Quantile Of Numeric

Find The: Quantile Of Numeric

quantile_of_numeric

Returns the quantile, in the range 0.0 to 1.0, of a numeric attribute on this event compared to the full population of past events within the configured Time Window.

Whereas Statistic Of Numeric detects anomalies in the behaviour of one user or device, Quantile Of Numeric detects them against the whole population.

For example:

Comparing the amount of a purchase or payment against past events to see if it is unusually large or small, without defining arbitrary thresholds that need constant monitoring.
Detecting outlier devices by comparing attributes such as profiling.tcp_connection['*'].rtt or profiling.tcp_connection['*'].mss.
Detecting outlier behaviour by comparing journey or biometric attributes such as time since last step.

Configuration

Feature Name: The "dictionary" name under which the feature value will be stored and referenced by in rules.
Numeric Attribute: The attribute to calculate the quantile of. It must be a continuous attribute with a Real or Integer data type; a discrete attribute is rejected at build time.

Optional Configuration

Default Value: A value that can be assigned to the feature if the attributes to calculate the features are not present.
Scope: Can be used to extend the event search from the local node, to all nodes within the same organization all the way to all across customers (global).
Time Window: Can be used to limit the Feature calculation to a specific time period/window. e.g 1 week. Default is all ; no time window

Use Case 1

Compare the amount of the current purchase against all purchases seen in the last 7 days.

Using the Feature in Rules and Investigations

Find purchases in the top 1% by amount for the last seven days.

feature('purchase_amt_quantile') >= 0.99

Implementing In The Feature Editor

  - name: purchase_amt_quantile
    default_value: 0
    time_window:
      days: 7
    find_the:
      quantile_of_numeric:
        attribute: purchase.amount
    with_scope: same_node_instance