Feedback:

anomalydetection

Use the anomalydetection operator to find outliers in values of numeric fields.

Syntax

|anomalydetection method="iqr" [anomalydet_action={"transform"|"remove"}] [<fields>]

method="iqr"

Required. Specify the detection method. The current available option: "iqr".
anomalydet_action={"transform"|"remove"}

Optional. Specify the action to be taken with the detected outliers. Possible options:

  • "transform": Transforms the fields based on the outliers.

  • "remove": Removes the outlier values.

If you specify the anomalydet_action attribute, the anomalydetection operator will work the same as the outlier operator.
<fields>

Optional. Specify a comma-separated or space-separated list of numeric fields to check for outliers.

Usage

You can use the anomalydetection operator to find anomalies in numeric fields and, if needed, modify them. Currently, this operator supports only the IQR method.

Examples

Example 1: Detect outliers using the interquartile range method.

Copy 
|anomalydetection method="iqr"

Example 2: Transform detected outlier values to boundaries.

Copy 
|anomalydetection method="iqr" anomalydet_action="transform"

Example 3: Remove outlier values from the results.

Copy 
|anomalydetection method="iqr" anomalydet_action