[R] vioplot kernel smooth via density kernel smooth

[Johannes Graumann]

Hi,

Vioplots have great appeal to me, as they manage to squeeze so much
information into so little space ...
Now some evaluation has made me suspicious about the implementation in the
package vioplot and I would like to hear what you say about the appended
results and the code going with it:

library(vioplot)
mydata <- read.table("data") # file "data" attached
plot(density(mydata[[1]])) # attached as densityplot.png
vioplot(mydata[[1]])  # attached as vioplot.png

Default settings for "density" seem to clearly indicate that the
distribution is bimodal, while "vioplot" total misses that. 
Can "vioplot" be made to react more "density"-like or is there an
alternative, more "density"-like implementation?

Thanks for your imput, Joh
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: data
Url: https://stat.ethz.ch/pipermail/r-help/attachments/20080425/e6296cef/attachment.pl 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: densityplot.png
Type: image/png
Size: 34087 bytes
Desc: not available
Url : https://stat.ethz.ch/pipermail/r-help/attachments/20080425/e6296cef/attachment.png 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: vioplot.png
Type: image/png
Size: 36948 bytes
Desc: not available
Url : https://stat.ethz.ch/pipermail/r-help/attachments/20080425/e6296cef/attachment-0001.png