Home

Additional learners for mlr3

The following learners are shipped in their own packages to keep the {mlr3learners} package small.
Not all of them have necessarily been developed by the mlr-org team.
Also, learners might possibly be maintained by external persons.

See the section in the {mlr3learners} README on how to add/request a new learner.

Fully implemented

CI	mlr3 Pkg / Alg. / Maintainer	mlr3 Learner
	mlr3learners.C50 C50::C5.0() @henrifnk	classif.C5.0
	mlr3learners.extratrees extraTrees::extraTrees() @be-marc	classif.extraTrees regr.extraTrees
	mlr3learners.fnn FNN::knn() @be-marc	classif.fnn regr.fnn
	mlr3learners.gbm gbm::gbm() @be-marc	classif.gbm regr.gbm
	mlr3learners.ksvm kernlab::ksvm() @be-marc	classif.ksvm regr.ksvm
	mlr3learners.liblinear LiblineaR::LiblineaR() @be-marc	classif.liblinearl1l2svc classif.liblinearl1logreg classif.liblinearl2l1svc classif.liblinearl2l2svc classif.liblinearl2logreg classif.liblinearmulticlasssvc regr.liblinearl2l1svr regr.liblinearl2l2svr
	mlr3learners.mboost mboost::gamboost() mboost::glmboost() @be-marc	classif.gamboost classif.glmboost regr.gamboost regr.glmboost
	mlr3learners.partytkit partykit::ctree() @be-marc	classif.ctree regr.ctree

In progress

CI	mlr3 Pkg / Alg. / Maintainer	mlr3 Learner
	mlr3learners.h2o h2o:glm() h2o:glmr() h2o:gbm() h2o:randomForest() h2o:deeplearning() @be-marc	classif.h2o.glm classif.h2o.glmr classif.h2o.gbm classif.h2o.randomForest classif.h2o.deeplearning regr.h2o.glm regr.h2o.glmr regr.h2o.gbm regr.h2o.randomForest regr.h2o.deeplearning
	mlr3learners.countglm stats::glm() @pkopper	classif.countglm regr.countglm
	mlr3learners.Rborist Rborist::Rborist() @pkopper	classif.Rborist regr.Rborist
	mlr3learners.RWeka RWeka::JRip() @henrifnk	classif.JRip

Installation

It is possible to have a CRAN-like installation via install.packages() by installing the learners from mlr3learners.drat.

Wishlist

• cforest, should be with ctree in one "partykit" package
• classif.nnet (in progress, @henrifnk)
• all algos from klaR (that mlr2 had)
• regr.cubits
• regr.earth
• classif | regr.catboost (+ install_catboost() function)
• every MODERN COOL gaussian process you can find in R