The receiver operating characteristic (ROC) is a graphical analysis that helps understand the performance of a binary [[Classification|classifier]] using different thresholds for the predictions. It consists in plotting the true positive rate (TPR) against the false positive rate (FPR) at each of the threshold settings. Therefore, is possible to understand how changing the threshold affects the bias towards [[Precision|precision]] or [[Recall|recall]] in the [[Classification|classification]], as well as the discrimination power of the classifier.