NEWS.md

In FoRecoML: Forecast Reconciliation with Machine Learning

FoRecoML 1.1.1

• Reworked the print() and summary() methods for rml_fit objects, which now return a more informative, structured overview of the fitted reconciliation (framework, machine-learning approach, problem dimensions, features, training sample size, combination matrix and trained models).
• csrml(), terml() and ctrml() now validate their arguments more strictly and fail with informative error messages that state the expected value and report the one actually supplied.

FoRecoML 1.1.0

• Every reconciliation function now returns an object of the new S3 class foreco, defined in FoReco. The objects are built through FoReco's exported new_foreco_class() constructor and therefore integrate seamlessly with FoReco's print(), summary(), plot() and components() methods.

• Revised the title field of every file so that all man page titles consistently follow title case.

FoRecoML 1.0.0

• Cross-sectional, temporal and cross-temporal forecast reconciliation with machine learning
• Initial CRAN submission.

FoRecoML documentation built on July 10, 2026, 1:08 a.m.