Categorical Feature

noun

say it: “kat-uh-GOR-ih-kul FEE-cher” /ˌkætəˈɡɔrɪkəl ˈfitʃər/

An input to a model whose values are names of groups, such as colors or cities, rather than measured amounts.

What is a categorical feature?

A categorical feature is a feature whose values are categories rather than amounts, such as a pet's species or a home's neighborhood. Most models only take numbers, so it has to be encoded first, usually with one-hot encoding; numbering the categories would invent an order they don't have.

In a sentence

  • “The model crashed on the neighborhood column, a categorical feature still holding text.”
  • “Species is a categorical feature: a cat isn't more or less than a dog.”

Level: Apprentice · More AI Basics terms at this level

Updated