# What is a feature?

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

In typical scenarios, the dataset ($\dataset$) consists of pairs $(\vec{\iinputvec{\idataset}}, \ioutputval{\idataset})$ where $\vec{\iinputvec{\idataset}}$ is a $\inputdim$-dimensional vector of inputs and $\ioutputval{\idataset}$ is called the output value.

The component $\sx_{\si\sj}$ is-called the $\sj$-th feature. The $\sj$-th feature vector is the vector:

Other names for the features are:

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