AI Model Performance Metrics

Outbreak Anomaly Detector v1.2 - Pilot Validation Data (2025)

← Back to Home

Sensitivity
84%
True positive detection rate
Specificity
91%
True negative accuracy
AUROC
0.89
Area under ROC curve
PPV
76%
Positive predictive value

Equity & Fairness Analysis

Performance across demographic subgroups (no statistically significant disparities detected at α=0.05)

Urban vs Rural p = 0.18 ✓
High vs Low SVI p = 0.23 ✓
Age Groups p = 0.31 ✓
Calibration Equity Δ = 0.03 ✓

Validation Context: Metrics based on pilot deployment data and literature review. Performance numbers represent projected outcomes from early validation studies and will be updated as additional peer-reviewed evidence becomes available. External validation pending Q1 2026. [16]