|Date: ||Thu, 20 Sep 2007 08:11:19 -0700|
|Sender: ||"SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>|
|Subject: ||Re: How to find anomolies in one's data|
|Content-Type: ||text/plain; charset="us-ascii"|
On Sep 19, 3:11 pm, art...@NETSCAPE.NET (Arthur Tabachneck) wrote:
> This may or may not be the place to ask this question, but I am under the
> impression that SAS has some components that may be able to provide the
> desired functionality.
> Given a situation where one gets daily transactional information (e.g.,
> vehicle information number (VIN), premium, # of monthly exposures, type of
> claim and loss incurred), from about 200 companies (providing
> approximately a total of 200 million records each year), can SAS be used
> to identify such deviations as changes in patterns of amounts reported for
> any or all of the above?
> Of course I realize that one could simply code lookup tables of previously
> reported amounts and test for differences, but I am under the impression
> that more advanced methods already exist.
> Do any of you use SAS for such a purpose and, if so, can you provide some
> Much appreciated as always,
Take a look at this . . . I've implemented this method and it works
extremely well. If this doesn't work for I'm certain it will
definitely get you thinking. Hope this helps.
'Detecting Anomalies in Your Data Using Benford's Law'