Could you try upgrading to the current version of mgcv, please? looking in the `changeLog' there were a couple of fixed problems that could be causing this.
On Wednesday 21 May 2008 16:14, Lung-Chang Chien wrote: > The version is mgcv 1.3-27. Last night, I dropped out many arguments from > my program to test which part > might be the source of causing this error message, and found that the > program works when just using corAR1(). > > This program has no this kind of problem in the beginning, but doesn't work > after I impute some missing data > in PM10 and tmean by multiple imputation approach. Hope this information > can also help identify the > element of this problem. > > LC > > > From: Simon Wood <[EMAIL PROTECTED]> > > To: r-help@r-project.org > > Date: Wed, 21 May 2008 08:55:02 +0100 > > Subject: Re: [R] an unknown error message when using gamm function > > What version of mgcv are you using, please? > > > > On Wednesday 21 May 2008 06:16, Lung-Chang Chien wrote: > > Dear everyone, > > > > I'm encountering an unknown error message when using gamm function: > > > fitoutput <- > > > > gamm(cvd~as.factor(dow)+pm10+s > > (time,bs="cr",k=15,fx=TRUE)+s(tmean,bs="cr",k > > >=7,fx=TRUE) + > > ,correlation=corAR1(form=~1|city),family=poisson,random=list(city=~pm10),da > > >ta=mimp) > > > > > > Maximum number of PQL iterations: 20 > > iteration 1 > > iteration 2 > > iteration 3 > > iteration 4 > > iteration 5 > > iteration 6 > > Error in `*tmp*`[[k]] : attempt to select less than one element > > > > I never saw this error message before, and there is also no related > > information about this error from the Internet. Hope some people who once > > encountered this problem can tell me what's wrong. > > > > Many thanks, > > > > LC -- > Simon Wood, Mathematical Sciences, University of Bath, Bath, BA2 7AY UK > +44 1225 386603 www.maths.bath.ac.uk/~sw283 ______________________________________________ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
