mlr3extralearners 1.5.2

Other

• Use CRAN version of survdistr.
• Use mlr3cmprsk version 0.0.5.
• Update crs parameters.

mlr3extralearners 1.5.1

Other

• Skip fastai and botorch tests on Windows and macOS where the Python
  backends crash or time out.
• Skip tabpfn tests until token work reliable again.
• Skip blockForest tests on macOS where SE predictions fail sanity checks.
• Skip h2o.glm classification tests on Windows due to Java NullPointerException.
• Skip GPfit tests on Windows where they crash under R-devel.
• Skip classif.aorsf sanity autotest due to inconsistent tie-breaking across
  predict types.
• Skip surv.flexreg sanity autotest on Windows due to initial parameter
  estimation failure.

mlr3extralearners 1.5.0

New Features

• New Learners:
  • LearnerCompRisksCoxboost
  • LearnerRegrGPfit
  • LearnerClassifMLP
  • LearnerClassifSaeDNN
  • LearnerClassifPlsdaCaret
  • LearnerSurvDNN
  • LearnerRegrH2ORandomForest
  • LearnerRegrH2OGLM
  • LearnerClassifH2OGLM
  • LearnerClassifH2OGBM
  • LearnerClassifH2ORandomForest
  • LearnerClassifH2ODeeplearning
  • LearnerRegrH2OGBM
  • LearnerRegrH2ODeeplearning
  • LearnerClassifLvq1
  • LearnerRegrBotorchFullyBayesian

• Added kernel and input/output transformations for LearnerRegrBotorchSingleTaskGP and LearnerRegrBotorchMixedSingleTaskGP.

Breaking Changes

• Moved LearnerSurvAkritas and LearnerSurvParametric to the attic.
  See #549.

Other

• Updated Extending vignette to incorporate information about skipping tests and considerations for testing Python learners
• survdistr is now on Suggests (used for constant interpolation of the Kaplan-Meier predictions of the partykit survival learners)
• Updated mlr3proba (0.8.8), pls and xgboost to the most recent CRAN versions

mlr3extralearners 1.4.0

New Features

• New Learners:
  • LearnerSurvGamCox
  • LearnerSurvFlexReg
  • LearnerSurvNCVsurv
  • LearnerRegrRRF
  • LearnerRegrPcr
  • LearnerRegrPlsr
  • LearnerRegrLaGP
  • LearnerRegrFrbs
  • LearnerRegrBcart
  • LearnerRegrBgp
  • LearnerRegrBgpllm
  • LearnerRegrBlm
  • LearnerRegrBtgp
  • LearnerRegrBtgpllm
  • LearnerRegrBtlm
  • LearnerRegrNCVReg
  • LearnerClassifDbnDNN
  • LearnerClassifNNTrain
  • LearnerClassifSparseLDA
  • LearnerClassifNCVreg

Contributors: @bblodfon @awinterstetter @be-marc

Breaking Changes

• lrn("surv.flexible") (LearnerSurvFlexible) was renamed to lrn("surv.flexsurvspline") (LearnerSurvFlexSpline) to properly reflect the wrapped train function (Royston/Parmar spline model).

Other

• CoxBoost is now on CRAN, so we removed it from Remotes
• lrn("surv.flexsurvspline") predicts linear predictors using predict.flexsurvreg(). We were doing manually the same exact prediction, so no functionality was changed.
• compatibility: xgboost 3.1.2.1 (survival learners)
• parameter updates for regr.lmer/glmer learners
• updates for randomForestSRC 3.5.0 (use.uno parameter)
• performance improvement: use of data.table::fifelse (@m-muecke)

1.3.1

• Update website to include citation information

1.3.0

• Add formula and anc params to surv.flexible learner, as well as response predict type (mean survival time).
• Fix regr.gamboost regression predictions (#498).

1.2.0

New Features

• New Learners:

  • LearnerCompRisksRandomForestSRC
  • LearnerSurvBlockForest
  • Learner{Classif,Regr,Surv}BlockForest
  • Learner{Classif,Regr}ExhaustiveSearch
  • LearnerClassifFastai
  • Learner{Classif,Regr}Penalized
  • Learner{Classif,Regr}Bst
  • LearnerClassifAdabag
  • LearnerClassifAdaBoosting
  • Learner{Classif,Regr}Evtree
  • LearnerClassifKnn
  • LearnerClassifRotationForest
  • LearnerRegrCrs
  • LearnerClassifStepPlr
  • LearnerClassifMda
  • LearnerClassifRferns
  • LearnerClassifNeuralnet
  • LearnerRegrBrnn
  • LearnerRegrBotorchSingleTaskGP
  • LearnerRegrBotorchMixedSingleTaskGP

• Add new control_custom_fun parameter in surv.aorsf

• New function learner_is_runnable() to check whether the
  required packages to train a learner are available.

• Added selected_features property to RandomForestSRC learners (prediction doesn't work if vars.used = 'all.trees')

Bug fixes

• Tests are now skipped when the suggested packages is not available.
  This will make local development much more convenient.
• Removed parameters from RandomForestSRC learners that weren't used + optimized tests
• Removed discrete parameter from surv.parametric, so that it is impossible to return distr6::VectorDistribution survival predictions (softly deprecated in mlr3proba@v0.8.1)

Breaking Changes

• All (extra) density learners are removed. These will be transferred to mlr3proba soon (see v0.8.2 or later).
• The create_learner() generator was removed, because it was hard to maintain and boilerplate code in the age of LLMs is easy enough to write.
• remove discrete parameter from surv.parametric, so that it is impossible to return distr6::VectorDistribution
  survival predictions (softly deprecated in mlr3proba@v0.8.1)
• classif.lightgbm now works with encapsulation with multiclass tasks
• the package no longer re-exports lrn and lrns, which should anyway
  be available to the user as the package depends on mlr3, where these
  functions are defined.
• Removed various learners:
  • randomPlantedForest was removed, because there is currently no way to
    save the model.
  • The deep learning methods from survivalmodels were removed, because
    they also cannot be saved and because the upstream package is archived.

Other

• The package now imports withr
• mlr3proba is now an import and no longer a suggested package.
• mlr3cmprsk is added as an import.
• The package no longer uses set.seed() in the tests and instead uses withr::local_seed()
  This means the auto tests will be stochastic like they should be.
• The CI now checks that RCMD-check passes when suggested packages are not available.
• distr6 dependency is removed. partykit survival learners use constant
  interpolation of the predicted Kaplan-Meier curves via survdistr::vec_interp()

1.1.0

See NEWS.md

1.0.0

• Add "Prediction types" doc section for all 30 survival learners + make sure it is consistent #347
• All survival learners have crank as main prediction type (and it is always returned) #331
• Added minimum working version for all survival learners in DESCRIPTION file
• Harmonized the use of times points for prediction as much as possible across survival learners #387
  • added gridify_times() function to coarse time points
  • fixed surv.parametric and surv.akritas use of ntime argument
• surv.parametric is now used by default with discrete = TRUE (no survival learner returns now distr6 vectorized distribution by default)
• Doc update for mlr3 (version 0.21.0)
• Fixed custom and initial values across all learners documentation pages
• Fixed doc examples that used learner$importance()
• Set n_thread = 1 for surv.aorsf and use unique event time points for predicted S(t)
• Add selected_features() for surv.penalized
• Fix surv.prioritylasso learner + add distr predictions via Breslow #344
• Survival SVM gamma.mu parameter was split to gamma and mu to enable easier tuning (surv.svm learner)

0.9.0

See NEWS.md