Microsoft ML DotNet SrCnn Anomaly Detection Method Signature Parameters
Detect Anomaly By SrCnn
I have a time series and i want to detect anomaly in that time series. But i don't have any knowledge on how to give the method parameters their values or as to how to tune these parameters in a way that i get correct output. I could leave them to their default value but that doesn't solve my problem, i need to know as if i change these parameters from their default value what would it affect and how do they actually behave in detecting the anomaly.
the parameters i want to have understanding of are:
-windowSize
-backAddWindowSize
-lookaheadWindowSize
-averagingWindowSize
-judgementWindowSize
-threshold
Also, i want to understand the algorithm working. I tried reading the research paper but i didn't understand it. If anybody can explain it to me in simpler terms.