How Demand Engine Calculates the Neighborhood in Outlier Calculation
Hi Experts,
for Outlier identification in Demand Management, there are certain Parameters like DetectOutlier, OutlierSensitivity, OutliersPercent, RemoveExtremeOutlier.
This Question is about OutlierSensitivity.
As per the Whitepaper, this OutlierSensitivity Specifies the sensitivity of outlier detection. The parameter defines how far (srd error units) the specific observation is from its near neighbors. Where SRD stands for Standardized Residual Deviation.
Calculation Looks okay once the Average and Standard Deviation done & SRD calculated.
But the Question is How Engine Decides How many Neighbors to Consider ? is there any any parameter for this ? or is it a static value given to Engine ?
thanks.