7.1-final-2001-spring-conflict¶
Final Exam (Conflict), 2001 (spring-conflict) · Section 7.1: Singular Values and Singular Vectors Source: finalexam_1.pdf, p. 76
Problem¶
6. (15pts.) Find the singular value decomposition of the 3 by 2 matrix
\[
A = \begin{bmatrix} 1 & 2 \\ -2 & -4 \\ 1 & 2 \end{bmatrix}.
\]
Solution
The second column of \(A\) is twice the first, so \(A\) has rank 1. Compute \(A^TA\): \[ A^TA = \begin{bmatrix} 6 & 12 \\ 12 & 24 \end{bmatrix}, \] with eigenvalues \(30\) and \(0\), so the singular values are \(\sigma_1 = \sqrt{30}\) and \(\sigma_2 = 0\). Eigenvectors of \(A^TA\): for \(\lambda = 30\), \(v_1 = \frac{1}{\sqrt{5}}\begin{bmatrix} 1 \\ 2 \end{bmatrix}\); for \(\lambda = 0\), \(v_2 = \frac{1}{\sqrt{5}}\begin{bmatrix} 2 \\ -1 \end{bmatrix}\). So \[ V = \frac{1}{\sqrt{5}} \begin{bmatrix} 1 & 2 \\ 2 & -1 \end{bmatrix}, \qquad \Sigma = \begin{bmatrix} \sqrt{30} & 0 \\ 0 & 0 \\ 0 & 0 \end{bmatrix}. \] Then \(u_1 = Av_1/\sigma_1 = (5, -10, 5)/(\sqrt{5}\sqrt{30}) = \frac{1}{\sqrt{6}}\begin{bmatrix} 1 \\ -2 \\ 1 \end{bmatrix}\). Complete to an orthonormal basis of \(R^3\): \[ u_2 = \frac{1}{\sqrt{5}} \begin{bmatrix} 2 \\ 1 \\ 0 \end{bmatrix}, \qquad u_3 = \frac{1}{\sqrt{30}} \begin{bmatrix} -1 \\ 2 \\ 5 \end{bmatrix}, \] both orthogonal to \(u_1\) (and to each other). Then \[ A = U\Sigma V^T = \begin{bmatrix} 1/\sqrt{6} & 2/\sqrt{5} & -1/\sqrt{30} \\ -2/\sqrt{6} & 1/\sqrt{5} & 2/\sqrt{30} \\ 1/\sqrt{6} & 0 & 5/\sqrt{30} \end{bmatrix} \begin{bmatrix} \sqrt{30} & 0 \\ 0 & 0 \\ 0 & 0 \end{bmatrix} \frac{1}{\sqrt{5}} \begin{bmatrix} 1 & 2 \\ 2 & -1 \end{bmatrix}. \] (The columns of \(U\) for the zero singular value are not unique; any orthonormal completion works.)Keywords: singular value decomposition, rank one matrix, singular values, orthonormal basis