EViews.com

I am trying to estimate an eigenvalue decomposition for Matrix A using the makepcomp command. However the eigenvectors (say vector M) and the eigenvectors (say matrix U) estimated do not give back the matrix A when I am trying to estimate UxMx@inverse(U). The commands I use are the following:
A.makepcomp(cov=cov, cpnorm, eigval=M, eigvec=U)
Matrix X=Ux@makediagonal(M)x@inverse(U)

makepcomp computes the principal components decomposition of the covariances between the columns of the matrix. I don't think that's the computation that you want.