Concepts

This section describes concepts that underpin the practical activities we will cover in this chapter.

Model evaluation metrics used in this chapter

We’ve already discussed the topic of model evaluation metrics in previous chapters. We first introduced the concept in Chapter 1, where we briefly discussed metrics such as the mean squared error (MSE) for regression use cases and accuracy for classification use cases. In Chapter 5, we used functions in scikit-learn to calculate some of these metrics for the models we created, and we suggested looking up additional metrics as a supplemental learning activity at the end of that chapter.

In this chapter, we will train models for a classification use case, and we will introduce some additional metrics to evaluate our models. The main metric we will use is something called AUC ROC, which stands for area under the receiver operating characteristic curve. That sounds like a lot, but don’t worry, we will explain...