Skip to content
Quantoros

What is feature scaling and which algorithms actually need it?

Classic MLLevel: FoundationClassical machine learning

What is feature scaling and which algorithms actually need it?

Hints
  • State the task, inputs and prediction target.
  • Compare with a simple baseline: assumptions, objective and independent validation.

Answer

Feature scaling (standardization/normalization) rescales features to comparable ranges. It matters for algorithms sensitive to feature magnitude/distance: KNN, K-Means, SVM, PCA, and gradient-descent-based models (logistic regression, neural nets) — unscaled features can dominate distance calculations or slow/destabilize convergence. It's generally not required for tree-based models (Decision Trees, Random Forest, Gradient Boosting) since splits are based on feature order, not magnitude.

Source and licence: Ishan Dutta — Awesome Data Science Interview Q&A · MIT

Related questions