Part 2. Building a reactive machine learning system

This is the heart of the book. This part will build up your knowledge of the components of a machine learning system, starting from raw data in the wild and looping all the way back around to acting on the real world.

Chapter 3 is about collecting data. It’s not a normal chapter for a machine learning book: instead of hand-waving away where data comes from, we’ll take a serious look at a range of data issues, concentrating on when that data is big, fast, and hairy.

Chapter 4 explores deriving useful representations of data, called features. This is one of the most important skills a machine learning systems developer can have and is often the largest part of the work.

Once you get to chapter 5, you should be ready to do what everyone focuses on in machine learning: learn some models. There are entire books about learning models, but this chapter represents a unique view and will give you an understanding of how this step connects to what came before and what comes after. I’ll introduce you to some useful techniques to employ when the pieces of your system aren’t as easy to join up as you’d like.

Chapter 6 covers the rich topic of how to make decisions about machine learning models that you’ve produced. Not all models are created equal. There’s a range of common errors that you can make in learning about models, so I’ll attempt to arm you with tools you can use to figure out the differences between a good model and a bad one.