icomb: Forecasting Hierarchical Time Series Using Information Combination

Implements the Information Combination (IComb) approach proposed by Nguyen, Vahid and Wickramasuriya (2025)<https://www.monash.edu/business/ebs/research/publications/ebs/2025/wp11-2025.pdf> for hierarchical forecast reconciliation. The method combines information from base forecasts constructed using different information sets while ensuring coherence. It is implemented using a penalized regression-based framework.

Version: 0.2.0
Depends: R (≥ 4.1.0)
Imports: cli, distributional, dplyr, fabletools (≥ 0.7.0), future.apply, generics, glmnet, progressr, purrr, tsibble, vctrs
Suggests: broom, fable, feasts, forecast, future, ggplot2, ggtime, hts, knitr, lubridate, plotly, rmarkdown, spelling, testthat (≥ 3.0.0), tibble, tidyr
Published: 2026-05-27
DOI: 10.32614/CRAN.package.icomb (may not be active yet)
Author: Shanika Wickramasuriya [aut, cre, cph], Mitchell O'Hara-Wild [ctb], Nuwani Palihawadana [ctb], Minh Nguyen [ctb]
Maintainer: Shanika Wickramasuriya <shanika.wickramasuriya at monash.edu>
BugReports: https://github.com/ShanikaLW/icomb/issues
License: GPL (≥ 3)
URL: https://shanikalw.github.io/icomb/, https://github.com/ShanikaLW/icomb
NeedsCompilation: no
Language: en-US
Citation: icomb citation info
Materials: README, NEWS
CRAN checks: icomb results

Documentation:

Reference manual: icomb.html , icomb.pdf
Vignettes: icomb (source, R code)

Downloads:

Package source: icomb_0.2.0.tar.gz
Windows binaries: r-devel: not available, r-release: not available, r-oldrel: not available
macOS binaries: r-release (arm64): icomb_0.2.0.tgz, r-oldrel (arm64): icomb_0.2.0.tgz, r-release (x86_64): icomb_0.2.0.tgz, r-oldrel (x86_64): icomb_0.2.0.tgz

Linking:

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