K-Means
Partitions data into K clusters by alternating assign-to-nearest and recompute-centroids.
- Lloyd iteration: assign points, then recompute centroids, repeat.
- Sensitive to feature scale and initialization (use k-means++).
- Assumes roughly spherical clusters of similar size.
When to use: Default first try for clustering; fast, scalable, and the baseline against which others are compared.
Example: Segment 200k customers on (recency, frequency, monetary) into k=5 clusters. K-Means converges in ~15 iterations and yields interpretable segments.