Dimensionality reduction

Principal component analysis

PCA as variance maximization and as a reconstruction problem; the SVD; choosing the number of components; PCR as a regression follow-up.

📖 Assigned reading: ISLR §6.3

---

📊 Slides

• 🖥️ Clean slides
• ✏️ Annotated slides (in-class version)

🧪 Lab

Live, in-browser R lab — runs entirely in your browser via webR, no local R install needed.

• 🧪 Lab: Breast tumors (live R) — edit and re-run any code cell: PCA by hand on the 30 Wisconsin tumor features (scaling, eigendecomposition, scree plot, loadings, a 2D map), then PCR with the number of components chosen by cross-validation. A “Download my code” button captures whatever you’ve written before you leave.