AstraAI

Machine learning ยท clustering

Can the computer find hidden groups?

In the curriculum:Clustering (unsupervised)See the map โ†’
๐ŸŽ“ School maths ยท Years 9โ€“10Statistics โ€” Spot clusters on a scatter graph of two measurements.Maths by year โ†’

The Case. Here are the sizes and weights of some animals, all jumbled together with no labels at all. Can a computer sort them into sensible groups on its own? This is how streaming apps quietly group listeners with similar taste, or shops cluster their customers โ€” no answer key needed.

Your tools. scikit-learn's KMeans is unsupervised learning: it finds patterns with no answer key.

Investigate. Press Run. The colours are the groups the computer discovered; the red X's mark the centre of each group.

  • Change n_clusters to 3 โ€” does it invent a third group?
  • This is how apps group customers or songs without being told the categories.

This case is part of the Family plan

Unlock every 12โ€“15 case and the full AI, machine-learning, and deep-learning path โ€” one subscription for the whole family.

Need a hint?

Unsupervised = no labels. KMeans just looks for points that clump together.