Estimate parameter of Frank Copula

Question

I'm analyzing some bivariate data with copulas. In order to simulate some copulas I need to estimate the corresponding parameters.

For example:

gumbel.fit<-fit.AC(Udata, "gumbel")
gumbel.parameter<-gumbel.fit$fit$par

or

clayton.fit<-fit.AC(Udata, "clayton")
clayton.parameter<-clayton.fit$fit$par

But this can't be applied to Frank copula, therefor I wonder how to estimate the paramter?

• Are there any function in r that allows me to do this?
• The only way to do this is programming the maximum likelihood?

Answer 1

You can estimate any copula parameter using for example "CDvine" package in R. In this case you can select any available method. There is a "Tau" estimation method or the pseudo-observation maximum likelihood method. Currently, to my best of knowledge, there are no further estimation methods available in the R.

Answer 2

Please have a look at the R package copula which provides various estimators including detailed examples (specifically, numerically stable MLE).

Answer 3

I found the answer in the CDVine package BiCopEst(u1,u2,5,method="mle")