Skip to content

4.2-exam2-2002

Quiz 2, 2002 · Section 4.2: Projections onto Lines and Subspaces Source: exam2_1.pdf, pp. 69, 75

Problem

1 (40 pts.) (a) Find the projection matrix \(P_C\) onto the column space of \(A\) (after looking closely at the matrix!)

\[ A = \begin{bmatrix} 3 & 3 & 6 \\ 1 & 1 & 2 \end{bmatrix} \]

(b) Find the 3 by 3 projection matrix \(P_R\) onto the row space of \(A\). What is the closest vector in the row space to the vector \(b = (1, 0, 0)\)?

(c) Multiply \(P_C A\) and then \(P_C A P_R\). Your answers should be a little surprising—can you explain?

(d) Find a basis for the subspace of all vectors orthogonal to the row space of \(A\).

Solution 1a. The columns of \(A\) are linearly dependent, but the column space is spanned by \(\begin{bmatrix} 3 \\ 1 \end{bmatrix}\). Use this matrix in the formula for the projection matrix: \[ P_C = A(A^T A)^{-1} A^T = \begin{bmatrix} 3 \\ 1 \end{bmatrix} \frac{1}{10} \begin{bmatrix} 3 & 1 \end{bmatrix} = \frac{1}{10} \begin{bmatrix} 9 & 3 \\ 3 & 1 \end{bmatrix}. \] 1b. As before but with \(A = \begin{bmatrix} 1 \\ 1 \\ 2 \end{bmatrix}\): \[ P_R = A(A^T A)^{-1} A^T = \begin{bmatrix} 1 \\ 1 \\ 2 \end{bmatrix} \frac{1}{6} \begin{bmatrix} 1 & 1 & 2 \end{bmatrix} = \frac{1}{6} \begin{bmatrix} 1 & 1 & 2 \\ 1 & 1 & 2 \\ 2 & 2 & 4 \end{bmatrix}. \] The vector in the row space of \(A\) closest to \(b = \begin{bmatrix} 1 & 0 & 0 \end{bmatrix}^T\) is \[ P_R b = \frac{1}{6} \begin{bmatrix} 1 \\ 1 \\ 2 \end{bmatrix}. \] 1c. \[ P_C A = \frac{1}{10} \begin{bmatrix} 9 & 3 \\ 3 & 1 \end{bmatrix} \begin{bmatrix} 3 & 3 & 6 \\ 1 & 1 & 2 \end{bmatrix} = \begin{bmatrix} 3 & 3 & 6 \\ 1 & 1 & 2 \end{bmatrix} = A \] \[ P_C A P_R = \begin{bmatrix} 3 & 3 & 6 \\ 1 & 1 & 2 \end{bmatrix} \frac{1}{6} \begin{bmatrix} 1 & 1 & 2 \\ 1 & 1 & 2 \\ 2 & 2 & 4 \end{bmatrix} = \begin{bmatrix} 3 & 3 & 6 \\ 1 & 1 & 2 \end{bmatrix} = A. \] The two multiplications project the columns/rows of \(A\) onto the column/row space of \(A\). This does not change the matrix, since the columns of \(A\) are already in its column space and the rows are already in its row space. 1d. All vectors orthogonal to the row space of \(A\) are in the nullspace of \(A\). A basis for the nullspace of \(A\) is (note dimension 2): \[ \begin{bmatrix} -1 \\ 1 \\ 0 \end{bmatrix}, \begin{bmatrix} -2 \\ 0 \\ 1 \end{bmatrix}. \]

Keywords: projection matrix, column space, row space, nullspace