, but the result
is taht correlation value is about 0.5.
thankyou very much for your responses.
dian
--
View this message in context:
http://www.nabble.com/-R--Generating-bivariate-or-multivariate-data-with-known-parameter-values-tf2405919.html#a11560289
Sent from the R help mailing list archive
Greetings,
I'm interested in generating data from various bivariate or
mulitivariate distributions (e.g. gamma, t, etc), where I can specify
the parameter values, including the correlations among the variables. I
haven't been able to dig anything up on the faq, but I probably missed
you can do that using the copula package, e.g.,
library(copula)
bivGamma - mvdc(claytonCopula(2), c(gamma, gamma),
list(list(shape = 1, rate = 2), list(shape = 2, rate = 1)))
dat - rmvdc(bivGamma, 1000)
regarding correlation, the Clayton copula with alpha = 2 gives
Kendall's tau =