Learning Curve

noun

say it: “LUR-ning KURV” /ˈlɝnɪŋ ˈkɝv/

A graph of a model's scores as it is trained on more and more examples, used to see whether more data would help.

What is a learning curve in machine learning?

A learning curve plots a model's training and validation scores as it's trained on more and more examples. If the two are still far apart, more data would likely help; if they've met at a poor score, the model is too simple and more data won't fix it. scikit-learn's learning_curve() works them out.

In a sentence

  • “A wide gap on the learning curve suggested the model needed more examples.”
  • “The learning curve flattened after 500 rows, so collecting more wouldn't help.”

Level: Freelancer · More AI Basics terms at this level

Updated