Hi Greg and Jim.
In the end I solved by modifying the function.
I was just wondering if there is some function similar to Matlab's subplot,
which let you change the active plot.
Next time I will run in the same issue I will try the split.screen
function proposed by Jim.
Thanks a lot for the help,
Cheers,
Luca
2014-06-17 18:00 GMT+02:00 Greg Snow <538...@gmail.com>:
> I am not familiar with the pamr.plotcv function, but in general if it
> uses par(mfrow=c(2,1)) to set up for the multiple plots then your
> problem with going back to the first plot is that you have lost the
> other information (such as user coordinates) needed to add to the 1st
> plot. You can see that with the following base graphics commands:
>
> par(mfrow=c(2,1))
> plot(runif(25),rnorm(25))
> tmp <- par(no.readonly=TRUE)
> hist(rnorm(1000))
>
> par(mfg=c(1,1))
> abline(h=0, col='red')
>
> The horizontal line is at what would be 0 in the lower plot, not the
> upper. Since I saved the graphics parameters I can do the correct
> thing with a command like:
>
> par(mfg=c(1,1), plt=tmp$plt, usr=tmp$usr)
> abline(h=0, col='blue')
>
> You can see the new line is in the correct place.
>
> Looking at the help for pamr.plotcv it does not look like it has any
> nice built-in ways to save the plotting parameters (some functions
> would let you plot just the top, edit, then plot just the bottom).
> Bot there is a hook to the plot.new function that can let us work
> around this. Try the following:
>
> mypars <- list()
> updatepars <- function() {
> n <- length( .GlobalEnv$mypars )
> .GlobalEnv$mypars[[ n + 1 ]] <- par(no.readonly=TRUE)
> }
> setHook('before.plot.new', updatepars)
>
> par(mfrow=c(2,1))
> plot(runif(25),rnorm(25))
> hist(rnorm(1000))
>
> setHook('before.plot.new',NULL, 'replace' ) ## clean up
>
> par( mfg=c(1,1), plt=mypars[[2]]$plt, usr=mypars[[2]]$usr )
> abline( h=0, col='blue' )
>
>
> This creates a global variable "mypars" that is an empty list (we
> should really figure out a way without using the global, but my
> current thoughts would make this much more complicated, any
> suggestions are welcome). Then it creates a small function that will
> add the current results of 'par' to that list. Then this function is
> set as a hook to be run before 'plot.new' so that any new plot will
> first save the previous parameter settings. Now we run the plotting
> commands and use the parameters that were saved into 'mypars'. I
> chose to save all the parameters from every old plot in case more than
> what is shown is needed, this could be used to go back 3 or 4 plots if
> more than just 2 are plotted.
>
> Try this with pamr.plotcv to see if it works (you may need to set some
> additional parameters depending on what all pamr.plotcv does).
>
>
> Whether this is the easy solution or not can be debated. Hope it helps,
>
>
>
>
> On Thu, Jun 12, 2014 at 9:06 AM, Luca Cerone <luca.cer...@gmail.com> wrote:
>> Dear all,
>> I am running some analysis using the pamr package (available on CRAN).
>>
>> One of the plots I produce is made using the function "pamr.plotcv".
>> This displays two plots in the same figure (using par(mfrow=c(2,1)).
>>
>> When the figure is created, I would like to be able to add some points
>> and lines, to the top plot.
>>
>> After producing the plot with pamr.cvplot, I have tried to add a line
>> doing something like:
>>
>> par(mfg=c(1,1))
>> lines(c(3,3), c(0,1), col = "blue", lty = 3)
>>
>> However this doesn't work and the line is shown in the bottom plot.
>> How can I add points and lines to the top plot?
>>
>> Thanks a lot in advance for the help,
>>
>> Cheers,
>> Luca
>>
>> ______________________________________________
>> 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.
>
>
>
> --
> Gregory (Greg) L. Snow Ph.D.
> 538...@gmail.com
--
Luca Cerone
Tel: +34 692 06 71 28
Skype: luca.cerone
______________________________________________
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.
