以下第11題到第13題中之所有的提問,都不需要高出推導過程,只要高出答案即可,答案正確 就得分。
11. (10%) Let
.
(b) (5%) When the equation Ax = b is unsolvable, we may consider the so-called least squares problen to find a set of solutions, having the least squares error, fom solving a normal equation. Suppose that rank(A) =k <min(m, n) and let A = BC be a full rank decomposition of A . Use the known matrices B, C, and b to describe the unique projection vector p of b onto R(A) with the least || b-pll2.