Prior to copula fitting, we must identify the dependency between two variable. What are the purpose of calculating the dependency? Some research said dependency value for example Kendall Tau value will use to estimate the copula parameter.
My supervisor said based on the Kendall Tau value we can decide which copula is suitable because difference type of copula have difference range of Kendall Tau value.
Can anyone of you explain about the above problem.
Each copula type has its own ability to capture one or two different types of dependency structures between variables. For example, Clayton can only capture negative tail dependencies while Frank can model negative and positive dependencies. There is different methods to find the shape of copula families prior to fit it to the data. For example, you can plot the data to see what is the most appropriate copula family that can describe these data. There is a one-to-one relationship between copula and Kendall tau (not all copula families). So, if you find the Kendall tau, you can calculate the copula parameter.
Hope this is helpful.
