R help - Jul 2007 - Error Coefficient matrix not invertible using nlme correlation

Hi,



I am learning R. I am using nlme function to my dataset. I am getting error
"Coefficient matrix not invertible". I don't know how to fix this
problem.


> bcows1.nlme<-nlme(Temp~model(newTime, beta, delta, kappa),
databcows.group,
+              fixed=beta+delta+kappa~1,

+              random=beta+delta+kappa~1,

+              start=list(fixed=c(coef(bcows.nls)[1], coef(bcows.nls)[2],
coef(bcows.nls)[3])))
> bcows2.nlme<-update(bcows1.nlme, correlation=corARMA(p=2))
Error in "coef<-.corARMA"(`*tmp*`, value = c(210.907034244201,
197.591708911204 :

        Coefficient matrix not invertible
>


Please anyone could tell me how to fix this problem.





Thanks

	[[alternative HTML version deleted]]