Dear List,
I fitted a bivariate extreme value distribution to my matrix (data). Now
I would like to add ?quantile curves? to the plot, similar to
those calculated with
qcbvnonpar()
But I need the
probability of the upper right corner instead of the lower left corner.
(So that the curves in the plot below would look like they were turned
by 180 degrees).
Is there a package that could help me? Or is
there an option within the ?evd? package that I didn?t see?
My
data:
a<-c(6.8044, 6.4422, 6.0900, 6.1778, 6.2778, 6.3978,
6.2156, 6.1512, 6.2712, 6.3112, 6.1590, 6.2368, 6.1968, 6.1746, 6.2202,
6.5836, 6.1970, 6.2870, 6.1126, 6.3460, 6.0616, 6.2016, 6.3294, 6.1494,
6.0350, 6.2006, 6.1106, 6.3506, 6.3706, 6.0484, 6.3140, 6.0618, 6.0418,
6.4618, 6.1996, 6.3196, 6.0996, 6.2174, 6.1574, 6.0774, 6.1874, 6.0230,
6.0608, 6.3386, 6.0686, 6.3042, 6.2042, 6.6242, 6.0720, 6.1276, 6.1410,
6.1010, 6.2688, 6.5788, 6.3566, 6.1766, 6.1444, 6.6200, 6.0178, 6.1856,
6.1634)
b<-c(20, 16, 12, 16, 10, 26, 11, 5, 12, 16, 6, 12, 9, 15, 13, 22, 10,
15, 10, 9, 5, 8, 30, 8, 5, 8, 7, 23, 18, 5, 8, 6, 5, 28, 6, 9, 7, 8, 6,
5, 6, 6, 6, 15, 6, 8, 10, 26, 7, 5, 8, 8, 6, 11, 8, 8, 11, 16, 6, 12,
12)
data<-cbind(a,b)
bifit<-fbvevd(data,model="log")
plot(bifit,which=5)
Thanks a lot,
Aishe