Feature names stored in `object` and `newdata` are different! using mlr package about mlr HOT 4 CLOSED

To my knowledge this Q should have been answered multiple times already on the web / other repos.

Besides, I suggest using mlr3 instead of mlr.

from mlr.

Can you provide me a link? I cannot find it anywhere.
Further, the mlr3 that you suggested does not support multilabel classification yet. I already aswered a previous post from you about that here https://stackoverflow.com/questions/67387608/warning-unknown-or-uninitialised-column-ntree-when-trying-to-pass-hyperpar?noredirect=1#comment119132750_67387608

from mlr.

I can't provide a link, sorry. I've seen this error in relation to XGBoost and ranger multiple times already and it usually came down to how variables were treated within the respective package.

I think it should be possible to find a solution in the web - I don't think this is a new issue. Not entirely sure though of course.

Thanks for your reprex. Next time, please include it as one block for easy c/p - preferably using the {reprex} package.

from mlr.

@agneswattel Sorry I could not be of any help here. Did you manage to find a way around?

Here is a link from the xgboost GH repo - though I am not sure if this is what I mean in my previous comment.

In mlr3 multilabel support does not yet exist as stated in the mlr3book - posting an issue in {mlr3multioutput} could help to lift the topic in priority :)

from mlr.