| rstable {QRMlib} | R Documentation |
random sample from stable distribution
rstable(n, alpha, beta=1)
n |
sample size |
alpha |
scalar parameter strictly larger than 0 and smaller than 2 (but avoid alpha=1) |
beta |
scalar parameter between -1 and 1 |
see pages 224 and 498 of QRM; default value beta=1 combined with an alpha value less than 1 gives positive stable distribution which we require for Gumbel copula generation; the case alpha=1 has not been implemented
sample of size n
Forthcoming John Nolan Book; see Definition 1.8 and Theorem 1.19
## Not run:
#Use rstable() method in copula simulation function
rcopula.Gumbel2Gp <- function(n = 1000,
gpsizes =c(2,2), theta =c(2,3,5))
{
Y <- rstable(n,1/theta[1])*(cos(pi/(2*theta[1])))^theta[1];
innerU1 <- rcopula.gumbel(n,theta[2]/theta[1],gpsizes[1]);
innerU2 <- rcopula.gumbel(n,theta[3]/theta[1],gpsizes[2]);
U <- cbind(innerU1,innerU2);
Y <- matrix(Y, nrow = n, ncol = sum(gpsizes));
out <- exp( - ( - log(U)/Y)^(1/theta[1]));
return(out);
}
## End(Not run)