概述

Feature engineering is the process of selecting from a dataset the relevant variables in their optimal form to be used as inputs in a predictive model.