Hi
I am looking at working with alarm thresholds.
Basically, I want to have an upper band and lower band which is derived from the values for last week.
Then if the values for this week breach these bands then I would look to do some trigger
This is my search and graph, just a basice search with 4 lines in the graph.
What, specifically, are you trying to detect?
So when the black line(shich is me drawing an example of the yellow line breaching the upper and lower bands) goes high and breaks the upper band at point x this would be a breech.
Likewise if the black line goes low and breaks the lower band at point y this would be a breech.
The values of this week are the yellow line, so as these realtime values come in, if they go above upper band or below lower band then these bands would be breached, therefore I want an alert.
I hope my comments help explaing the graph/picture better now.
I wanted to try and do it in machine learning as this has the outliers chart, which it seems is only available in the machine learning app. The outliers chart looks useful but it may not be of value, i am thinking at this early stage
note: I have asked a similar question here before, But it is slightly different here.
