Christian Lorentzen

> I checked `lightgbm` and `xgboost`, they both take a list of tuples as validation sets, which makes me think why, and if this proposal is enough! Let's ask them...

This will be a great addition! Thanks @mayer79

The failing test is in the python 3.9 env with numpy 1.22.0 polars 1.0.0 scipy 1.10.0 pandas 1.5.3 pyarrow 11.0.0 Could you check if increasing one of the versions fixes...

fyi, CI will fail due to new versions of polars and numpy. I am working on a fix.

Fix in #203, you need to sync (e.g. merge) with the main branch (and maybe `hatch env prune` on your local machine).

fyi: I am preparing to bump the minimum versions of python to 3.11 and numpy to 2. This implies polars 1.1.0, pandas >= 2.2.2 and pyarrow >= 16, see #206.

You may be interested in a similar discussion in scikit-learn, see https://github.com/scikit-learn/scikit-learn/issues/14251. There, the discussion about entry gate for dataframes, the `ColumnTransformer`, was resolved by taking care about column order...

While feature names with `FuncTransformer` certainly helps, the still open question is: Do we want to add such date time related transformers/estimators in scikit-learn? Options: 1. Include a dedicated `DateTimeTransformer`...

The link function is the inverse of what you usually apply in reality: - Binary Logistic: `return of predict_proba = expit(raw)` and link function is `logit` - (Log-)-Poisson regression: `return...

In principle, loss and link function are independent. It's only an efficient implementation that needs to merge them together. As written in the RFC, having link and inverse link function...