Skip to main content
. 2016 Dec 7;11(4):791–799. doi: 10.1177/1932296816681584

Figure 3.

Figure 3.

(i) ROC curves of algorithms (A), (B), (C), and (E). (ii) The PS curves of the algorithms. We drew PS curves and evaluated whether the PPV maintained a high score while retaining a high sensitivity. The axes of the ROC curve consist of specificity = (1 – (TN / (FP + TN))) and sensitivity = (TP / (TP + FN)). The axes of the PS curve consist of precision = PPV and sensitivity, where TN = true negative, FP = false positive, TP = true positive, and FN = false negative.