4.2-exam2-2013¶
Quiz 2, 2013 · Section 4.2: Projections onto Lines and Subspaces Source: exam2_2.pdf, pp. 52-53, 60
Problem¶
2 (30 pts.)
Let \(Q = [q_1 \ q_2 \ q_3]\) be an \(m \times 3\) real matrix with \(m > 3\) and \(Q^TQ = I_3\), the \(3 \times 3\) identity.
Let \(P = QQ^T\).
a) (7 pts.) What are all possible values of \(\det(P)\)?
b) (7 pts.) What are all the eigenvalues of the \(m \times m\) matrix \(P\) including multiplicities?
c) (8 pts.) Find one eigenvalue, eigenvector pair of the non-symmetric \(m \times m\) matrix \(q_1q_2^T\).
d) (8 pts.) What are the four fundamental subspaces of \(M = I - P\) in terms of the column space of \(P\)?
Solution
a) \(P\) is the projection onto a 3-dimensional subspace of an \(m > 3\)-dimensional vector space, so \(\det(P) = 0\). b) The eigenvalues are 1 (multiplicity 3) and 0 (multiplicity \(m - 3\)). c) An eigenvector is \(q_1\) with eigenvalue \(q_2^Tq_1 = q_2 \cdot q_1\), since \((q_1q_2^T)q_1 = q_1(q_2^Tq_1) = (q_2^Tq_1)q_1\). d) \(P\) is symmetric, so \(M\) is symmetric. The left nullspace = right nullspace = the column space of \(P\), and the column space = row space = the orthogonal complement of the column space of \(P\).Keywords: projection matrix, eigenvalues, orthogonal columns, fundamental subspaces