[R] Problems with coda

[Geir Olve Storvik]

Thanks to ShumTim for comments on the problems regarding CODA. 
As noted, the "missing" problem is not the only one in CODA, and
correcting this
only made CODA to crash at another place. It therefore seems best to
wait for
an updated version of CODA

Geir

LUK ShunTim wrote:
> 
> Geir Olve Storvik wrote:
> >
> > I have successfully installed coda in R, and I am also able to read
> > output from a Bugs run into
> > coda. When I try to plots from CODA Output Analysis Menu, I do however
> > _get_ problems:
> > CODA Output Analysis Menu
> > *************************
> >
> > 1:Plots
> > 2:Statistics
> > 3:List/Change Options
> > 4:Return to Main Menu
> > Selection: 1
> > Error: "missing" illegal use of missing
> >
> > Quitting CODA....
> >
> > Do you want to save the BUGS output as an R object file(y/n) ?
> > 1:
> >
> > I tried to use the same data in the Splus version of coda, and that
> > worked fine.
> > I have also got the same type of problem with coda in R with output from
> > a completely different model.
> >
> 
> Geir,
> 
> I had the same problem and got it fixed by adding an additional argument
> "bwf" to the "densplot" function like this:
> 
> "densplot" <-function (x, combine.chains = F, bwf)
> {
>         x <- as.mcmc(x)
> # *** Argument bwf required        do.densplot <- function(j, i) {
> #     Like the next line
>          do.densplot <- function(j, i, bwf) {
>                 #NB Using R scoping rules
>                 #
>                 if (missing(i))
>                         y <- as.vector(x[, j, , drop = T])
>                 else y <- as.vector(x[, j, i, drop = T])
>                 y.orig <- y
> # *** R complains about the next line, "illegal use of missing"
>                 if (missing(bwf)) {
>                         N <- length(!is.na(y))
>                         bw <- 1.06 * min(sd(y), IQR(y)/1.34) *
>                                 N^-0.2
>                 }
>                 else {
>                         bw <- bwf(y)
>                         if (bw == 0)
>                                 bw <- 0.1
>                 }
> ...
> 
> There are a couple of other bugs, too.
> I have notified the author/maintainer of the package and according to
> him, it may be due to changes in R between 0.62 and 0.63 and he is
> trying to look at the problem.
> 
> Regards,
> 
> ST
> --

-- 
          Geir Storvik                        Tlf: (+47) 22855894
          1. Amanuensis                       Fax: (+47) 22854349
          Department of Mathematics           email: geirs at math.uio.no
          Statistics Division
          University of Oslo
          P.O.Box 1053 Blindern
          N-0316 Oslo
          Norway
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._