Move from party to partykit

[HeidiSeibold]

partykit is still being actively moved forwards while party is not. So mlr should switch to using partykit algorithms (see also #1481). For this we need to update:

• [x] classif_ctree
• [x] regr_ctree
• [x] classif_cforest
• [x] regr_cforest
• [x] multilabel_cforest
• [ ] regr_mob
• [ ] tests for classif_ctree
• [ ] tests for regr_ctree
• [ ] tests for classif_cforest
• [ ] tests for regr_cforest
• [ ] tests for multilabel_cforest
• [ ] tests for regr_mob

What would be nice would be to add

• [ ] classif_glmtree
• [ ] regr_glmtree
• [ ] regr_lmtree

I started working on this in https://github.com/mlr-org/mlr/tree/partykit-transition

[HeidiSeibold]

I could use help with:

• [ ] Checking if the parameter sets are correct. In particular I was unsure with perturb in cforest (see https://github.com/mlr-org/mlr/blob/partykit-transition/R/RLearner_classif_cforest.R#L16)
• [ ] How to update the tests. It is not totally clear to me why the tests currently are the way they are.

[stale[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.