Hands-On Artificial Intelligence for Cybersecurity

AUC (Area Under the ROC curve)

The ROC curve allows us to evaluate the performance of a classifier by plotting TPR against FPR (where each point of the curve corresponds to a different classification threshold).

We can also compare different classifiers to find out which one is more accurate, using the area under the ROC curve.

To understand the logic of this comparison, we must consider that the optimal classifier within the ROC space is identified by the coordinates of points x = 0 and y = 1 (which correspond to the limit case of no false negatives and no false positives).

To compare different classifiers, we can calculate the value of the Area Under the ROC Curve (AUC) associated with each classifier; the classifier that obtains the highest ...