Introducing our multi-class classification model

Previously, we covered binary classification models, which are named that way because they are predicting a categorical variable that’s going to be either one of two possible options.

A multi-class classification model is similar to a binary classification model except that there are three or more classes that something could fall into. For example, predicting which season something will occur in has four classes: spring, summer, fall, and winter.

In this chapter, we’ll be training a multi-class classification model that takes in information about a football club and its opponent in an upcoming match and predicts the winner of the game.

Since football matches can end in draws, this is a multi-class classification problem with 3 distinct classes: Win, Loss, or Draw.

We’ll start in our usual manner by importing ML.NET, AutoML, and the DataFrame with this hopefully familiar C# code block:

#r "nuget...