Sebastian Raschka
Sebastian Raschka
Yeah, I think with the new version it will be way easier to track. Regarding "both" there is a difference between floating-forward and floating-backward though. They can both give you...
Thanks for the updates. Actually, some of the failing tests are related to sklearn 1.0. Due to bug fixes or changes, their have been very slight changes in some of...
I am not sure how many people use the floating versions. One option could be to deprecate/remove them. We could maybe add a module `mlxtend.legacy` for that where we can...
Finally got around fixing the CI. Then, I was trying to rebase this PR but since it is not in a feature branch but in the master branch itself, I...
Actually not sure why the workflow CI doesn't run. It worked fine in another PR that I just closed a few minutes ago. Maybe it would be easiest to just...
I am not completely sure if this is related, but maybe the recent change in response to #605 fixes this. I.e., it could have been an issue related to the...
That could probably be relatively easily addressed and definitely sounds useful to avoid confusion in terms of interpreting the results. For future reference to reproduce this issue ```python from mlxtend.plotting...
Kind of related to #589
Thanks for the suggestion, having a stratified bootstrap option could indeed be useful.
Thanks for posting. This sounds annoying indeed and needs to be improved. Maybe a `refresh_rate` parameter could help here, or just moving to `tqdm` like you suggested. It's commonly used...