Christian Lorentzen

Is this now blocked by #28931?

Can you detect the minimum needed pandas version to avoid any error/test failure? (If you can reproduce locally at your computer, you can install the different versions and run the...

A few more: - For classifiers, make `predict` return probabilities and introduce `decide` with an explicit threshold/cutoff argument returning classes. - Make it possible to produce/create a **metadata object** in...

Another idea: - Based on #11865, make one-vs-rest classifier a meta-estimator and take it away from inside `LogisticRegression`. Edit: We already have [`sklearn.multiclass.OneVsRestClassifier`](https://scikit-learn.org/stable/modules/classes.html#module-sklearn.multiclass)

Very interesting, indeed. Let‘s bring some structure. #### 1. To the original question whether to switch pandas examples to polars examples I think sooner or later we should have a...

Ok, it seems that we have the "2. Making scikit-learn dataframe native" discussion in this issue anyway. I few personal thoughts: **I would love to have a dataframe native ML...

Those doc updates often take longer than anticipated :smirk:

> The comments of my first pass of review were marked as resolved without any pushed code change to address them nor any discussion. Sorry, I forgot to push. Will...

One problem with the current way of just setting sample weight to zero is that the sample counts in histograms is wrong and `min_samples_leaf` does not work as it should....

@marenwestermann Could NuSVC added in another existing example? The general idea is to have less examples. I could also live with just remaining this one ( => I approved).