R help - Apr 2003 - superimposing graphs

Dear People,

I have a data set of data x from a probability distribution, and I have a
function, mydensity, of the pdf of that distribution.

I'm asking for help in superimposing the histogram of x and the plot of
mydensity.

In the function below, I call truehist and curve, but these are plotted in
different figures.

I'd like them to be plotted on the same figure, and to use common axes, of
course. Can anyone suggest a simple, robust way to do this? I'd prefer not
to do micro-fiddling of graphical parameters if possible.

I'm currently searching R-help mail archives to see if I can turn up
something.

A couple of people suggest par(new=TRUE), but this superimposes
everything, including the axes.

I notice that there is something very similar in the trellis demo, namely

 print(densityplot(~x))
  print(histogram(x, type = "density",
                  panel = function(x, ...) {
                    panel.histogram(x, ...)
                    panel.densityplot(x, col = "brown",
plot.points = FALSE)
                  }))

which plots a histogram and a density plot on top of that. However, I
cannot get the syntax to work right. (I'm not at all familar with trellis
graphics).  Does anyone know whether I could turn this to my purposes?
If so, the exact syntax required would be helpful.

Thanks in advance.

                                                      Faheem.

***********************************************************************

mg.hist <- function(len,theta,pos,size)
{
  x <- empmargdistvec(len,theta,pos,size)


  postscript(file="plot.ps", horizontal = FALSE, onefile = FALSE,
paper
             = "special", width=6, height=4)
  par(mfcol=c(1,2),pch=20)

  truehist(x, nbins=100)

  mydensityfn <- function(x)
    {
      mydensity(x,theta,pos,len)
    }
  curve(mydensityfn)

  dev.off()

Faheem this has been asked several times in the past couple of months.

check ?curve and the parameter 'add' as in

curve(fc(x), 0, 10, col = "darkblue", add = TRUE) # from the reference
manual

If this isn't what you want then search Jonathon Baron's archive at 
http://finzi.psych.upenn.edu/search.html for numerous other solutions,



Richard Rowe
Senior Lecturer
Department of Zoology and Tropical Ecology, James Cook University
Townsville, Queensland 4811, Australia
fax (61)7 47 25 1570
phone (61)7 47 81 4851
e-mail: Richard.Rowe at jcu.edu.au
http://www.jcu.edu.au/school/tbiol/zoology/homepage.html

On Fri, Apr 18, 2003 at 07:22:58PM -0400, Faheem Mitha
wrote:
> In the function below, I call truehist and curve, but these are plotted in
> different figures.
> 
> I'd like them to be plotted on the same figure, and to use common axes,
of
> course. Can anyone suggest a simple, robust way to do this? I'd prefer
not
> to do micro-fiddling of graphical parameters if possible.
Here is one way with an estimated density function, IIRC more a less a copy
of what is in MASS.
> library(MASS)
> X<-rt(250,2)
> Xdensity <- density(X, width=width.SJ(X, method="dpi"), n=200)
> truehist(X, col="lightgray", ymax=max(Xdensity$y))
> lines(Xdensity)
> title(main="Histogram and Density")
> box()
Peter has another solution in ISwR, but that may have been particular to a
Normal distribution, and my copy is at work anyway.

Hth,  Dirk

-- 
Don't drink and derive. Alcohol and algebra don't mix.

You can use curve(), but you need to add the `add = TRUE' argument.  For
example,

library(MASS)
x <- rnorm(100)
truehist(x)
curve(dnorm, -3, 3, add = TRUE)


-roger
_______________________________
UCLA Department of Statistics
http://www.stat.ucla.edu/~rpeng

On Fri, 18 Apr 2003, Faheem Mitha wrote:
> 
> Dear People,
> 
> I have a data set of data x from a probability distribution, and I have a
> function, mydensity, of the pdf of that distribution.
> 
> I'm asking for help in superimposing the histogram of x and the plot of
> mydensity.
> 
> In the function below, I call truehist and curve, but these are plotted in
> different figures.
> 
> I'd like them to be plotted on the same figure, and to use common axes,
of
> course. Can anyone suggest a simple, robust way to do this? I'd prefer
not
> to do micro-fiddling of graphical parameters if possible.
> 
> I'm currently searching R-help mail archives to see if I can turn up
> something.
> 
> A couple of people suggest par(new=TRUE), but this superimposes
> everything, including the axes.
> 
> I notice that there is something very similar in the trellis demo, namely
> 
>  print(densityplot(~x))
>   print(histogram(x, type = "density",
>                   panel = function(x, ...) {
>                     panel.histogram(x, ...)
>                     panel.densityplot(x, col = "brown",
> plot.points = FALSE)
>                   }))
> 
> which plots a histogram and a density plot on top of that. However, I
> cannot get the syntax to work right. (I'm not at all familar with
trellis
> graphics).  Does anyone know whether I could turn this to my purposes?
> If so, the exact syntax required would be helpful.
> 
> Thanks in advance.
> 
>                                                       Faheem.
> 
> ***********************************************************************
> 
> mg.hist <- function(len,theta,pos,size)
> {
>   x <- empmargdistvec(len,theta,pos,size)
> 
> 
>   postscript(file="plot.ps", horizontal = FALSE, onefile = FALSE,
paper
>              = "special", width=6, height=4)
>   par(mfcol=c(1,2),pch=20)
> 
>   truehist(x, nbins=100)
> 
>   mydensityfn <- function(x)
>     {
>       mydensity(x,theta,pos,len)
>     }
>   curve(mydensityfn)
> 
>   dev.off()
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://www.stat.math.ethz.ch/mailman/listinfo/r-help
>