Thomas J. Fan

At this point, I think it's worth adding a new example section for "Developing Estimators".

I am +1 with adding a more informative error message for transformers with `sparse_output`.

I agree the warning is a false positive. I opened https://github.com/scikit-learn/scikit-learn/pull/26944 to fix it.

I think the comments from @ogrisel is addressed. Specifically, the updating wording for "cross-fitting" addresses the open comment in https://github.com/scikit-learn/scikit-learn/pull/26677#discussion_r1241886607. With that I am going to merge, I think the...

> casing of the home page navbar was intentional or not. Maybe @thomasjpfan knows? It was intentional, but I am okay with this PR which makes it consistent.

We have went with deprecations over breaking changes for 1.0. I feel like it will be hard to change this philosophy for 2.0. If we must have breaking changes, I...

I am +1 on downgrading our PyPy support.

Panda's block manager that supports 2D arrays is still nicer memory-wise when it comes to the roundtrip: `numpy` -> `dataframe` -> `numpy`. Pandas can hold on to numerical F-contiguous or...

@MarcoGorelli > - preprocess in Polars > - call .to_pandas() before passing to a scikit-learn Pipeline There are two types of examples: - If the preprocessing in polar does not...

Currently, we use `copy=False` when wrapping the `ndarray` to avoid the copy: ```python import pandas as pd import numpy as np pd.set_option("mode.copy_on_write", True) rows = int(1e8) cols = 5 x...