What is data normalization and how is it different from standardization?
Updated May 17, 2026
Short answer
Normalization scales data to a fixed range, while standardization centers data around mean with unit variance.
Deep explanation
Normalization (Min-Max scaling) transforms data into a fixed range like [0,1], preserving distribution shape but scaling values. Standardization transforms data to have mean 0 and standard deviation 1, making it suitable for algorithms assuming Gaussian distributions. Standardization is more robust to outliers compared to normalization.
Real-world example
Scaling pixel values in image processing before training neural networks.
Common mistakes
- Applying normalization blindly without checking algorithm requirements.
Follow-up questions
- When should you prefer standardization?
- Does scaling affect tree-based models?