Skip to content
Quantoros

You want PCA features and selected original features from the same matrix. How does FeatureUnion differ from a…

Classic MLLevel: IntermediateMultiple views of the same inputs

You want PCA features and selected original features from the same matrix. How does FeatureUnion differ from a sequential pipeline?

Hints
  • Parallel views versus a transformation chain.
  • Concatenation increases feature width.

Answer

It fits independent transformer branches on the same input and concatenates their outputs. A sequential pipeline would feed one representation into the next, losing the independent branches.

Source and licence: scikit-learn 1.7.2 — compose · BSD-3-Clause · answer: Quantoros editorial

Related questions