approximate

thumbnail
(0)

Version 1.1.1  by WSO2

Category : Siddhi Execution
Supported Product Version : SP 4.0.0 SP 4.1.0 SP 4.2.0 SP 4.3.0 SP 4.4.0  

Summary

The siddhi-execution-approximate is an extension to Siddhi that performs approximate computing on event streams.

Features of siddhi-execution-approximate extension.


  • Count

    This extension applies the Count-min-sketch algorithm to a Siddhi window.
    The algorithm is set based on a specific relative error and a confidence value to calculate the approximate count(frequency) of events.
    Note that using this extension without a window may cause out of memory.

  • distinctCount

    This applies the HyperLogLog algorithm to a Siddhi window.
    The algorithm is set with a relative error and a confidence value to calculate the number of distinct events with an accepted level of accuracy.
    Note that if this extension is used without a window, it may cause out of memory errors.
    If you need to perform these calculations without windows, use the approximate:distinctCountEver extension.

  • distinctCountEver

    This extension applies the HyperLogLog algorithm set on a specific relative error and a confidence value to a Siddhi window in order toon a streaming data set based on a specific relative error calculate the number of distinct events.
    Note that this extension returns erroneous values if is is used with a Siddhi window.
    If you want to perform these calculations with a window, you need to use the approximate:distinctCountEver extension.

Social Sites