SOURCE-LINKED INTELLIGENCE

Collapsibility of Performance Metrics in Clinical Predictive AI

arXiv · AI, language, vision and robotics · article · Aug 31, 2026 · UTC

Background: Population level assessments of predictive artificial intelligence (AI) can conceal performance disparities across subgroups. Fairness evaluations commonly rely on performance analyses across subgroups. However, some performance metrics are non-collapsible, meaning that the overall population performance value does not equal the weighted average of subgroup specific values. Objective: To examine the collapsibility properties of commonly reported performance metrics in predictive AI, with a focus on the area under the receiver operating characteristic curve (AUC, also known as c-sta

Read original source ↗ Open in workspace

recordType

paper

region

Global