auc_mtl: Area Under the ROC Curve for Multi-Task Predictions

In orthoMTL: Multi-Task Learning with Orthogonal Constraints

Area Under the ROC Curve for Multi-Task Predictions

Description

Pooled AUC over every non-NA cell, computed via the Mann-Whitney U statistic (with 0.5 credit for ties), for orthoMTL fits in classification mode. The positive class is defined by true label > 0. AUC is invariant to monotone transforms, so raw "link" scores and "response" probabilities give identical results.

Usage

auc_mtl(true.label.mat, pred.label.mat)

Arguments

true.label.mat	A numeric matrix of true responses (n x numTasks). NA cells are ignored.
pred.label.mat	A numeric matrix of predictions of the same dimensions (as produced by predict.orthoMTL).

Value

A single numeric value in [0, 1] (higher is better; 0.5 = random).

See Also

accuracy_mtl, cindex_mtl

Examples

set.seed(1)
Y <- matrix(sample(c(-1, 1), 30, replace = TRUE), 10, 3)
P <- Y + matrix(rnorm(30), 10, 3)
auc_mtl(Y, P)

orthoMTL documentation built on Aug. 23, 2026, 5:10 p.m.