(a) ROC curves for two hypothetical classification systems. The red curve for system A is better than the blue curve for system B. We plot the true positive rate (TPR) vs the false positive rate (FPR) as we vary the threshold $\tau$. We also indicate the equal error rate (EER) with the red and blue dots, and the area under the curve (AUC) for classifier B by the shaded area. Generated by roc_plot.ipynb . (b) A precision-recall curve for two hypothetical classification systems. The red curve for system A is better than the blue curve for system B. Generated by pr_plot.ipynb .