Built with Mathigon

Glossary

Select one of the keywords on the left…

Precision and RecallNot the End

Reading time: ~5 min

Congratulations, you've survived the tradeoff! You are now equipped with the main tools needed to evaluate a classification model. However, the world of machine learning is massive, and precision, recall, and the F1-score are by no means the only tools in the shed.

There are many other popular diagnostic techniques out there, such as evaluating specificity, building expectation frameworks, or looking at likelihood ratios. Furthermore, while we've only talked about simple 'this-or-that' binary scenarios today, everything you just learned can be perfectly scaled up to handle complex, settings!

What's Next?

We've explored how altering the classification threshold changes an AI's priorities. In an upcoming chapter, we're going to introduce a powerful framework that evaluates a model's performance across every possible threshold simultaneously.

Keep an eye out for our deep dive into ROC and AUC!

Sina