Why should standardized features with negative values not be fed directly into chi2 feature selection?
Classic MLLevel: IntermediateChi-square feature constraints
Why should standardized features with negative values not be fed directly into chi2 feature selection?
Hints
- The score function has input assumptions.
- Scaling can invalidate them.
Answer
This test expects nonnegative feature values such as counts or frequencies. Use appropriate original/count features or choose a test suitable for the transformed data.
Source and licence: scikit-learn 1.7.2 — feature selection · BSD-3-Clause · answer: Quantoros editorial