Rare Category Miss Score (RCMS)

GMI-MAS-013Maximo Application SuiteIntegration-RequiredPattern R-2: NLP Classification

Definition

The false-negative rate of the incident classifier on incident type and category combinations that are rare in the training population but consequential in operation.

The Measurement Gap

There is no metric that tells you the false-negative rate of the incident classifier on incident types that are rare but consequential, which means rare categories are absorbed into frequent ones and disappear from trend analysis.

Formula

RCMS = FN_rare / (FN_rare + TP_rare)

Proposed formula, pending validation pilots.

At a Glance

Unit
Ratio (0 to 1)
Trend Direction
Lower
Category
Classification and Extraction Fidelity

Decision Supported

Whether the classifier can be trusted on the incident types that matter most, and whether the training population needs deliberate enrichment rather than more volume.

← Inference Filter Coverage (IFC)Suggestion Acceptance Quality (SAQ) →

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