02 — Dimensionality Reduction

Methods for projecting data from a high-dimensional space into a lower-dimensional representation while preserving structure — either globally (linear methods) or locally (non-linear methods).

Notes

• Dimensionality Reduction — PCA, truncated SVD, t-SNE, UMAP, random projections, applications and trade-offs
• Dimensionality Reduction — Implementation — PCA (scree plot, whitening), TruncatedSVD (sparse), t-SNE, UMAP, SparseRandomProjection with scikit-learn

Links

← 01 — Clustering → 03 — Density Estimation

• SVD — mathematical foundation of PCA
• Unsupervised Learning