Skip to content

Help for L-moment Ratio Diagram

3 messages · amna khan, Eric Thompson, William Asquith

#
I think there are a few packages for doing this. I have used "lmomco":
the function lmom.ub() will calculate the sample lmoments, and
lmrdia() gives theoretical lmoments for different distributions.

Hope this is helpful.

Eric
On 12/4/06, amna khan <amnakhan493 at gmail.com> wrote:
#
Amina,
If you can follow this code, you will see that the black dots scatter  
around an L-skew of zero and L-kurtosis of 0.122602, which are the  
theoretical L-moments of the standard normal distribution.

library(lmomco)
lmrdia <- lmrdia()
plotlmrdia(lmrdia,ylim=c(-.2,.2),xlim=c(-.2,.2))
lms <- lmoms(rnorm(500))
points(lms$ratios[3],lms$ratios[4],pch=16)
lms <- lmoms(rnorm(500))
points(lms$ratios[3],lms$ratios[4],pch=16)
lms <- lmoms(rnorm(500))
points(lms$ratios[3],lms$ratios[4],pch=16)
lms <- lmoms(rnorm(500))
points(lms$ratios[3],lms$ratios[4],pch=16)

Further interpretion of the diagram is well documented in Hosking's  
1990  paper and the references shown on the help page.

William
On Dec 4, 2006, at 12:05 PM, Eric Thompson wrote: