Estimation of Structured t-Copulas: LogLik(x,Nu,Sigma) - File Exchange

Code covered by the BSD License

Highlights from
Estimation of Structured t-Copulas

LogLik(x,Nu,Sigma)
MleRecursionForT(x,Nu,K,Toler... this function computes recursively the ML estimators of the correlation
[Nu,C]=StrucTMLE(X,K,Toleranc... this function computes recursively the ML estimators of the correlation
[W,F,U]=SeparateMargCop(X) this function separates the empirical copula from the marginal distribution
S_ReadMe.m
View all files

from Estimation of Structured t-Copulas by Attilio Meucci
Recursive routine to estimate structured correlation matrix and degrees of freedom

LogLik(x,Nu,Sigma)

function LL = LogLik(x,Nu,Sigma)

InvS=inv(Sigma);
N=size(x,2);

Norm=-N/2*log(Nu*pi) + log(gamma((Nu+N)/2)) - log(gamma(Nu/2)) - 1/2*log(det(Sigma));

LL=0;
for t=1:size(x,1)
    Ma2 = x(t,:)*InvS*x(t,:)';
    LL = LL + Norm - (Nu+N)/2 * log(1+Ma2/Nu);
end

Highlights from Estimation of Structured t-Copulas

Highlights from
Estimation of Structured t-Copulas