[Research] Use grouping of infrequent categories in `OneHotEncoder`

[eddiebergman]

Feature preprocessing.OneHotEncoder now supports grouping infrequent categories into a single feature. Grouping infrequent categories is enabled by specifying how to select infrequent categories with min_frequency or max_categories. https://github.com/scikit-learn/scikit-learn/pull/16018 by Thomas Fan.

Mentioned in #1611

This would need to be researched but could perform great opportunities in low memory / high feature class count cases