What is a feature?

Features are individual measurable characteristics of an object. It might be a numeric, string, or any other value. Usually, data is presented in the form of many features, each of which describes a particular property of an object. 

In machine learning, features are variables used by the model to draw insights from the data. Therefore, selecting the right features is crucial for model training. For the example of the stocks, features might include dates, prices, and companies’ names, while in tv viewing example, channels, times, duration of watching, etc., would come as features.

Synonyms: variable, explanatory variable.

Why is it important?

A feature is a set of data describing a particular part or property of an object. Data is usually presented in features, which makes them an integral part of a machine learning pipeline.