turn on suggestions

Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.

Showing results for

Find a Community

- Home
- /
- SAS Programming
- /
- SAS Procedures
- /
- Nested random effects in GLIMMIX

Topic Options

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

09-24-2008 03:46 AM

Hello,

I have two categorical variables (A and B) that I want to add as Random factor in GLIMMIX. A has 4 categories and B has 44. Finally B is nested in A.

When I added as random factor, the model results said that "Estimate G matrix is not positive definite". How can I solve this problem? Did I add the variables in a correct way??

**proc GLIMMIX data=Resumen;**

class A B Cat;

model total = intra flux Cat length / dist=poisson link=log solution;

random B B(A);

run;

I have two categorical variables (A and B) that I want to add as Random factor in GLIMMIX. A has 4 categories and B has 44. Finally B is nested in A.

When I added as random factor, the model results said that "Estimate G matrix is not positive definite". How can I solve this problem? Did I add the variables in a correct way??

class A B Cat;

model total = intra flux Cat length / dist=poisson link=log solution;

random B B(A);

run;

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

Posted in reply to deleted_user

09-26-2008 08:48 AM

While this may not solve the problem completely, note that the statement:

random B B(A);

specifies factor B twice. That may be the source of the error message.

Try

random A B(A);

This should specify A as a random effect, and B within A as a second effect.

Good luck.

random B B(A);

specifies factor B twice. That may be the source of the error message.

Try

random A B(A);

This should specify A as a random effect, and B within A as a second effect.

Good luck.