The ROC: A Mystery Unveiled

The image displays two machine learning performance graphs side-by-side: a precision-recall curve on the left showing a 'Perfect Classifier' as a horizontal line at precision 1, a 'Pretty Good' classifier as a descending step line, and a 'Baseline Classifier' as a dashed horizontal line at precision 0.5. On the right is an ROC curve with a 'Perfect classifier' at the top-left corner (1,1), a 'Better' curve above the diagonal, a 'Worse' curve below the diagonal, and a 'Random classifier' along the diagonal. The text 'Baseline – predict all instances as +ve class' is at the bottom left.

Report this content