Programming the statistical procedures from SAS

Proc severity that results only with Kernel

Accepted Solution Solved
Reply
New Contributor
Posts: 3
Accepted Solution

Proc severity that results only with Kernel

Hi everyone,

I am currently working on a project with the SAS software.

The point is to model the operational risk of a financial institution.

I am using the proc severity function to estimate the distribution of my loss data. If I rely on the many examples on the Internet, the proc severity function, used with DIST=_predefined_, should result with statistical study of each of the distribution and which one is the best. However, when I am using the function, it seems like it is only checking for the Kernel distribution while I would like it to check for every predefined distribution.

I attach you the program, log and results for you to see.

Program.PNGlog.PNGresults.PNG

I simplified the program for it to be as simple as possible on my issue.

As you can see, the histogram reveals only the kernel distribution but I would like it to compare all the predefined distributions.

Can anyone help me?

I thank you in advance for your help.


Accepted Solutions
Solution
‎05-09-2015 12:55 PM
SAS Employee
Posts: 5

Re: Proc severity that results only with Kernel

Please remove the '=' sign after 'dist' and try again.

View solution in original post


All Replies
New Contributor
Posts: 3

Re: Proc severity that results only with Kernel

Furthermore, when I use the proc severity with the option OUTEST, the outest comes up empty...

Maybe is it because of the way my input database is formatted (I put only the variable name with ~15 observations below it for this simplified example).

Thank you in advance !

Solution
‎05-09-2015 12:55 PM
SAS Employee
Posts: 5

Re: Proc severity that results only with Kernel

Please remove the '=' sign after 'dist' and try again.

New Contributor
Posts: 3

Re: Proc severity that results only with Kernel

Thank you a lot !

☑ This topic is SOLVED.

Need further help from the community? Please ask a new question.

Discussion stats
  • 3 replies
  • 308 views
  • 0 likes
  • 2 in conversation