What do you trade when increasing ncomponents in a Nystroem approximation?
Classic MLLevel: IntermediateApproximation width budget
What do you trade when increasing n_components in a Nystroem approximation?
Hints
- The basis size is also output dimensionality.
- Approximation and computational budgets are linked.
Answer
More basis samples and output features can improve kernel approximation but increase memory, transformation cost and downstream fitting cost. Validate predictive benefit rather than equating width with guaranteed accuracy.
Source and licence: scikit-learn 1.7.2 — kernel approximation · BSD-3-Clause · answer: Quantoros editorial