Precision gauges how a model behaves by quantifying the percentage of instances correctly classified as a specific class, relative to all instances predicted as the same class. It is calculated as follows:

Recall is another name for sensitivity. The harmonic mean of precision and recall is called the F1 score and is calculated as follows:

The reason to use the harmonic mean instead of a simple average is that the harmonic mean is greatly affected by imbalances between the two values (precision and recall). Thus, if either precision or recall is significantly smaller than the other, the F1 score will reflect this imbalance.