Standardization vs normalization — what's the difference?
easyAnswer
- Standardization (Z-score): (x - μ) / σ → mean 0, SD 1.
- Preserves shape, robust when features have similar scales after transform.
- Normalization / MinMax: (x - min) / (max - min) → bounded [0, 1].
- Sensitive to outliers (a single extreme value squashes everything).
- Use standardization for linear models, PCA, k-means, neural networks.
- Use normalization when a bounded range matters (image pixel scaling to [0,1]).
- Robust scaling: (x - median) / IQR — outlier-robust alternative.
Check yourself — multiple choice
- Same operation
- Standardize: (x-μ)/σ → mean 0 / SD 1 (shape-preserving); Normalize: MinMax → [0,1] (outlier-sensitive); Robust: (x-median)/IQR
- Only for classification
- Never used
Standardize: Z-score; Normalize: MinMax; Robust scaling with median/IQR.
#descriptive#eda
Practise Statistics Fundamentals
215 interview questions in this topic.