What is a feature?

In machine-learning, a feature is an input variable to a supervised learning problem.

In typical scenarios, the dataset () consists of pairs where is a -dimensional vector of inputs and is called the output value.

The component is-called the -th feature. The -th feature vector is the vector:

Other names for the features are:

  • covariates,
  • independent variables,
  • explanatory variables,
  • exogenous variables,
  • predictors,
  • regressors