Feature Engineering

Feature engineering is the process of creating or transforming features from raw data, as we mentioned in Chapter 3, Binary Classification. These features get fed into various machine learning models to generate our desired business outcomes. Feature engineering is one of the most important steps in the data science life cycle and is even more important than the models themselves. The veracity of the models depends on what goes into the models, which are the features you build for the dataset.

Building the best set of features is dependent largely on domain understanding and the intuitions derived from the data during the exploratory data analysis phase. There is a lot of creativity involved in creating and transforming features, and therefore feature engineering can be considered both as an art and a science. However, performing feature engineering manually is quite an arduous and time-consuming process. This is where automated feature engineering plays a significant...