Skip to content
BENIBIX Learn

Feature (Machine Learning)

Also known as Input Variable, Attribute, Predictor

One fact about a case that a model reads. In a table, it is one column.

Think of it like

Think of it like the clues in a guessing game. Is it big? Is it red?

Example

Emma's model guesses house prices. It uses three features: size in square metres, number of rooms and distance to the station.

Why it matters

A model can only use what the features tell it. A missing feature can make a good answer impossible.

Definition

A feature is one piece of information about an example that the model looks at. In a table, each feature is a column. For a house, the features might be its size, its number of rooms and its distance to town. The model learns how the features link to the answer. Good features make learning much easier.
The first three columns are features. They are what the model reads to make its guess.

Common mistake

Adding a feature that holds the answer in disguise. A column like final sale date gives away facts the model will not have later.

Key takeaways

  1. 1
    One fact. A single piece of information.
  2. 2
    A column. One per example, in a table.
  3. 3
    Feeds the guess. The model reads it to answer.

Learn it properly