Skip to content
ML to LLMs
Learning path

Module 8

Dimensionality reduction

Compress many features into a few informative directions with PCA, and use t-SNE and UMAP carefully.

  1. 1. PCA, t-SNE, and UMAPCompute principal components and explained variance for a small dataset, project onto them, and read nonlinear embedding plots without fooling yourself.35 min

High-dimensional data is hard to visualize, slow to process, and suffers from the curse of dimensionality. PCA finds the directions of greatest variance and projects onto them. Nonlinear methods (t-SNE, UMAP) make striking 2D maps that are easy to over-interpret.

#LessonYou will be able to
1PCA, t-SNE, and UMAPcompute principal components and explained variance for a small dataset, and read embedding plots critically
Next moduleModule 9: Classical NLP