Skip to content
BENIBIX Learn

Clustering

Also known as Cluster Analysis

Putting things that are alike into groups. Nobody names the groups first.

Think of it like

Think of it like kids at a party. Soon they stand in small groups of friends.

Example

Lucy's music app clusters 2 million songs by sound. One cluster turns out to be calm piano music, which becomes a new playlist.

Why it matters

It shows you groups you did not know were there, like new kinds of buyers or new kinds of faults.

Definition

Clustering is the task of putting similar examples into groups without being told the groups first. The model looks at how close examples are to each other. Examples that are alike end up in the same group, called a cluster. There are no labels, so it is unsupervised. People then look at each cluster and decide what it means.
Before, the dots are one cloud. After, they form three clusters, each with its own centre.

Common mistake

Mixing it up with classification. That sorts things into groups you named first. Clustering finds the groups itself.

Key takeaways

  1. 1
    No groups given. The model finds them.
  2. 2
    Alike sits together. Groups by closeness.
  3. 3
    People name them. You decide what each means.

Compare

Learn it properly