Hi all, I have been trying the nlm function but received an error message which reads: Error in nlm(intensities ~ f, c(epsilon.spec.start, epsilon.unspec.start, : invalid function value in 'nlm' optimizer The message is generated somewhere in the compiled part, apparently within the function static void fcn(int n, const double x[], double *f, function_info *state) where a jump to a "badvalue" label is caused on different conditions: - if the type of a value generated from the call from R is neither of type INTSXP nor REALSXP, - if the length of this value is not 1. The error message is issued after the second call to the function that nlm minimizes over. Could you give any suggestions on what I could be mishandling? Best wishes Johannes
Johannes -
There's something special about the way control parameters are
accepted by nlm() and nlme(). Try searching very recent help
archives for nlme() or control. Try duplicating exactly one of
the examples at the bottom of the help page help("nlm"). My
recollection is that the parameter that you actually pass has
to be literally a function, maybe one with a specific name.
But that's jsut vague recollection. Others will have better
information.
- tom blackwell - u michigan medical school - ann arbor -
On Mon, 1 Sep 2003, [iso-8859-1] "Hüsing, Johannes" wrote:
> I have been trying the nlm function but received an error message
> which reads:
>
> Error in nlm(intensities ~ f, c(epsilon.spec.start,
> epsilon.unspec.start, :
> invalid function value in 'nlm' optimizer
>
> The error message is issued after the second call to the function that
> nlm minimizes over.
>
> Could you give any suggestions on what I could be mishandling?
> Johannes
Hi all, I've been struggling with the following problem... Normally, if you build a package, the help files are created automatically (by 'converting' the Rd-files) in pdf- and/or html-format. I have some additional information, about the general purpose and how-to-use of the package, actually, it's a scheme. This is made in LaTeX. Can anyone give me a clue of how to include this scheme (and some text) in the pdf- (and probably html-)help files while building the package? tnx, Kurt.
On 01 Sep 2003 17:17:32 -0500, you wrote:>Duncan Murdoch <dmurdoch at pair.com> writes: > >> On Mon, 01 Sep 2003 17:24:11 -0400, you wrote: >> >> >> >You can put additional documentation in a "doc" subdirectory, >> >> That should be info/doc in the source... > >I think you mean inst/docJust not my day for accuracy, I guess! Thanks. Duncan