mlr3 package updates - q4/2022

This posts gives an overview by listing the recent release notes of mlr3 packages from the last quarter. Cover photo by Aaron Burden.


January 12, 2023

Due to the high amount of packages in the mlr3 ecosystem, it is hard to keep up with the latest changes across all packages. This posts gives an overview by listing the recent release notes of mlr3 packages from the last quarter. Note that only CRAN packages are listed here and the sort order is alphabetically.

mlr3cluster 0.1.6

Description Cluster Extension for ‘mlr3’

  • Add states as row names to usarrest task.
  • Remove dictionary items after unloading package.

mlr3cluster 0.1.5

  • Added Mclust learner
  • Fix error associated with new dbscan release

mlr3tuning 0.17.2

Description Hyperparameter Optimization for ‘mlr3’

  • feat: AutoTuner accepts instantiated resamplings now. The AutoTuner checks if all row ids of the inner resampling are present in the outer resampling train set when nested resampling is performed.
  • fix: Standalone Tuner did not create a ContextOptimization.

mlr3tuning 0.17.1

  • fix: The ti() function did not accept callbacks.

mlr3tuning 0.17.0

  • feat: The methods $importance(), $selected_features(), $oob_error() and $loglik() are forwarded from the final model to the AutoTuner now.
  • refactor: The AutoTuner stores the instance and benchmark result if store_models = TRUE.
  • refactor: The AutoTuner stores the instance if store_benchmark_result = TRUE.

mlr3tuning 0.16.0

  • feat: Add new callback that enables early stopping while tuning to mlr_callbacks.
  • feat: Add new callback that backups the benchmark result to disk after each batch.
  • feat: Create custom callbacks with the callback_tuning() function.

mlr3tuning 0.15.0

  • fix: AutoTuner did not accept TuningSpace objects as search spaces.
  • feat: Add ti() function to create a TuningInstanceSingleCrit or TuningInstanceMultiCrit.
  • docs: Documentation has a technical details section now.
  • feat: New option for extract_inner_tuning_results() to return the tuning instances.

mlr3fselect 0.9.0

Description Feature Selection for ‘mlr3’

  • fix: Add genalg to required packages of FSelectorGeneticSearch.
  • feat: Add new callback that backups the benchmark result to disk after each batch.
  • feat: Create custom callbacks with the callback_fselect() function.

mlr3fselect 0.8.0

  • refactor: FSelectorRFE throws an error if the learner does not support the $importance() method.
  • refactor: The AutoFSelector stores the instance and benchmark result if store_models = TRUE.
  • refactor: The AutoFSelector stores the instance if store_benchmark_result = TRUE.
  • feat: Add missing parameters from AutoFSelector to auto_fselect().
  • feat: Add fsi() function to create a FSelectInstanceSingleCrit or FSelectInstanceMultiCrit.
  • refactor: Remove unnest option from function.

mlr3mbo 0.1.1

Description Flexible Bayesian Optimization

  • Initial CRAN upload

mlr3spatial 0.3.1

Description Support for Spatial Objects Within the ‘mlr3’ Ecosystem

  • chore: Remove rgdal dependency and require raster version 3.6-11.

mlr3spatial 0.3.0

  • feat: Add prediction on vector data to spatial_predict().

mlr3oml 0.7.0

Description Connector Between ‘mlr3’ and ‘OpenML’

  • feature: Add argument task_type to function list_oml_tasks().
  • fix: strings and nominals are distinguished for parquet files
  • docs: Fixed some OpenML links
  • docs: Renamed the docs for OpenML objects
  • Renamed the sugar functions from:
    • oml_data() is now odt()
    • oml_task()is now otsk()
    • oml_flow() is now oflw()
    • oml_run() is now orn
    • oml_collection() is now ocl()
  • Addresses a CRAN issue: examples fail gracefully if OpenML server is busy.

bbotk 0.7.2

Description Black-Box Optimization Toolkit

  • fix: Standalone Tuner and FSelector were rejected by ContextOptimization.

bbotk 0.7.1

  • feat: Data unrelated to a specific point evaluation can be written to Archive$data_extra.

bbotk 0.7.0

  • fix: Terminator$format(with_params = TRUE) printed an empty list when no parameter was set.
  • refactor: OptimizerIrace automatically added the instances parameter to Objective$constants. From now on, the instances parameter can be also set manually.
  • BREAKING CHANGE: branin(xs) is now branin(x1, x2, noise) and branin_wu(x1, x2, fidelity).
  • feat: Add ObjectiveRFunMany that can evaluate a list of configurations with a user supplied function.
  • fix: If all configurations were missing a parameter, ObjectiveRFunDt$eval_many() did not create a column with NA for the missing parameter.
  • refactor: The default of digits in OptimizerIrace is 15 now to avoid rounding errors.
  • refactor: The bounds of double parameters were processed with only 4 decimal places in OptimizerIrace. By default, the bounds of double parameters are represented with 15 decimal places now. The digits parameter of OptimizerIrace also changes number of decimal places of the bounds now.

bbotk 0.6.0

  • fix: OptimizerIrace did not work with parameters with multiple dependencies.
  • feat: Add new callback that backups the archive to disk to mlr_callbacks.
  • feat: Create custom callbacks with the callback_optimization() function.

mlr3tuningspaces 0.3.3

Description Search Spaces for ‘mlr3’

  • fix: Extra paradox::TuneToken in lts() were not passed to learners created with $get_learner().
  • docs: Add lts() return.

mlr3tuningspaces 0.3.2

  • docs: Add mlr_tuning_spaces prefix to aliases.

mlr3tuningspaces 0.3.1

  • docs: Add glmnet description.

mlr3hyperband 0.4.4

Description Hyperband for ‘mlr3’

  • fix: Remove emoa from required packages of OptimizerSuccessiveHalving.

mlr3hyperband 0.4.3

  • docs: Examples use branin_wu() function now.

mlr3spatiotempcv 2.0.3

Description Spatiotemporal Resampling Methods for ‘mlr3’

  • add label support for built-in tasks
  • adhere to CRAN “noSuggests” policy

mlr3verse 0.2.7

Description Easily Install and Load the ‘mlr3’ Package Family

  • Updated reexports.

mlr3verse 0.2.6

  • Updated reexports.

mlr3 0.14.1

Description Machine Learning in R - Next Generation

  • Removed depdency on package distr6.
  • Fixed reassembling of GraphLearner.
  • Fixed bug where the measured elapsed time was 0:
  • Fixed as_prediction_classif() for data.frame() input (#872).
  • Improved the error message when predict type of fallback learner does not match the predict type of the learner (mlr-org/mlr3extralearners#241).
  • The test set is now available to the Learner during train for early stopping.

mlr3benchmark 0.1.4

Description Analysis and Visualisation of Benchmark Experiments

  • Add friedman_global argument to posthoc tests and to autoplots to allow methods and plots to run even if the global Friedman test fails (i.e. don’t reject null)
  • New maintainer: Sebastian Fischer
  • Fix documentation

paradox 0.11.0

Description Handling parameter spaces

  • Minor Bug Fixes