Vad betyder "Supervised"?

Same Data, Two Worlds

~15 min

Below is the same dataset. But watch what happens when you toggle labels on and off.

⚠️ You MUST try all 4 modes

This lesson won't make sense until you see what happens with wrong labelsand removed labels. Don't skip ahead!

🔍 Without Labels

Model sees patterns (gray circles show natural groupings) but has NO GOAL. It can cluster, but doesn't know what's "right" or "wrong".

Behavior: Groups similar things together, but can't predict labels.

Try all four modes to unlock the insight →

What You Just Saw

Labels don't just describe the data — they steer what the model learns.

  • With correct labels: Model learns the actual pattern
  • With wrong labels: Model learns the wrong pattern confidently
  • Without labels: Model has no direction at all

💡 The Deep Implication

This is why data labeling is so important. Whoever creates the labels determines what the model will learn. The "correct" answer is a human choice, not an objective truth.