Training a classification model with ML.NET AutoML

AutoML is maintained in a separate NuGet package from the main ML.NET package, so we’ll need to install that with the #r magic command, as shown in this C# code cell:

#r "nuget:Microsoft.ML.AutoML,0.21.1"
using Microsoft.ML.AutoML;

With ML.NET AutoML installed and imported, we can set up our classification experiment. In this chapter, we’ll be classifying football players as likely to retire next season or not.

Because there are only two categories something could be in, this is what’s formally called a binary classification experiment, since each prediction is either going to be in one category or the other.

If there are three or more categories something might be categorized in, this is called a multi-class classification experiment. An example from our football dataset might be looking at a prospective player and predicting which position in your football club they’d perform best...