Closed mnwright closed 4 years ago
Hi @mnwright, thanks for raising the issue! Is it as simple as merging this pull request, or do I need to implement the other half of the condition for ranger >=0.11.5?
Yes, this should work with both versions. We don't have the outcome variable y
in the split variable IDs anymore because we separated x
and y
data structures, so that NA
in the front is not needed anymore.
Your tests passed with both old and new ranger versions.
That's great. I have merged your pull request. Thanks for the help!
Fixes issue here https://cran.r-project.org/web/checks/check_results_tree.interpreter.html.
We changed the internal data structure in ranger (now separate X and Y data).