Singular Value Decomposition (SVD). where: rows of Vt are eigenvectors of DtD = basis functions. S is diagonal, with sii = sqrt(li) (ith eigenvalue). rows of U are ...