## samples from the multivariate Cauchy distribution:
bfp:::rmvt(20)
## here the covariance exists:
sigma <- matrix(c(1, 0.5, 0.5, 1), nrow=2)
df <- 10
## theoretical covariance:
sigma * df / (df - 2)
## this should be close:
cov(bfp:::rmvt(n=100000, sigma=sigma, df=df))
Run the code above in your browser using DataLab