You are on page 1of 1

Foundations of data science, 2nd midterm RETAKE, December 16, 2021

Problem 1
The norms in this problem denote the matrix / operator norms induced by the corresponding
vector norm.
a 1
[10%] Find a ∈ ℝ such that   has the smallest value.
2 3 ∞
a 1
[15%] Find a ∈ ℝ such that   has the smallest value.
3 a 2

Problem 2 [15%]
Let us consider the real vector space V := C([0, 1]), that is, V = { f : [0, 1] →  f is continuous } with
the usual operations. The functional ·∞ : V → ℝ is defined by f∞ := 2 maxx∈[0,1] f(x) (for any
f ∈ V). Show that ·∞ satisfies all norm axioms.

Problem 3
u1 v1
Let us consider the vector space V := 3 with the usual operations. For any u = u2 and v = v2 ,
u3 v3
3 2
we define 〈u, v〉 := ∑k=1 k uk vk .
[10%] Show that the above formula determines an inner product.
1 1 1
[10%] Show that the vectors a := 0 , b := 1 , and c := 1 are linearly independent, then apply
0 0 1
the Gram–Schmidt process to them.

Problem 4
1 2
[25%] Find the SVD of the matrix A :=  .
3 6
[5%] Then, by using the SVD, find its Moore–Penrose pseudoinverse A+ .
2
[10%] From this, compute P := A A+ . Finally, evaluate P b with b :=  , and show that P b is indeed
4
the orthogonal projection of the vector b onto the subspace spanned by the column vectors of the
matrix A.

You might also like