A true negative occurs when a data point with a negative label is correctly classified as negative. In a health test, a healthy person diagnosed as healthy is a true negative. section 7.5
Spam filtering
If spam is the positive class, a ham email correctly classified as ham is a true negative. On a classification diagram, it appears as a negative-labeled point in the negative zone. section 7.5
Keep the label and prediction separate:
- Negative label plus negative prediction equals true negative. section 7.5
- Negative label plus positive prediction equals false positive. section 7.5
Figure 7.1 Examples of two models that are widely used in real life and that we’ll use throughout this chapter. On the left, a coronavirus model where the people are diagnosed as healthy or sick. On the right, a spam-detection model where the emails are classified as spam or ham. For each model, we have highlighted some of their errors and separated them as false positives and false negatives.
Compare healthy-versus-sick diagnoses and spam-versus-ham predictions, including the locations of true negatives and other outcomes. section 7.5

