Reportable Event Classification Accuracy (RECA-AGR)

GMI-AGR-055Agriculture and Farm OperationsPilot-ReadyPattern R-2: NLP Classification

Definition

The share of intake records (customer and neighbor complaints, worker exposure and injury reports, foreign-material findings, food safety deviations, pesticide drift reports) that met a reportability threshold under FSMA, the Worker Protection Standard, OSHA, or state pesticide rules and were classified as routine by the NLP or AI intake system, measured on an audited sample.

The Measurement Gap

There is no metric that tells you how often an AI-classified intake record (drift complaint, worker exposure, foreign material, food safety deviation) that met a reportability threshold was classified as routine, which means reportable events leave the system as closed complaints.

Formula

RECA-AGR = 1 − N_reportable_classified_routine / N_reportable_in_sample, on a stratified sample weighted toward rare categories; companion: escape count per quarter extrapolated from the sample

Proposed formula, pending validation pilots.

At a Glance

Unit
Ratio (0 to 1)
Trend Direction
Higher
Category
Packing and First-Buyer Interface

Decision Supported

Whether the intake system's routine queue contains reportable events, and which categories need a human reportability review regardless of the classifier's label.

← Recall Trace Scope Silence Rate (RTSSR)Risk-Weighted Pathogen Testing Coverage (RWPTC) →

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