CDF: Centroid Decision Forest for High-Dimensional Classification

Implements the Centroid Decision Forest (CDF) as a single user-facing function CDF(). The method selects discriminative features via a multi-class class separability score (CSS), splits by nearest class centroid, and aggregates tree votes to produce predictions and class probabilities. Returns CSS-based feature importance as well. Amjad Ali, Saeed Aldahmani, Zardad Khan (2025) <doi:10.48550/arXiv.2503.19306>.

Version:	0.1.0
Depends:	R (≥ 3.6)
Imports:	matrixStats, utils
Published:	2025-11-10
DOI:	10.32614/CRAN.package.CDF
Author:	Amjad Ali [aut, cre], Saeed Aldahmani [aut], Zardad Khan [aut]
Maintainer:	Amjad Ali <amjadali at uaeu.ac.ae>
License:	MIT + file LICENSE
NeedsCompilation:	no
CRAN checks:	CDF results

Documentation:

Reference manual:

CDF.html , CDF.pdf

Downloads:

Package source:	CDF_0.1.0.tar.gz
Windows binaries:	r-devel: CDF_0.1.0.zip, r-release: CDF_0.1.0.zip, r-oldrel: CDF_0.1.0.zip
macOS binaries:	r-release (arm64): CDF_0.1.0.tgz, r-oldrel (arm64): CDF_0.1.0.tgz, r-release (x86_64): CDF_0.1.0.tgz, r-oldrel (x86_64): CDF_0.1.0.tgz

Linking:

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