R help - Dec 2006 - plot.mids / Error in plot.new() : figure margins too large

Hello R-Users,

 

I would like to check the convergence of my imputations. However, when I use
the function the plot.mids(), I always obtain the following error message

 

Error in plot.new() : figure margins too large

I read the same question in thread from November 2005 (see below). I
actually have the same problem. Is it now possible to plot subsets of
mids.objects. If yes, how? 

 

My datasets has 8 variables with missing values.

 

Thanks in advance,

 

Simon

 

Attachment:

 

From: Roel de Jong <dejongroel
<mailto:dejongroel?Subject=Re:%20%5bR%5d%20plot%20of%20mice.mids%20objects>
> 
Date: Fri, 04 Nov 2005 12:57:05 +0100

Hi, 

it is impossible to plot a subset of the variables at the moment, but we 
could add it. 

p.s. The latest version of mice is v1.14, which can be downloaded from 
CRAN. 

Roel de Jong. 

Leo G?rtler wrote: 
> Hello... 
> 
> how can I plot mice.mids objects as described by Buuren (2000) 
> http://web.inter.nl.net/users/S.van.Buuren/mi/docs/Manual.pdf 
> page 17, if there are many variables (~80) with NAs included? 
> mice runs well, but the plot is not possible because it seems that there 
> are too many variables. 
> 
> > imp.temp <- mice(data.withnas, m=5) 
> > plot(imp.temp) 
> Fehler in plot.new() : Grafikr?nder zu gro? 
> = error in plot.new(): boundaries of graphics too big/huge 
> 
> Is it possible to run the plot() sequentially, so that only a selected 
> part of the variables are plotted? 
> Of course I can run mice() on a reduced dataset, but then the 
> estimations are different what is not useful here. 
> 
> thanks + best, 
> 
> leo 
> 
> 
 

 

 

 

 


	[[alternative HTML version deleted]]