Fig. 10From: A novel estimator for the two-way partial AUCReal data illustrations. Panel a compares the ROC curves generated from the predictions of a logistic regression and a random forest classifier on the Diabetic Retinopathy Debrecen dataset, while panel b presents the analogous comparison on the Sepsis Survival dataset. In both panels the red square captures the area of interest of the ROC space obtained by setting the sensitivity and specificity thresholds to \(b_{se} = 0.4\) and \(b_{sp} = 0.4\), respectivelyBack to article page