FEA Support missing-values in ExtraTrees*

[adam2392]

Reference Issues/PRs

Follow-up to: #27966

Closes: #27931

What does this implement/fix? Explain your changes.

• Adds support and tests for missing-values in the ExtraTrees*

Any other comments?

This should be a very easy review as this just activates the usage of missing value support and adds unit-test parameterizations for the ExtraTree Forest. Just requesting reviews from @thomasjpfan, @glemaitre and @OmarManzoor since you all were able to review the code in #27966. Hope this is easy :)

Waiting for #27966 to be merged first. This just shows the relevant code changes that introduces the missing-value support overall for ExtraTrees. Note the bulk of the work for going from tree->forest missing-value support was added in #26391. The main change here is to introduce certain unit tests.

• benchmark script ran to demonstrate ExtraTrees* perform much better on data with missing values compared to naive imputation on main
• benchmark script to compare ExtraTree* compared to naive imputation on main

Benchmarks on ExtraTrees demonstrating there is no runtime regression

[github-actions]

✔️ Linting Passed

All linting checks passed. Your pull request is in excellent shape! ☀️

_{Generated for commit: 963de46. Link to the linter CI: here}

[OmarManzoor]

LGTM. Thanks @adam2392. Would it be suitable for ExtraTreesRegressor to be added to test_non_supported_criterion_raises_error_with_missing_values?

[adam2392]

Good point @OmarManzoor!

Added in 1441925

[thomasjpfan]

Otherwise, LGTM