ROC Curve and AUC: the Mass Above the Diagonal
An interactive live film: negative and positive score bells stand on two walls, a threshold cuts their tails into FPR and TPR, and the bell of score pairs shows AUC = P(S+ > S-) as the mass above the diagonal. Find the best threshold.
Narration
A classifier gives every example a score. Here negative examples score like a standard normal, on the back wall in red, and positive examples score like a normal shifted by d, on the left wall in blue.
Pick a threshold tau and call everything above it positive. The blue tail beyond tau is the true positive rate; the red tail beyond tau is the false positive rate.
The rest of the narration is in the book's Companion.
More from Chapter 13
Interactive1:35Normalizing Flows: Chaining Invertible MapsChapter 13
Preview1:38Independence as an Outer ProductChapter 13
Preview1:40Conditional Gaussian: A Slice of the BellChapter 13
Preview1:22Random Fourier FeaturesChapter 13
Preview1:46Change of Variables: Same Probability, Different AreaChapter 13
Preview1:35The von Mises-Fisher Distribution: A Gaussian for DirectionsChapter 13
Preview1:42The Dirichlet Distribution: A Density on the SimplexChapter 13
Preview1:14Gaussian Mixture Model: Multimodal TerrainChapter 13
Preview1:41Whitening: PCA and ZCAChapter 13
Preview1:36Beta and Bernoulli: Updating with Pseudo-CountsChapter 13
Preview1:40Hoeffding's Inequality: How Fast a Sample Mean SettlesChapter 13
Preview1:25Temperature Scaling: Calibrating a SoftmaxChapter 13
Preview1:34Mahalanobis Distance: The Ellipsoids of a GaussianChapter 13
Preview2:09Random Directions Are Nearly OrthogonalChapter 13