Eval Metrics

all demo data

Powered by Arize Phoenix · Last updated: just now

Cases processed

20

Urgency accuracy

90%

Guideline accuracy

88%

Avg confidence

85%

Routed to human

10%

Patients reached

85%

Confusion matrix — Predicted vs. actual urgency

Diagonal = correct predictions

Predicted →UrgentShort intervalRoutineNo follow-up
Urgent5000
Short interval0610
Routine0050
No follow-up0003

Per-case breakdown

CasePredictedActualConfidenceResult
RR-001UrgentUrgent
91%
Correct
RR-002RoutineRoutine
87%
Correct
RR-003UrgentUrgent
78%
Correct
RR-004No follow-upNo follow-up
93%
Correct
RR-005Short intervalShort interval
68%
Correct