How can OneHotEncoder limit a high-cardinality feature without manually listing rare values?
Classic MLLevel: IntermediateCategory cardinality control
How can OneHotEncoder limit a high-cardinality feature without manually listing rare values?
Hints
- Group before encoding.
- Frequency rules must be learned on training rows.
Answer
Use min_frequency or max_categories to group infrequent categories. Fit grouping only on training data and check how unseen values map at inference.
Source and licence: scikit-learn 1.7.2 — preprocessing · BSD-3-Clause · answer: Quantoros editorial