epkde: Bayesian Bandwidth Selection for Multivariate KDE via Expectation Propagation

Implements the approximate Bayesian method for bandwidth selection in multivariate kernel density estimation (KDE) proposed in Filippone & Sanguinetti (2011) <doi:10.1016/j.csda.2011.05.023>. The method uses the Expectation Propagation (EP) algorithm to approximate the posterior distribution of the inverse bandwidth (precision matrix) under a leave-one-out cross-validated likelihood. Three covariance structures are supported: isotropic (scalar precision), diagonal, and full precision matrix. Online Bayesian updating is supported for the isotropic case. The approximate posterior can be used for bandwidth selection, model comparison (via the model evidence / Bayes factor), and online learning.

Version:	0.1.0
Depends:	R (≥ 4.0.0)
Suggests:	ks, testthat (≥ 3.0.0), knitr, rmarkdown
Published:	2026-09-10
DOI:	10.32614/CRAN.package.epkde (may not be active yet)
Author:	Maurizio Filippone [aut, cre]
Maintainer:	Maurizio Filippone <maurizio.filippone at kaust.edu.sa>
BugReports:	https://github.com/mauriziofilippone/epkde/issues
License:	GPL-3
URL:	https://github.com/mauriziofilippone/epkde
NeedsCompilation:	no
CRAN checks:	epkde results

Documentation:

Reference manual:	epkde.html , epkde.pdf
Vignettes:	Introduction to epkde (source, R code)

Downloads:

Package source:	epkde_0.1.0.tar.gz
Windows binaries:	r-devel: not available, r-release: not available, r-oldrel: not available
macOS binaries:	r-release (arm64): not available, r-oldrel (arm64): epkde_0.1.0.tgz, r-release (x86_64): not available, r-oldrel (x86_64): epkde_0.1.0.tgz

Linking:

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