Stopped because of infinate likelihood

Reply
Contributor
Posts: 44

Stopped because of infinate likelihood

Hi all,

i know this message will come
      

  1. If duplicates exist in our data – i don't have duplicates
  2. Don’t have enough data to estimate – i have less observations for perticular visit.
  3. Nonpositive residual variance estimate - bcz of less obs i am getting this warning.

so in this case how i come out from this problem?? please suggest me what i have to do?

Best Regards,

Hima


 
 
 
 
 
 
 
 
 






 
 
 
 
 
 
 
 





Grand Advisor
Posts: 9,576

Re: Stopped because of infinate likelihood

What model are you using ? consider using exact method if you could . post it at would be better.

SAS Statistical Procedures

Respected Advisor
Posts: 2,655

Re: Stopped because of infinate likelihood

It could be an overspecified G side and thus the R side infinte likelihood error arises.

Steve Denham

Contributor
Posts: 44

Re: Stopped because of infinate likelihood

i used proc mixed model (MMRM) to my data . how i come out of this problem!

please give me detailed discription..

Respected Advisor
Posts: 2,655

Re: Stopped because of infinate likelihood

First, please share the code you are using.  It is very difficult to troubleshoot when we don't know what you are doing currently.  Second, give us an idea of the design, and of how your data are currently shaped.  Then we can see if it is a dataframing problem or an instance of extremely pathological data.

Steve Denham

Frequent Contributor
Posts: 144

Re: Stopped because of infinate likelihood

Have you check that all variables that appears in the CLASS have more than a level?

Ask a Question
Discussion stats
  • 5 replies
  • 213 views
  • 2 likes
  • 4 in conversation