By Jared Webb, BlueConduit
Now that the EPA has included statistical prediction as a service line investigation method, it’s worth considering what makes a “good” prediction. Like so many things, there are trade-offs and choices to be made. One choice is the prioritization of precision and recall. It may sound pedantic, but it matters for real-world outcomes.
In this article, explore the precision-recall tradeoff and the importance of understanding this tradeoff when using statistical tools.