Elbow Method
noun
say it: “EL-boh METH-ud” /ˈɛlˌboʊ ˈmɛθəd/
A way to choose the number of groups for clustering by graphing how much each extra group helps and picking the point where the gains level off.
What is the elbow method?
The elbow method chooses k for k-means by plotting inertia for k = 1, 2, 3 and so on, and picking the k where the line stops dropping steeply: the bend that looks like an elbow. Past it, more clusters only split real groups. The bend is often soft, so it's a guide, checked against what the clusters mean.
In a sentence
- “Her plot had no clear bend, so the elbow method couldn't settle k on its own.”
- “The elbow method pointed to three groups of customers.”
Level: Journeyman · More AI Basics terms at this level
Updated