Extends the 'mlr3' package and ecosystem to time series
forecasting. Provides forecasting tasks, learners, resampling
strategies, performance measures, and 'mlr3pipelines' operators for
time-series feature engineering. Machine learning regression learners
can be turned into forecasters through recursive and direct multi-step
strategies.
| Version: |
0.1.0 |
| Depends: |
mlr3 (≥ 1.7.0), R (≥ 3.6.0) |
| Imports: |
backports (≥ 1.5.0), checkmate (≥ 2.0.0), cli, data.table (≥ 1.18.0), generics (≥ 0.1.2), lgr, mlr3misc (≥ 0.22.0), mlr3pipelines (≥ 0.11.0), paradox (≥ 1.0.1), R6 (≥ 2.4.1), stats, utils |
| Suggests: |
distributional, fabletools, feasts, forecast (≥ 9.0.2), ggplot2 (≥ 3.4.0), greybox, mlr3tuning, nnfor (≥ 0.9.9), prophet (≥ 1.1.7), Rcatch22, Rlgt (≥ 0.2.3), rpart, smooth (≥ 4.4.0), testthat (≥ 3.2.0), tidyselect, timeSeries, tsbox, tscount (≥ 1.4.3), tsfeatures, tsibble, tsibbledata, vctrs, vdiffr (≥ 1.0.0), withr (≥ 3.0.0), xts, zoo |
| Published: |
2026-07-22 |
| DOI: |
10.32614/CRAN.package.mlr3forecast |
| Author: |
Maximilian Mücke 
[aut, cre],
Marc Becker [aut],
Bernd Bischl
[aut] |
| Maintainer: |
Maximilian Mücke <muecke.maximilian at gmail.com> |
| BugReports: |
https://github.com/mlr-org/mlr3forecast/issues |
| License: |
LGPL-3 |
| URL: |
https://mlr3forecast.mlr-org.com,
https://github.com/mlr-org/mlr3forecast |
| NeedsCompilation: |
no |
| Materials: |
README, NEWS |
| CRAN checks: |
mlr3forecast results |