Programming Machine Learning

What You Just Learned

We went through a lot of information in these first two chapters, and I introduced many new terms. Let’s recap.

In this chapter we wrote our first supervised learning program. A supervised learning system learns from examples composed of input variables and labels. In our case, the input variables were the number of reservations, and the labels were the number of pizzas.

Supervised learning works by approximating the examples with a function, also called the model. In our first program, the model is a line identified by two parameters—the weight and the bias. This idea of approximating the examples with a line is called linear regression.

The first phase of supervised learning is the training phase, when the system tweaks ...