aceNormalized risk-based evaluation of machine learning-based classification models: A multiclass approach with applications in medical devices
ObjectiveMachine learning is increasingly integrated into safety-critical domains such as medical applications. In this context, regulatory frameworks require the assessment and minimization of risks associated with incorrect model predictions. However, classical evaluation methods often focus on quantifying error frequencies, which do not reflect the heterogeneous impact of different types of errors. To address this deficiency, we elaborate an approach for assessing machine learning-based multiclass classificat…
Weighted balanced accuracy (WBAn) provides a regulatory-aligned, risk-based evaluation for multiclass medical AI that weights errors by severity and links development to real-world performance, demonstrated in X-ray lung disease detection.
Standard evaluation that counts error frequencies fails to reflect differing clinical severity of error types, leaving regulatory risk requirements incompletely implemented and real-world clinical impact inadequately assessed.
Demonstration is limited to a single use case, which constrains generalizability of the practical applicability claim.
Evidence
- Peer-reviewedJournal of International Medical Research2026-09-28
How should this claim be treated?
Truvace Impact Record TRV-2026-1215, v1: “Normalized risk-based evaluation of machine learning-based classification models: A multiclass approach with applications in medical devices.” Truvace, 2026-09-30. /record/TRV-2026-1215 (accessed at citation time). sha256 ada2a9ed1e1b136d…
Calibration history
Every change to this record since certification, in the open. None yet — the reading has held since it entered the record.
Certified into the record
How to verify without trusting this page
Fetch the canonical text of any version from /api/record/TRV-2026-1215 and hash it yourself — for example shasum -a 256 on the saved canonical field. The result must equal content_hash, and each version’s text ends with prev:followed by the prior version’s hash (version 1 chains to 64 zeros). If a single character of any version had been altered since certification, the chain would not reproduce.