mlr3
mlr3 copied to clipboard
run_paramtest should be able to generate parameters depending on the tasks
Some learners require (such as priority lass from mlr3extralearners) require to set a parameter to a specific value, that depends on properties of the tasks (in this case, the number of features).
Currently the run_autotest
function does not really support this.
Suggestion: add a parfn
argument that takes in arguments task, learner
and outputs a list with the new parameters.