Skip to contents

Changelog

Source: NEWS.md

mlr3fda (development version)

  • fix: Stricter parameter checks with informative errors in PipeOpFDACor, PipeOpFDADerive, PipeOpFDAExtract, PipeOpFDAIntegrate, PipeOpFDAInterpol, PipeOpFDAScaleRange, PipeOpFDAWavelets, and PipeOpFDAZoom.
  • fix: PipeOpFDABsignal now uses the argument values instead of index positions, which changes the features for grids other than 1, ..., p.
  • fix: PipeOpFDARegister no longer fails during prediction when the "srvf" method gets penalty_method in args.
  • fix: PipeOpFDAScaleRange now keeps the evaluator and maps the domain to [lower, upper].
  • fix: PipeOpFDAScaleRange now stores its state under $state$trafos.
  • fix: PipeOpFDAWavelets no longer supports irregular functional data, where coefficients of curves with different lengths were misaligned. Use PipeOpFDAInterpol first.
  • fix: PipeOpFPCA now uses the n_components from training during prediction.

mlr3fda 0.7.2

CRAN release: 2026-08-21

  • docs: Various documentation improvements.
  • fix: PipeOps no longer forward affect_columns to their backend functions.
  • fix: PipeOpFDAInterpol now reports which grid, left, and right combination is invalid instead of failing with an opaque assertion.
  • fix: PipeOpFDAInterpol now errors when left and right fall outside the domain of a functional column.

mlr3fda 0.7.1

CRAN release: 2026-07-15

  • Compatibility with new tf version 0.5.0.

mlr3fda 0.7.0

CRAN release: 2026-07-02

  • fix: PipeOpFDAWavelets no longer errors when filter is a wt.filter object or a numeric vector. Generated columns are now named <column>_wav_<i> regardless of the filter.
  • fix: PipeOpFDAZoom now accepts a one-sided begin or end instead of requiring both.
  • fix: as.data.table(mlr_pipeops) no longer errors with object 'value' not found when mlr3fda is loaded.
  • feat: New PipeOpFDACatch22 for extracting the catch22 time series features from functional columns via Rcatch22::catch22_all().
  • feat: New PipeOpFDAIntegrate for extracting the definite integral of functional columns via tf::tf_integrate().

mlr3fda 0.6.0

CRAN release: 2026-06-01

  • fix: Add mlr3fda to mlr_reflections$loaded_packages to fix errors when using mlr3fda in parallel.
  • feat: New PipeOpFDADepth for computing the data depth of functional columns via tf::tf_depth().
  • feat: New PipeOpFDADerive for computing derivatives of functional columns via tf::tf_derive().
  • feat: PipeOpFDAExtract gained the "sd" feature for extracting the standard deviation.
  • perf: PipeOpFDAFourier is now several times faster.

mlr3fda 0.5.0

CRAN release: 2026-05-03

  • feat: New PipeOpFDARegister for aligning functional columns to a learned template via tf::tf_register().

mlr3fda 0.4.0

CRAN release: 2026-03-30

  • fix: The fuel task now correctly names the scalar column h2o instead of h20.
  • feat: Mlr3Error and Mlr3Warning classes for errors and warnings.
  • feat: New PipeOpFDAFourier for extracting fast Fourier transform features from functional columns.

mlr3fda 0.3.0

CRAN release: 2025-10-15

  • mlr3fda now depends on R 4.1.0 instead of R 3.1.0 to reflect tf requiring 4.1.0
  • New PipeOps:
    • PipeOpFDABsignal
    • PipeOpFDARandomEffect
    • PipeOpFDATsfeatures
    • PipeOpFDAWavelets
    • PipeOpFDAZoom

mlr3fda 0.2.0

CRAN release: 2024-07-22

  • New PipeOps:
    • PipeOpFDACor
    • PipeOpFDAScaleRange
  • mlr3fda now depends on mlr3pipelines

mlr3fda 0.1.1

CRAN release: 2024-04-09

  • Fix CRAN issues

mlr3fda 0.1.0

CRAN release: 2024-04-04

Initial CRAN release.