mlr 2.17.0 2020-01-10

plotting

Functional Data

PR: #2638 (@pfistl)

  • Added several learners for regression and classification on functional data

    • classif.classiFunc.(kernel|knn) (knn/kernel using various semi-metrics)
    • (classif|regr).fgam (Functional generalized additive models)
    • (classif|regr).FDboost (Boosted functional generalized additive models)
  • Added preprocessing steps for feature extraction from functional data

    • extractFDAFourier (Fourier transform)
    • extractFDAWavelets (Wavelet features)
    • extractFDAFPCA (Principal components)
    • extractFDATsfeatures (Time-Series features from tsfeatures package)
    • extractFDADTWKernel (Dynamic Time-Warping Kernel)
    • extractFDAMultiResFeatures (Compute features at multiple resolutions)
  • Fixed a bug where multiclass to binaryclass reduction techniques did not work with functional data.

  • Several other minor bug fixes and code improvements

  • Extended and clarified documentation for several fda components.

learners - general

  • xgboost: added options ‘auto’, ‘approx’ and ‘gpu_hist’ to param tree_method (@albersonmiranda, #2701)
  • getFeatureImportance() now returns a long data.frame with columns variable and importance. Beforehand, a wide data.frame was returned with each variable representing a column (@pat-s, #1755).

filters - general

  • Allow a custom threholding function to be passed to filterFeatures and makeFilterWrapper (@annette987, #2686)
  • Allow ensemble filters to include multiple base filters of the same type (@annette987, #2688)

filters - bugfixes

mlr 2.17.1 2020-03-24

Learners - bugfixes

  • remove regr_slim learner due to pkg (flare) being orphaned on CRAN

Measures - bugixes

  • remove measure clValid::dunn and its tests (package orphaned) (#2742)
  • Bugfix: tuneThreshold() now accounts for the direction of the measure. Beforehand, the performance measure was always minimized (#2732).
  • Remove adjusted Rsq measure (arsq), fixes #2711

Filters - bugfixes

  • Fixed an issue which caused the random forest minimal depth filter to only return NA values when using thresholding. NAs should only be returned for features below the given threshold. (@annette987, #2710)
  • Fixed problem which prevented passing filter options via argument more.args for simple filters (@annette987, #2709)

Feature selection - bugfixes

Misc

  • pkgdown: Move changelog to Appendix

  • Account for {checkmate} v2.0.0 update (#2734)

  • Refactor function calls from packages (<pkg::fun>) within ParamSets (#2730) to avoid errors in listLearners() if those pkgs are not installed

  • listLearners() should not fail if a package is not installed (#2717)