Fraud ML System
Live drift surveillance, model registry status, and seeded prediction telemetry from the production backend.
Stable
Observed Fraud Ratepolicy
0%
0 recent predictions sampled
Registry Versionslayers
0
No versions returned
Latest Confidencetarget
n/a
Waiting for prediction data
Last Syncsync
pending
Refreshing now...
Prediction Confidence Trend
Latest production and shadow confidence values
No data
Prediction Mix
Fraud vs normal decisions
0% fraud
No data
Normal
0
Fraud
0
System Health
Real-Time Drift
0.000
Threshold: 0.1
Low Drift
Retrain Pipeline
Last Statusidle
TriggerNone
New ModelN/A
Last RunN/A
Model Registry
Version control and promotion workflow
Active Production Model
No active production model found
Available Candidate Models
No candidate models available.
Drift Score Over Time
Evolution of data distribution divergence
No data
Confidence Distribution
Density of production model probability scores
No data