A customer deals with a holiday rush of traffic during November each year, but does not want to be flooded with alerts when this happens. The increase in traffic is expected and consistent each…

Question

Options

• AOutlier Detection
• BStatic Threshold
• CCalendar Window
• DHistorical Anomaly

• A
  5% (3)
• B
  7% (4)
• C
  16% (9)
• D
  72% (41)

Explanation

historical anomaly is a detector condition that allows you to trigger an alert when a signal deviates from its historical pattern. Historical anomaly uses machine learning to learn the normal behavior of a signal based on its past data, and then compares the current value of the signal with the expected value based on the learned pattern. You can use historical anomaly to detect unusual changes in a signal that are not explained by seasonality, trends, or cycles. Historical anomaly is suitable for creating a detector for the customer's data, because it can account for the expected and consistent increase in traffic during November each year. Historical anomaly can learn that the traffic pattern has a seasonal component that peaks in November, and then adjust the expected value of the traffic accordingly. This way, historical anomaly can avoid triggering alerts when the traffic increases in November, as this is not an anomaly, but rather a normal variation. However, historical anomaly can still trigger alerts when the traffic deviates from the historical pattern in other ways, such as if it drops significantly or spikes unexpectedly.

No community discussion yet for this question.