5.2.9 Anomaly Detection
This topic describes the information about the Anomaly Detection.
Anomaly detection identifies data instances that deviate from the normal data pattern. Anomalies are unusual patterns not necessarily fraud.
For example: Suppose you are monitoring financial transactions with the columns: TRANSACTION_ID, AMOUNT, LOCATION, TIME.
Table 5-8 Example – Transaction Details
| TRANSACTION_ID | AMOUNT | LOCATION | TIME |
|---|---|---|---|
| TXN001 | 100 | NY | 14:00 |
| TXN002 | 150 | NY | 14:05 |
| TXN003 | 105 | NY | 14:12 |
| TXN004 | 9000 | CA | 03:00 |
Anomaly detection models may flag TXN004 as an anomaly due to the unusually high amount and odd time/location compared to typical transaction patterns.
Parent topic: Frameworks Supported