Ability to define dates categories for more accurate anomalies detection

[Kostantinos Zafiropoulos]

We have a periodical operation during a month, when the first/last working day should be compared with other first/last working dates of other months/years. In addition there are some other dates with extra workload (Black Friday).

We would like to pre-define these days with expected high volumes on watch Tower with specific flags (e.g first or last working date of month) so the detection anomaly algorithm can produce more accurate results.

Kind regards

[Ian Kolker]

Great feedback and indeed this feature is on our roadmap. It is currently known and Special days. Our plan is to introduce dedicated ML models for specific areas. Please be patient on this as we need to deliver foundation abilities before this may be implemented.