Algorithms
Predictive Insights anomaly detection algorithms use numerous statistical and analytics techniques to detect anomalies when Predictive Insights is sure that anomalous behavior is occurring.
Anomaly detection algorithms
| Name | Description |
|---|---|
| Robust Bounds | Detects an anomaly when the value of a metric deviates from the metric's baseline. The baseline is the normal operating range the algorithm dynamically maintains for a metric. |
| Flatline | Detects an anomaly when the value of a metric that is normally variable becomes flat. |
| Variant/Invariant | Detects an anomaly when the variance between a metric's high and low values reduces significantly. |
| Granger | Finds causal relationships between metrics and detects an anomaly if that causal relationships changes. |
| Finite Domain | Detects an anomaly when a metric value elevates to a level not reached previously. |
| Predominant Range | Detects an anomaly when the variation in a metric value exceeds the range within which the metric normally varies. |
| Discrete Values | For a metric that normally has a defined number of possible values, detects an anomaly when the probability of the current value is rare. |