In this question it has been described how to generate a random variable with a defined correlation to an existing variable. My question is what happens if I want to have a specific pattern of correlation, meaning having a copula connecting the two variables. For example if the two variables are normally distributed then it comes up a normal copula. But what if I want to have the two random variables normally distributed but connected with a student T copula. How the answer in this question be generalised to that case.
$\begingroup$
$\endgroup$
2
-
$\begingroup$ If the desired correlation is compatible with a $t$ copula, the explanation is the same. $\endgroup$– Xi'anCommented Sep 16, 2022 at 16:07
-
1$\begingroup$ I do not understand the comment. How you introduce in the explanation the copula? Could you be more specific? $\endgroup$– krikitosCommented Sep 16, 2022 at 19:20
Add a comment
|