Question
Why should features usually be scaled before SVM training?
Select an option. Your answer will be checked instantly.
Correct Answer: A. Feature magnitude affects dot products, distances, and regularization
Explanation:
Large-scale features can dominate the geometry of the optimization.
Consistent scaling also makes C and gamma tuning more meaningful.
Leave a Reply