Controlling Risks Using Test-Driven Development
There are risks, such as selecting unreliable models, associated with creating models and technologies built on top of our models. The question is, could we avoid them and better manage the risks associated with machine learning modeling? In this chapter, we will talk about programming strategies such as unit testing, which could help us not only in developing and selecting better models but also in reducing risks associated with modeling.
In this chapter, we will cover the following topics:
- Test-driven development
- Machine learning differential testing
- Tracking machine learning experiments
By the end of this chapter, you will have learned how to reduce the risk of unreliable modeling and software development using unit and differential testing and how to reliably build upon previous attempts in model training and evaluation using machine learning experiment tracking.
