Compare the sensitivity and specificity of different SMOR-based metrics for trio sharing (K=3) using the ROC curve

False-positive rate (FPR) = FP / N, where N = TN + FP; FPR = 1 – specificity True-positive rate (TPR) = TP / P, where P = TP + FN; TPR = sensitivity

As long as you can get a set of FPR and TPR values (by sweeping through the range of your statistic picking continuously changing values as corresponding thresholds), you would be able to do everything introduced in this post.

References

• Calculating AUC: the area under a ROC Curve
• How to determine best cutoff point and its confidence interval using ROC curve in R?