Overfitting
When a machine learning model learns training data too well, including noise, reducing its generalization ability.
Definition
Overfitting occurs when a machine learning model learns the training data too well — including noise and random fluctuations — resulting in excellent training accuracy but poor performance on new, unseen data. An overfit model fails to generalize. It typically happens when a model is too complex relative to the amount of training data. Techniques to combat overfitting include regularization (L1/L2), dropout, early stopping, cross-validation, data augmentation, and using simpler models.
Example
“A model trained on 100 medical images achieves 100% training accuracy but only 60% on new images — it memorized training examples rather than learning general patterns.”
Synonyms
- memorization
- over-training
- model over-specialization
Antonyms / Opposites
- underfitting
- generalization
- regularization
Images
CC-licensed · free to useVideo
Related Terms
- underfitting
- regularization
- cross-validation
- training-data