and the correlation matrices problem data 
and 
which results from the function ccama.m and returns the state-space realization of the linear filter which generates the suitable colored-in-time forcing into the linear dynamics.% Filter realization % % Written by Armin Zare and Mihailo Jovanovic, April 2016 %
function [Af,Bf,Cf,Df] = linfilter(A,X,Z) % Spectal decomposition % B is the input matrix; H is the cross-correlation of state and forcing [B, H, S] = Qdecomposition(Z); % nb is the rank of B or required nubmer of input channels [mb, nb] = size(B); % Rho, covariance of white noise Rho = eye(nb); % Filter realization Cf = (-.5*Rho*B'+ H')/X; Af = A + B*Cf; Bf = B; Df = eye(nb); end