hi,i'm trying to build a multivariate distribution by vine copula.no need to tell that i'm rookie at r. well i have the stock price of three company for 3 years and I simply calculate the returns out of it for every single company and after getting the tree now i want to estimate the copulas and their parameter but i got this error :
cree = read.csv("cree.csv",header = TRUE)
yahoo = read.csv("yahoo.csv",header = TRUE)
google = read.csv("goog.csv",header = TRUE)
library(VineCopula)
x = pobs(as.matrix(cbind(cree,yahoo,google)))[,1]
y = pobs(as.matrix(cbind(cree,yahoo,google)))[,2]
z = pobs(as.matrix(cbind(cree,yahoo,google)))[,3]
data = cbind(x,y,z)
rvm = RVineStructureSelect(data = data,familyset = NA,type = "RVine",selectioncrit = "AIC")
RVineSeqEst(data = data,RVM = rvm,method = "itau")
Error in BiCopTau2Par(family, tau) : some tau is too close to -1 or 1
and when i try to estimate the copula i get this :
cop = RVineCopSelect(data = data,familyset = NA,Matrix = rvm,selectioncrit ="AIC",indeptest = FALSE,level = 0.5)
Error in if (nrow(args$Matrix) != ncol(args$Matrix)) stop("\n In ", args$call[1], :
argument is of length zero
any help?