R help - Oct 2005 - nlme Singularity in backsolve at level 0, block 1

Hi,
 
I am hoping some one can help with this.
 
I am using nlme to fit a random coefficients model. It ran for hours before
returning
 
Error: Singularity in backsolve at level 0, block 1

The model is 
>
plavix.nlme<-nlme(PLX_NRX~loglike(PLX_NRX,PD4_42D,GAT_34D,VIS_42D,MSL_42D,SPE_ROL,XM2_DUM,THX_DUM,b0,b1,b2,b3,b4,b5,b6,b7,alpha),
+ data=data,
+ fixed=list(b0 + b1+b2+b3+b4+b5+b6+b7+alpha~1),
+ random=b0+b1+b2+b3+b4+b5+b6+b7~1|menum,
+ 
+ start=c(b0=0,b1=0,b2=0,b3=0,b4=0,b5=0,b6=0,b7=0,alpha=5)
+ )

Can anyone tell me what this error means and how I can run the model?
 
Thanks,
 
Elizabeth Lawson

		
---------------------------------

	[[alternative HTML version deleted]]

RSiteSearch("Singularity in backsolve") produced 33 hits, the second 
of which was the following:

http://finzi.psych.upenn.edu/R/Rhelp02a/archive/62691.html

	  This reply from Peter Dalgaard dates 8 Oct. 2005 asks, Which version 
of R and NLME? R 2.2.0 ships with a version where the internal optimizer 
is changed to nlminb(). As I understand it, this was in response to 
reports where code that worked in S-PLUS refused to
work in R."

	  If you are using R 2.2.0 and the latest version of nlme, then it's 
possible (likely?) that your model is overparameterized, and can't be 
fit with nlme.  Can you fit the model using, e.g., nls ignoring the 
random effects model, as suggested in ch. 8 of Pinheiro and Bates (2000) 
Mixed-Effects Models for S and S-Plus (Springer)?  If no, it's virtually 
certain that nlme won't work, either.  If I had trouble diagnosing the 
problem, I might recode it for optim(..., hessian=T), then look at the 
eigenvalues and vectors of the hessian to understand the deficiencies in 
the model.

	  If nls works for you but not nlsList, there still might be hope, but 
it would be more difficult.  If I had that problem and it were 
sufficiently important, I would step through nlme until I could figure 
out how to modify the code so it would continue, as does optim, and 
provide answers that would help me diagnose the singularity.  (I'd also 
include a facility for providing a prior distribution that should 
eliminate the singularities.)

	  Hope this helps.
	  spencer graves

Elizabeth Lawson wrote:
> Hi,
>  
> I am hoping some one can help with this.
>  
> I am using nlme to fit a random coefficients model. It ran for hours before
returning
>  
> Error: Singularity in backsolve at level 0, block 1
> 
> The model is 
> 
>>plavix.nlme<-nlme(PLX_NRX~loglike(
PLX_NRX,PD4_42D,GAT_34D,VIS_42D,MSL_42D,SPE_ROL,XM2_DUM,THX_DUM,b0,b1,b2,b3,b4,b5,b6,b7,alpha),
> 
> + data=data,
> + fixed=list(b0 + b1+b2+b3+b4+b5+b6+b7+alpha~1),
> + random=b0+b1+b2+b3+b4+b5+b6+b7~1|menum,
> + 
> + start=c(b0=0,b1=0,b2=0,b3=0,b4=0,b5=0,b6=0,b7=0,alpha=5)
> + )
> 
> Can anyone tell me what this error means and how I can run the model?
>  
> Thanks,
>  
> Elizabeth Lawson
> 
> 		
> ---------------------------------
> 
> 	[[alternative HTML version deleted]]
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide!
http://www.R-project.org/posting-guide.html
-- 
Spencer Graves, PhD
Senior Development Engineer
PDF Solutions, Inc.
333 West San Carlos Street Suite 700
San Jose, CA 95110, USA

spencer.graves at pdf.com
www.pdf.com <http://www.pdf.com>
Tel:  408-938-4420
Fax: 408-280-7915