Math 104 Fall 2019

Homework 6
Due Thursday, November 14 (before class)

Homework should be submitted via gradescope.

1. Let U ∈ Rm×m and V ∈ Rn×n be othogonal matrices. Show that for any A ∈ Rm×n , kU AV kF = kAkF
and kU AV k2 = kAk2 . Is it true that that we also always have kU AV k1 = kAk1 ? Why or why not?
2. Let A ∈ Rm×n and recall that the p-norm of A is defined as

kAkp = max kAxkp .

kxkp =1

(a) Show that

m
X
kAk1 = max cj , cj = |aij |.
1≤j≤n
i=1

(b) Show that

n
X
kAk∞ = max ri , ri = |aij |.
1≤i≤m
j=1

3. Very simple linear regression. Consider the three points

     
0 1 2
, , .
1 −1 −2

What is the equation of the line that best fits those three points, in the sense of least squares?
4. The least-squares solution minimizes kb − Axk2 . In some applications, we would like to favor ‘low-
energy’ solutions and minimize instead

kb − Axk2 + λkxk2 , (1)

where λ is a positive parameter chosen by the scientist/engineer.

(a) Find the solution to this problem. [Hint: The SVD of A might be quite handy here.]
(b) Suppose you have a routine that returns the SVD of any matrix A; that is, it returns U , Σ and
V . Can you use it to compute the solution? If yes, explain how.
(c) Suppose you have a routine that returns the solution to any least-squares problem; that is, it
takes as input an arbitrary pair A, b and returns an x minimizing kb − Axk2 . Can you use it to
compute the solution? If yes, explain how.