#Quantile
qq1<-quantile(P100m,probs=seq(0.95,1,0.01),type=1)
qq1
qq<-matrix(0,9,6)
for (i in 1:9) qq[i,]<-quantile(P100m,probs=seq(0.95,1,0.01),type=i)
qq
r1<-rnorm(20)
for (i in 1:9) qq[i,]<-quantile(r1,probs=seq(0.95,1,0.01),type=i)
qq

#Vergleich Gleichverteilung und Normal
ru<-runif(100,0,1)
rn<-rnorm(100,1/2,1/(12**0.5))
par(mfrow=c(3,1))
hist(ru)
hist(rn)
qqplot(ru,rn)

#Vergleich zweier empirischen Verteilungen (Gosset Gerste)
par(mfrow=c(3,1))
hist(GG$REG)
hist(GG$KILN)
summary(GG$REG)
summary(GG$KILN)
qqplot(GG$REG,GG$KILN,xlim=c(1250,2500),ylim=c(1250,2500))

#Vergleich zweier empirischen Verteilungen (ZK)
qqplot(P100m,Plj,xlim=c(min(min(P100m),min(Plj)),max(max(P100m),max(Plj))),ylim=c(min(min(P100m),min(Plj)),max(max(P100m),max(Plj))))
abline(0,1)

#1-d DichteschŠtzer Hidalgo Beispiel
ihist(H$ThouStampTh)
dh<-density(H$ThouStampTh)
plot(dh)
dh<-density(H$ThouStampTh, width=2)
plot(dh)

#2-d DichtschŠtzer
plot(P100m,Plj)
plot(P100m,Plj,pch=19)

#Kernel
f1<-kde2d(P100m,Plj)
JavaGD()
plot(f1)
contour(f1, xlab = "P100m",
+ ylab = "Plj", levels  =  c(0.05, 0.1, 0.2, 0.4) )
image(f1, zlim = c(0, 0.05))

summary(f1$z)
max(f1$z)

contour(f1, xlab = "P100m",ylab = "Plj", levels  =  c(0.00003,0.000015,0.000005,0.0000025,0.000001))
image(f1, zlim = c(0, 0.000035))
image(f1, zlim = c(0, 0.000005))
image(f1, zlim = c(0, 0.00005))

#ASH
library(ash)
pp<-cbind(P100m,Plj)
b2<-bin2(pp,n=c(50,50))
fash<-ash2(b2)
image(fash$x,fash$y,fash$z)
contour(fash$x,fash$y,fash$z,add=TRUE)