Subscribe to DSC Newsletter


 Presently we are working on building outlier detection model using SAS.As I don't have much experience I don't now how to go ahead in this. I have few ideas like cluster analysis...can u please suggest any method or procedure that we can adopt to build this model or any stuff that I can read..

Again today I read one interesting article based upon outlier detection

but I still unable to go ahead as I don't know how to Calculate distance of each point of a cluster from the centroid of the cluster

PS: By outlier I mean fraud & abuse that happen in health care industry

Thanks in Advance


Views: 3253

Reply to This

Replies to This Discussion

The definition of an outlier depends on the model you assume underlies your data:  Outliers under one model may not be outliers under another.


In SAS, a quick-and-dirty approach to outliers using cluster analysis is PROC FASTCLUS.  The documentation for this procedure includes an example to detect outliers.  Try several different options (especially the LEAST option of the PROC FASTCLUS statement) to determine whether this procedure works for your data.  Note that PROC FASTCLUS works only with interval/ratio/continuous valued variables, unless you stratify your data using separate analyses for nominal variables using the SAS BY statement.

There was a good discussion on Outlier topics on LinkedIn:


On Data Science Central

© 2021   TechTarget, Inc.   Powered by

Badges  |  Report an Issue  |  Privacy Policy  |  Terms of Service