fit.tcopula.rank {QRMlib} | R Documentation |
fits t copula to pseudo-copula data
fit.tcopula.rank(Udata, method="Kendall")
Udata |
matrix of pseudo-copula data where rows are vector observations with all values in unit interval |
method |
method to use for calculating rank correlations; default is "Kendall", which is theoretically justified |
see pages 229-231 of QRM
list containing parameter estimates and details of fit
fit.tcopula
,
fit.gausscopula
,
fit.Archcopula2d
# Multivariate Fitting with Gauss and t: Simulation # Create an equicorrelation matrix: P <- equicorr(3,0.6); set.seed(113); #Generate a new set of random data from a t-copula (10df) with the same Sigma matrix: Udatasim2 <- rcopula.t(1000,df=10,Sigma=P); #Now fit the copula to the simulated data using (Kendall) rank correlations #and the fit.tcopula.rank() method: mod.t2 <- fit.tcopula.rank(Udatasim2); mod.t2;