[R] How to evaluate a learner with multiple resampling in mlr3?

[Neha gupta]

Hello everyone,

I want to evaluate the cart learner with different (i.e. ideally 10
different) resampling methods. My questions are:

(1) How should I specify the list of resampling methods?
(2) Should I specify the resampling lists in the beginning or in the
benchmark?
(3) Currently mlr3 shows bootstrap method, can we use the variants of
bootstrap like boot632, optimistic boot like we use in caret?

d=readARFF("mall.csv")

task=TaskRegr$new("d", d, target = "Price")
learner= lrn("regr.rpart")
resampling = rsmp("bootstrap", repeats=100)

grid = benchmark_grid( task = task,  learner = learner,  resampling)
bmr = benchmark(grid)

	[[alternative HTML version deleted]]