Continuous features

Continuous features can have infinite possible values. Unlike nominal and ordinal features, which can only have a discrete set of values, continuous variables are numerical variables, and are not compatible with some ML algorithms. However, continuous features can be converted into ordinal features using a technique called discretization.

Although we will not discuss techniques to transform features from one form to another here, we will demonstrate how it can be done in our example sections. We have selected example datasets in this book where feature transformation is required. You should not only learn about these various transformation techniques from this book, but also observe how a data scientist analyzes a dataset and uses specific feature transformation techniques based on the application. We have also provided examples to apply these techniques at scale in Python and AWS SageMaker.