R help - Jun 2005 - how to use plot.lda included in MASS package?

Hi there,

I am a master student in Denmark, and apply R to analyze the drug data.

I use the function 'lda' to obtain a linear discriminant object, and
then wish to use 'plot.lda' to do the plot on this object like below.

drug.lda <- lda(Inhibition ~ NET_CHARGE + PKA_1 + MW + MLOGP, 
                      data = drug.class, method = "moment")

plot.lda(drug.lda, panel = panel.lda, cex = 0.7, dimen = 1, abbrev = FALSE, 
            type ="density")

but I got an error message: Error: couldn't find function
"plot.lda"
and then I tried 

plot(drug.lda, panel = panel.lda, cex = 0.7, dimen = 1, abbrev = FALSE, 
            type ="density")

still an error message was returned: Error in plot.default(drug.lda,
panel = panel.lda, cex = 0.7,  : argument 2 matches multiple formal
arguments

Also I can't find this panel function, panel.lda anywhere. Do you know
how to use this plot function?

Thanks a lot!
wsz

The main answer is to follow the examples in the book for which this is 
support software.  (That answer is in the R posting guide, too.)

plot.lda is in the MASS namespace, and can be seen via getAnywhere() (so 
you can find them `anywhere').  panel.lda is defined in plot.lda.

You should almost never call methods like plot.lda() directly.  If you ask 
for help on it, you get

      ## S3 method for class 'lda':
      plot(x, panel = panel.lda, ..., cex = 0.7, dimen,
           abbrev = FALSE, xlab = "LD1", ylab = "LD2")

Note that you are supplying default arguments.  Just call

plot(drug.lda, dimen = 1, type ="density")

In particular, you have given a panel function for a 1D plot which does 
not use one.

I am unable to reproduce anything like the message you got, e.g.

example(z)
plot(z, panel = panel.lda, cex = 0.7, dimen = 1, abbrev = FALSE,
      type ="density")

works.  Perhaps your R or MASS package is way out of date?  (You haven't 
told us: see the R posting guide.)

On Thu, 16 Jun 2005, wu sz wrote:
> Hi there,
>
> I am a master student in Denmark, and apply R to analyze the drug data.
>
> I use the function 'lda' to obtain a linear discriminant object,
and
> then wish to use 'plot.lda' to do the plot on this object like
below.
>
> drug.lda <- lda(Inhibition ~ NET_CHARGE + PKA_1 + MW + MLOGP,
>                      data = drug.class, method = "moment")
>
> plot.lda(drug.lda, panel = panel.lda, cex = 0.7, dimen = 1, abbrev = FALSE,
>            type ="density")
>
> but I got an error message: Error: couldn't find function
"plot.lda"
> and then I tried
>
> plot(drug.lda, panel = panel.lda, cex = 0.7, dimen = 1, abbrev = FALSE,
>            type ="density")
>
> still an error message was returned: Error in plot.default(drug.lda,
> panel = panel.lda, cex = 0.7,  : argument 2 matches multiple formal
> arguments
>
> Also I can't find this panel function, panel.lda anywhere. Do you know
> how to use this plot function?
-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595