Overfitting

How it is used

You detect, avoid or reduce overfitting, for example with simpler models, more data, regularization or cross-validation. The verb is 'overfit': 'the model overfits'. The opposite problem is underfitting.

Example sentences

1. Training accuracy is 99% but test accuracy is 70%; that's classic overfitting.

2. With only 200 rows and 50 features, overfitting is a real risk.

Common collocations

• avoid overfitting
• reduce overfitting
• sign of overfitting
• prone to overfitting

Related terms

• training dataThe examples, with known inputs and usually known answers, that a model learns from; they are kept separate from the data used to test the model.
• test setA part of the data that is held back and never used in training, so that it can show at the end how well a model works on data it has not seen.
• cross-validationA way of testing a model by splitting the data into several parts, training on all but one, testing on the remaining part and repeating until each part has been used for testing.
• modelA mathematical representation of a real process, built from data, that is used to explain relationships or to make predictions about new cases.

This word in other fields

The same English word can have a different meaning in another field:

• overfitting · Software Engineering
• overfitting · Artificial Intelligence & Machine Learning

More terms: Modelling and Machine Learning

• feature
• regression
• classification
• accuracy
• precision
• recall
• baseline
• feature engineering
• target variable
• clustering