juniorPCA
What is Principal Component Analysis (PCA)?
Updated May 17, 2026
Short answer
PCA is a dimensionality reduction technique that transforms correlated features into a smaller set of uncorrelated components.
Deep explanation
PCA works by identifying directions (principal components) in feature space where data variance is maximized. It re-expresses data in a new coordinate system where axes are orthogonal and ranked by variance contribution. This reduces dimensionality while preserving most information.
Real-world example
Reducing 1000-dimensional text embeddings into 100 dimensions for faster search.
Common mistakes
- Assuming PCA selects important features instead of creating new ones.
Follow-up questions
- Is PCA supervised or unsupervised?
- Does PCA preserve labels?