[R-sig-ME] Help! "Singular convergence problem" and "Singular precision matrix"!

Ben Bolker bbolker at gmail.com
Sat May 12 06:57:21 CEST 2012


Michael <comtech.usa at ...> writes:

> Error in lme.formula(y ~ X - 1, data = tmp, weights = mycov, random =
> myrnd, : nlminb problem, convergence error code = 1
>   message = singular convergence (7)
> 
> Also, on different data sets, I've got the following:
> 
> Warning messages:
> 1: In logLik.reStruct(object, conLin) :
>   Singular precision matrix in level -1, block 3
> 
> Could you please shed some lights on this?

  You've probably got an overfitted model or wonky data, but it's
very hard to say without a reproducible example ... see e.g
http://tinyurl.com/reproducible-000



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