Skip to content

4.3-exam2-2007

Quiz 2, 2007 · Section 4.3: Least Squares Approximations Source: exam2_1.pdf, p. 144

Problem

2 (25 points)

(a) Compute the projection of the vector \(b = (1, 1, 1)\) onto the column space of

\[ A = \begin{bmatrix} 1 & 0 & 2 \\ 0 & 1 & 5 \\ -1 & 0 & -2 \end{bmatrix}. \]

(Hint: first check whether \(A\) has linearly independent columns.)

(b) Find the least-squares solution \(\hat{x}\) for the system

\[ \begin{bmatrix} 1 & 0 \\ 0 & 1 \\ -1 & 0 \end{bmatrix} x = \begin{bmatrix} 1 \\ 1 \\ 1 \end{bmatrix}. \]

(c) Find the projection of the vector \(b = (1, 1, 1)\) onto the column space of

\[ \begin{bmatrix} 10000 & 0 & 2 \\ 0 & 1 & 5 \\ -1 & 0 & -2 \end{bmatrix}. \]

(Hint: No computations!)

Solution 2a. The third column \((2, 5, -2)\) is in the span of the first two columns: \((2, 5, -2) = 2(1, 0, -1) + 5(0, 1, 0)\). So in order to calculate the projection we need to use the matrix \[ B = \begin{bmatrix} 1 & 0 \\ 0 & 1 \\ -1 & 0 \end{bmatrix}. \] The projection is \(p = B(B^T B)^{-1} B^T b = (0, 1, 0)^T\). 2b. \[ \hat{x} = (B^T B)^{-1} B^T b = (0, 1)^T. \] 2c. The columns of the \(3 \times 3\) matrix are linearly independent (the third column is not a combination of the first two), so its column space is all of \(\mathbb{R}^3\) and the projection of \((1, 1, 1)\) is itself: \(p = (1, 1, 1)^T\).

Keywords: projection, least squares, independent columns, column space