mlr3tuningspaces (version 0.5.2)

mlr_tuning_spaces_default: Default Tuning Spaces

Description

Tuning spaces from the Bischl (2023) article.

glmnet tuning space

• s \([1e-04, 10000]\) Logscale

• alpha \([0, 1]\)

kknn tuning space

• k \([1, 50]\) Logscale

• distance \([1, 5]\)

• kernel [“rectangular”, “optimal”, “epanechnikov”, “biweight”, “triweight”, “cos”, “inv”, “gaussian”, “rank”]

ranger tuning space

• mtry.ratio \([0, 1]\)

• replace [TRUE,FALSE]

• sample.fraction \([0.1, 1]\)

• num.trees \([1, 2000]\)

rpart tuning space

• minsplit \([2, 128]\) Logscale

• minbucket \([1, 64]\) Logscale

• cp \([1e-04, 0.1]\) Logscale

svm tuning space

• cost \([1e-04, 10000]\) Logscale

• kernel [“polynomial”, “radial”, “sigmoid”, “linear”]

• degree \([2, 5]\)

• gamma \([1e-04, 10000]\) Logscale

xgboost tuning space

• eta \([1e-04, 1]\) Logscale

• nrounds \([1, 5000]\)

• max_depth \([1, 20]\)

• colsample_bytree \([0.1, 1]\)

• colsample_bylevel \([0.1, 1]\)

• lambda \([0.001, 1000]\) Logscale

• alpha \([0.001, 1000]\) Logscale

• subsample \([0.1, 1]\)