Feature engineering is the process of using raw data to create features that will be used for predictive modeling. Using, transforming, and combining existing features to define new features are also considered to be feature engineering. Depending on the data and the problem you are working with, it could be a key activity in the predictive analytics process; it can make or break your model. There are many standard techniques for performing feature engineering. We will explain and use some of them in later sections; however, in most contexts this activity is based on common sense, intuition, and mainly domain knowledge, and it can be considered more of an art.
There is a debate over which transformations ...
