Contents

function [y,x] = create_dummies(lamda,tau,delta,epsilon,p,mu,sigma,n)

% Creates matrices of dummy observations [...];
%lamda tightness parameter
%tau  prior on sum of coefficients
%delta prior mean for VAR coefficients
% epsilon tigtness of the prior around constant
% mu sample mean of the data
% sigma AR residual variances for the data




% Initialise output (necessary for final concatenation to work when tau=0):
x = [];
y = [];
yd1 = [];
yd2 = [];
xd1 = [];
xd2 = [];

Get dummy matrices in equation (5) of Banbura et al. 2007:

if lamda>0
    if epsilon >0
	yd1=[diag(sigma.*delta)./lamda;
         zeros(n*(p-1),n);
         diag(sigma);
         zeros(1,n)];

	jp=diag(1:p);

	xd1=[kron(jp,diag(sigma)./lamda) zeros((n*p),1);
         zeros(n,(n*p)+1);
         zeros(1,n*p) epsilon];


else

   yd1=[diag(sigma.*delta)./lamda;
         zeros(n*(p-1),n);
         diag(sigma)];

	jp=diag(1:p);

	xd1=[kron(jp,diag(sigma)./lamda);
         zeros(n,(n*p))];
    end
end

Input argument "lamda" is undefined.

Error in ==> create_dummies at 22
if lamda>0

Get additional dummy matrices - see equation (9) of Banbura et al. 2007:

if tau>0
    if epsilon>0
	yd2=diag(delta.*mu)./tau;
	xd2=[kron((1:p),yd2) zeros(n,1)];
    else
      yd2=diag(delta.*mu)./tau;
	xd2=[kron((1:p),yd2)];
    end
end

y=[yd1;yd2];
x=[xd1;xd2];

Published with MATLAB® 7.9