Consequence-Weighted Resolution Error (CWRE)

GMI-EGD-005Evidence-Governed DataIntegration-RequiredPattern O-27: Risk-Weighted Allocation

Definition

The rate of false merges and false splits in entity resolution, weighted by the consequence of the claims each error affects, expressed against the unweighted rate.

The Measurement Gap

There is no metric that tells you the consequence-weighted rate of false merges and false splits in entity resolution, which means match quality is reported on a labeled sample while the merges that destroy evidence channels and the splits that halve corroboration carry the same weight as trivial ones.

Formula

CWRE = ( Σ_e w_e ) / ( Σ_j w_j ) , reported alongside E_unweighted = |errors| / |decisions|, with the amplification A = CWRE / E_unweighted

Proposed formula, pending validation pilots.

At a Glance

Unit
Ratio (0 to 1)
Trend Direction
Lower
Category
Evidence Acquisition and Admissibility

Decision Supported

Whether entity resolution thresholds should be tuned globally or per consequence band.

Context Boundary Assignment Accuracy (CBAA) →

Assess Your Organization

The AI-GPS Assessment evaluates your AI governance maturity and recommends KPIs like this one for your implementation roadmap.

Take the Free AI-GPS Assessment