Skip to content

4.2-final-2005

Final Exam, 2005 · Section 4.2: Projections onto Lines and Subspaces Source: finalexam_1.pdf, p. 112

Problem

2 (12 pts.) Keep the same matrix \(A\) as in Problem 1.

(a) Find the matrix \(P\) that projects any vector in \(R^4\) onto the subspace spanned by the first column of \(A\).

(b) Describe the nullspace of \(I - P\) and the nullspace of \(PA\).

(c) Find all the eigenvalues of \(P\).

Solution (a) \[ P = \frac{aa^T}{a^Ta} = \frac{1}{7} \begin{bmatrix} 4 & 2 & 2 & 2 \\ 2 & 1 & 1 & 1 \\ 2 & 1 & 1 & 1 \\ 2 & 1 & 1 & 1 \end{bmatrix}. \] Trick: computation easiest using \(a = (2, 1, 1, 1)^T\), the first column of \(A\). (b) The nullspace of \(I - P\) consists of all multiples of \(a\). (One view is that \(x = Px\). Another view is that it is the orthogonal complement of the nullspace of \(P\), which consists of all vectors orthogonal to \(a\).) The official solution does not give the nullspace of \(PA\); since \(PAx = \frac{a}{7} a^TAx\) and \(a \neq 0\), we have \(PAx = 0\) exactly when \(a^TAx = 0\), i.e. \[ (2x_1 + x_2 + x_3 + x_4) + 5(x_1 + x_2 + x_3 + x_4) = 7x_1 + 6x_2 + 6x_3 + 6x_4 = 0. \] Thus \(N(PA) = \{x : 7x_1 + 6x_2 + 6x_3 + 6x_4 = 0\}\), a 3-dimensional subspace of \(R^4\). (c) The matrix has rank 1 and is a projector, so one eigenvalue is 1 and the rest are 0.

Keywords: projection, nullspace, eigenvalues, rank one