Increase the number of splitting data on machine learning


(Angelos) #1


Is there any way to increase the result of splitting data in machine learning?
Now, when you split the data in multi metric job, the result has a limit of 50 values(5, 10, 25, 50).
Can we change that?

Thank you in advance!


(rich collier) #2


This view isn't showing the number of splits for analysis (because in theory there could be hundreds or thousands) - rather this display it is showing the top N (5, 10, 25, or 50) most unusual entities in the timeframe selected.

If you truly want to see anomalies for ALL entities within a timeframe, this information is available via queries to the results index (.ml-anomalies-*) or via the API. The following blog gives some explanation and examples:

(Mark Walkom) #3