[R-sig-ME] Coefficient matrix not invertible

Rian Dickson rdd at sfu.ca
Tue Jun 21 03:19:05 CEST 2011

hello again,

Thanks for the recent answers to my other questions - they've been quite useful.
I've run into another problem a little further along. 

When I try to run this model:

global.ri.vf1.cs1.lme <- lme(Log.minH ~ site + f.year + cohort + emerg + emergSQ + pri + priSQ
  + start.t + startSQ + sea + tidem + tided, data = SUSCforage, random = ~1|scoterID, 
  weights = varExp(form =~Log.minH), correlation = corAR1(form =~ Date|scoterID), method = "REML")

I get this error message:
Error in `coef<-.corARMA`(`*tmp*`, value = 20.3619280472088) : 
  Coefficient matrix not invertible

However, if I try the same model except without the variance weighting, it works fine. 

Also, with the variance weighting included, 
corARMA(form =~ Date|scoterID, p=0, q=1) works
corARMA(form =~ Date|scoterID, p=1, q=1) returns a similar error message as above.
Error in `coef<-.corARMA`(`*tmp*`, value = c(21.0236910353017, -6.13825550757903 : 
  Coefficient matrix not invertible

I'm not sure what this means, or if there is anything I can do about it. 
Any advice would be appreciated.

thank you,



Rian Dickson
M.Sc. candidate
Centre for Wildlife Ecology
Department of Biological Sciences
Simon Fraser University
8888 University Drive
Burnaby BC V5A 1S6

More information about the R-sig-mixed-models mailing list