7.1-final-2001¶
Final Exam, 2001 · Section 7.1: Singular Values and Singular Vectors Source: finalexam_1.pdf, p. 58
Problem¶
6 (36 pts.) Let \(A\) be a symmetric matrix of size 3-by-3 with 3rd column \(\begin{bmatrix} 1 \\ 0 \\ 0 \end{bmatrix}\).
(a) Assume \(A\) has the cofactors \(C_{13} = 0\) and \(C_{23} = 1\), and
\[
A \begin{bmatrix} 1 \\ 1 \\ 1 \end{bmatrix} = \begin{bmatrix} 3 \\ 1 \\ 1 \end{bmatrix}.
\]
Give \(A\) explicitly. Is \(A\) invertible?
(b) Compute the SVD of \(A\).
(c) With the help of the SVD, give an orthonormal basis of the column space of \(A\).
Solution
(a) Symmetry and the third column \((1, 0, 0)\) give \[ A = \begin{bmatrix} a & b & 1 \\ b & d & 0 \\ 1 & 0 & 0 \end{bmatrix}. \] The cofactor \(C_{13} = \det\begin{bmatrix} b & d \\ 1 & 0 \end{bmatrix} = -d = 0\), so \(d = 0\). The cofactor \(C_{23} = -\det\begin{bmatrix} a & b \\ 1 & 0 \end{bmatrix} = b = 1\). Then \(A(1, 1, 1) = (a + 2, 1, 1) = (3, 1, 1)\) gives \(a = 1\). Thus \[ A = \begin{bmatrix} 1 & 1 & 1 \\ 1 & 0 & 0 \\ 1 & 0 & 0 \end{bmatrix}. \] Rows 2 and 3 are equal, so \(\det A = 0\): \(A\) is not invertible (rank 2). (b) \(A^TA = \begin{bmatrix} 3 & 1 & 1 \\ 1 & 1 & 1 \\ 1 & 1 & 1 \end{bmatrix}\) has eigenvalues 4, 1, 0 (the 2-by-2 block on the subspace \(x_2 = x_3\) has characteristic \(\lambda^2 - 5\lambda + 4\)), so the singular values are \(\sigma = 2, 1, 0\) with \[ V = \begin{bmatrix} 2/\sqrt{6} & -1/\sqrt{3} & 0 \\ 1/\sqrt{6} & 1/\sqrt{3} & 1/\sqrt{2} \\ 1/\sqrt{6} & 1/\sqrt{3} & -1/\sqrt{2} \end{bmatrix}, \qquad \Sigma = \begin{bmatrix} 2 & 0 & 0 \\ 0 & 1 & 0 \\ 0 & 0 & 0 \end{bmatrix}. \] Then \(u_1 = Av_1/2 = (2, 1, 1)/\sqrt{6}\), \(u_2 = Av_2 = (1, -1, -1)/\sqrt{3}\), and \(u_3 = (0, 1, -1)/\sqrt{2}\) completes an orthonormal basis of \(R^3\). The SVD is \[ A = U\Sigma V^T, \qquad U = \begin{bmatrix} 2/\sqrt{6} & 1/\sqrt{3} & 0 \\ 1/\sqrt{6} & -1/\sqrt{3} & 1/\sqrt{2} \\ 1/\sqrt{6} & -1/\sqrt{3} & -1/\sqrt{2} \end{bmatrix}. \] (c) The columns of \(U\) corresponding to the nonzero singular values form an orthonormal basis of \(C(A)\): \[ u_1 = \frac{1}{\sqrt{6}}\begin{bmatrix} 2 \\ 1 \\ 1 \end{bmatrix}, \qquad u_2 = \frac{1}{\sqrt{3}}\begin{bmatrix} 1 \\ -1 \\ -1 \end{bmatrix}. \]Keywords: singular value decomposition, cofactors, symmetric matrix, column space