Anomaly Detection
Automated detection of unusual trade patterns using statistical analysis
Total Anomalies
High Severity
Medium Severity
Low Severity
Detected Anomalies
0 anomalies
Loading anomalies...
Recent High-Priority Alerts
How Detection Works
1
We calculate a 12-month rolling average and standard deviation for each HS code.
2
Current values are compared against the expected range using Z-score analysis.
3
Anomalies are flagged based on deviation thresholds: |Z| > 2.5 (High), |Z| > 2.0 (Medium), |Z| > 1.5 (Low).
Severity Levels
high
Z-score > 2.5 or < -2.5
medium
Z-score > 2.0 or < -2.0
low
Z-score > 1.5 or < -1.5
Anomaly Trends
Historical pattern of detected anomalies
| Period | High | Medium | Low | Total | Trend |
|---|---|---|---|---|---|