Synthetic Feature

Synthetic feature in machine learning refers to an attribute constructed from one or more existing features in the dataset. These are engineered features designed to provide additional information that the original features may not explicitly contain.

How Synthetic Features are Created

For instance, if a dataset contains features like 'height' and 'weight', a synthetic feature could be 'body mass index (BMI)', calculated from these two. Synthetic features are often used to enhance the performance of a model by incorporating domain knowledge or capturing interactions between existing features.