be backward compatible to the 'party' implementation. The package is compatible with random forests fit using the 'party' and the 'randomForest' package. The methods are described in Strobl et al. (2007) <doi:10.1186/1471-2105-8-25> and Debeer and Strobl (2020) <doi:10.1186/s12859-020-03622-2>.

Version: 1.0-2
Depends: R (≥ 3.6.0)
Imports: graphics, grDevices, ipred (≥ 0.9-6), methods, party (≥ 1.3-3), randomForest (≥ 4.6-14), stats, survival (≥ 2.44-1.1), utils
Suggests: knitr, rmarkdown, scales (≥ 0.5.0), testthat
Published: 2021-09-13
DOI: 10.32614/CRAN.package.permimp
Author: Dries Debeer [aut, cre], Torsten Hothorn [aut], Carolin Strobl [aut]
Maintainer: Dries Debeer <debeer.dries at>
License: GPL-2 | GPL-3
NeedsCompilation: no
Materials: NEWS
CRAN checks: permimp results


Reference manual: permimp.pdf
Vignettes: The permimp-package


Package source: permimp_1.0-2.tar.gz
Windows binaries: r-devel:, r-release:, r-oldrel:
macOS binaries: r-release (arm64): permimp_1.0-2.tgz, r-oldrel (arm64): permimp_1.0-2.tgz, r-release (x86_64): permimp_1.0-2.tgz, r-oldrel (x86_64): permimp_1.0-2.tgz
Old sources: permimp archive

Reverse dependencies:

Reverse imports: rfvimptest


Please use the canonical form to link to this page.