Reference Work Entry

Encyclopedia of Machine Learning

pp 577-580

Learning Curves in Machine Learning

  • Claudia Perlich

Synonyms

Error curve; Experience curve; Improvement curve; Training curve

Definition

A learning curve shows a measure of predictive performance on a given domain as a function of some measure of varying amounts of learning effort. The most common form of learning curves in the general field of machine learning shows predictive accuracy on the test examples as a function of the number of training examples as in Fig. 1.
https://static-content.springer.com/image/prt%3A978-0-387-30164-8%2F13/MediaObjects/978-0-387-30164-8_13_Part_Fig1-452_HTML.gif
Learning Curves in Machine Learning. Figure 1

Stylized learning curve showing the model accuracy on test examples as a function of the number of training examples

Motivation and Background

Learning curves were initially introduced in educational and behavioral/cognitive psychology. The first person to describe the learning curve was Hermann Ebbinghaus in 1885 (Wozniak, 1999). He found that the time required to memorize a nonsense syllable increased sharply as the number of syllables increased. Wright (1936) described the effect of lear ...

This is an excerpt from the content