What is overfitting in machine learning?

An overfit model essentially memorizes its training examples rather than learning the general pattern behind them. It looks excellent on the data it was trained on, but fails to generalize when it encounters new, slightly different examples.

A simple analogy

Imagine a student who memorizes the exact answers to last year's practice exam questions instead of understanding the underlying concepts — they'll do great if this year's exam is identical, but poorly the moment the questions change even slightly.

Common ways to prevent it

• Using more training data so patterns generalize better
• Simplifying the model so it can't memorize excessive detail
• Regularization techniques that penalize overly complex models
• Testing on a separate validation set the model never trained on, to catch overfitting early