Lesson 22 — k-means step (expert)

k-means: find groups without labels

Clustering finds groups in data that has no labels. k-means keeps k centres and repeats two steps: (1) assign every point to its nearest centre; (2) update every centre to the mean of the points assigned to it. Repeat until the centres stop moving.

groups = {0: [1, 2], 1: [9, 10]}
new_centres = [sum(g) / len(g) for g in groups.values()]   # [1.5, 9.5]

Why it matters: clustering is used for customer segments, image compression and spotting unusual data.