Help: CDF plot

 Thanks again.  I want to create a plot  with  cdf of all realizations
overlay with observed  cdf value  just for comparing the distribution
of  all realizations  with observed data (instead of histogram). I have
already done threshold part. I sed 1st, 5th and 9th decile values as
thresholds.   Thanks
Zia

On 10/21/2010 3:12 PM, Edzer Pebesma wrote:

On 10/21/2010 08:42 PM, Zia Ahmed wrote:

I want to write a R- function to make a CDF plot of all realizations
(sim1, sim2, sim3........) with CDF of observed data (suppose zinc of
muse data). Help will be appreciated

It can be me, but after reading your wish several times, I still don't
understand what you want. Do you want to make a cdf plot of each
realization?

plot(ecdf(lzn.sim[[1]])) # etc

Or for each location over all realizations? For threshold z:

z = 5.0
nsim = 6
lzn.sim$z = apply(as.data.frame(lzn.sim)[-(1:2)]<= z, 1, sum)/nsim
spplot(lzn.sim["z"])

Or of the observed values?

plot(ecdf(meuse$zinc))

Thanks
Zia

library(gstat)

data(meuse)

data(meuse.grid)

coordinates(meuse) = ~x + y

coordinates(meuse.grid) = ~x + y

lzn.vgm<- variogram(log(zinc) ~ 1, meuse)

lzn.fit<- fit.variogram(lzn.vgm, model = vgm(1, "Sph", 900, 1))

lzn.sim<- krige(log(zinc) ~ 1, meuse, meuse.grid, v.fit, nsim = 6, nmax
= 40)

/lzn.sim at data<- exp(lzn.sim at data)/

/ /


/ /