Skip to content
BENIBIX Learn

Overfitting

Also known as Overfit, Overtraining

The model learns its old data by heart. Then it fails on new data.

Think of it like

Think of it like a kid who learns the test answers by heart. The kid has not learned the subject.

Example

Jack's model scores 99 percent on its training photos but only 70 percent on new ones. The gap shows it has overfit.

Why it matters

It is the most common way a model fails in real use. A great training score means little until new data agrees.

Definition

Overfitting is when a model learns its training data too closely, noise and all. It scores very well on the data it trained on. But on new data, it does much worse. It has memorised the examples, not the real pattern. More data, a simpler model or stopping training earlier can all help.
Overfit is the right panel. The line bends to touch every dot, noise included.

Common mistake

Judging a model by its training score. Always check it on a validation set it has never seen.

Key takeaways

  1. 1
    Too close a fit. Learns the noise, not just the pattern.
  2. 2
    Great on old data. The training score is high.
  3. 3
    Bad on new data. The real test fails.

Compare

Learn it properly