Michael Mayer
Michael Mayer
At least the path variable was pointing to C:/rtools40... let's see what I get with clean environment variables.
A different error message pops up: Is g++ version too high? ``` D:\xgboost>R CMD INSTALL --with-keep.source ./R-package * installing to library 'C:/Users/Michael/AppData/Local/R/win-library/4.3' * installing *source* package 'xgboost' ... ** using...
@trivialfis Thanks your your patient assistance! A simple `--preclean` was now sufficient to build the library :-). ``` R CMD INSTALL --preclean --with-keep.source ./R-package ```
@jameslamb Maybe we can dump/parse single trees (or m=NULL trees) instead of the full model.
@jameslamb: Very neat! I can work on this after https://github.com/microsoft/LightGBM/pull/6364 is merged.
Okay, agreed. The wish to apply such transform in the explainers will raise as soon as we get interaction importance measues like Friedman's H. There, it will be essential to...
> > I don't get that suggestion. y in fit and the return value of predict must > > be on the same scale, or metrics don't work. > >...
This would be extremely useful. In the currently available inspection tools, its indeed only partial dependence that would need to be touched. (Later also H-statistics drafted in this [PR](https://github.com/scikit-learn/scikit-learn/pull/28375)). In...
Try "binary_logloss" instead of "logloss".
Very good idea. I have in mind that it is also in the `fit()` method of the Scikit-learn API. Do we need it there because that is the first time...