wolffd@0: function p = matrix_T_pdf(A, M, V, K, n)
wolffd@0: % MATRIX_T_PDF Evaluate the density of a matrix under a Matrix-T distribution
wolffd@0: % p = matrix_T_pdf(A, M, V, K, n)
wolffd@0: 
wolffd@0: % See "Bayesian Linear Regression", T. Minka, MIT Tech Report, 2001
wolffd@0: 
wolffd@0: [d m] = size(K);
wolffd@0: is = 1:d;
wolffd@0: c1 = prod(gamma((n+1-is)/2)) / prod(gamma((n-m+1-is)/2));
wolffd@0: c2 = det(K)^(d/2) / det(pi*V)^(m/2); %% pi or 2pi?
wolffd@0: p = c1 * c2 * det((A-M)'*inv(V)*(A-M)*K + eye(m))^(-n/2);
wolffd@0: