The Receiver Operating Characteristic (ROC) curve is a tool commonly used by data scientists to understand the performance of classifiers. It shows the true positive rate vs. the false positive rate for a range of propensity thresholds going from 0 to 1 (remember, your propensity score will range between 0 and 1). It is often used to compare models in an apples-to-apples way since it does not depend on your choice of propensity threshold, whereas many other error metrics do.
This article briefly discusses what an ROC curve is
Updated over 11 months ago