· Introduction of EER(Equal Error Rate)

Deep learning articles generally use the Error probability such as EER(Equal Error Rate) as an objective standard to measure the classifier, and the ROC curve in the blog explains how to calculate EER.

Below is a brief introduction to EER calculation

EER (Average Error probability) is a biometric security system algorithm used to pre-determine its error acceptance rate and its error rejection rate threshold. When rates are equal, the common value is called the equal error rate. This value indicates that the proportion of false acceptances equals the proportion of false rejections. The lower the value of equal error rate, the higher the accuracy of biometric identification system.

Other evaluation criteria using ROC

• Area under thecurve (AUC) is also the area under the ROC curve. The larger the area is, the better the classifier is. The maximum value is 1
• EER (equal error rate) is the value of FPR=FNR. Since FNR= 1-tpr, it is possible to draw A straight line from (0,1) to (1,0) and find the intersection point, A and B in the figure.