Developer Guide for Intel® Data Analytics Acceleration Library 2019 Update 5
Outlier detection methods aim to identify observation points that are abnormally distant from other observation points. A univariate outlier is an occurrence of an abnormal value within a single observation point.
For more information on the concepts behind the algorithm, see "Details" section.
For more information on the algorithm's parameters for a specific computation mode and examples of its usage, see "Batch Processing", "Online Processing" and "Distributed Processing" sections.