[R] Density Plots

Jason Turner jasont at indigoindustrial.co.nz
Mon Jan 5 02:05:47 CET 2004


David Tyler wrote (using an e-mail client that doesn't wrap lines):

> I am using an older version of R (1.6.2) to run a Monte Carlo 
 >
> simulation, generating 10,000 samples per 'run'.  When I plot 
 > histograms I get the expected 'bins' on the x-axis and the
 > frequency distribution on the y-axis. However when I ask R
 > to plot the SAME data set with a density curve the x-axis
 > emains the same but the y-axis can generate values of up to 1e8 etc.

> Can anyone (a) explain why this might be so and/or (b) suggest a fix?

try

hist(..., freq=FALSE)

This should give the same numbers as the density plots' y-axes.

It sounds like you've got a narrow range of x-axis values (small 
numbers, or small differences between them, or both).  The total area 
under a density estimate curve must equal 1 by definition, so nothing's 
really "broken".  The only fix is to re-scale the x axis to different 
units, or draw a different y-axis on after the fact.  Something like...

foo <- density(...)
plot(foo, yaxt="n")
axis(...)  # something that means something to you here.

Since this isn't a density plot any longer, it would help to be clear to 
your readers what's going on with the plots.

Hope that helps

Cheers

Jason
-- 
Indigo Industrial Controls Ltd.
http://www.indigoindustrial.co.nz
64-21-343-545
jasont at indigoindustrial.co.nz




More information about the R-help mailing list