MSBD 5004 Mathematical Methods for Data Analysis Homework 4
Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit
MSBD 5004 Mathematical Methods for Data Analysis
Homework 4
Due date: 11 November, Friday
1. Consider the function f : R2 → R given by f(x) = x1 + ex2 −x1 .
(a) Find the gradient of f .
(b) Find the Taylor approximation fˆ of f near the point z = (1, 2).
(c) Compare f(x) and fˆ(x) for the following values of x:
x = (1, 2), x = (0.96, 1.98), x = (1.10, 2.11), x = (0.85, 2.05), x = (1.25, 2.41). Make a brief comment about the accuracy of the Taylor approximation in each case.
2. Find the gradient of the following functions, where the space Rn and Rn ×n are equipped with the standard inner product.
(a) f(x) = ∥Ax − b∥2(2) + λ∥x∥2(2) , where A ∈ Rm ×n , b ∈ Rm , and λ > 0 are given. (b) f(X) = bT Xc, where X ∈ Rn ×n and b, c ∈ Rn .
(c) f(x) = xT Ax, where x ∈ Rn , and A ∈ Rn ×n is non-symmetric.
(d) f(X) = bT XT Xc, where X ∈ Rn ×n and b, c ∈ Rn . (e) f(X) = trace(XAXB), where X , A, B ∈ Rn ×n .
3. Consider the vector space l∞ equipped with the norm || · || ∞ . Define the operator T : l∞ → l∞ by T({xn }n∈N) = {yn }n∈N where yn = xn+1 .
(a) Prove that T is a linear operator.
(b) Prove that T is a bounded operator.
(c) Prove that ∥T∥ = 1.
4. Let V1 and V2 be two Hilbert spaces with the inner products ⟨ · , · ⟩ V1 and ⟨ · , · ⟩ V2 , respectively. Let T : V1 → V2 be a bounded linear operator, and let T ∗ be the adjoint operator of T, i.e., ⟨Tx, y⟩V2 = ⟨x,T∗ y⟩V1 for any x ∈ V1 and y ∈ V2 .
(a) Prove that T ∗ is a linear operator.
(b) Prove that (T∗ )∗ = T.
(c) Prove that ∥T∗ ∥ ≤ ∥T∥ .
5. Find the Jacobian matrix of the following vector-valued multi-variable functions.
(a) f : Rn → Rm is defined by f(x) = Ax − b, where x ∈ Rn , A ∈ Rm ×n , b ∈ Rm . (b) f : Rn → Rn is defined by f(x) = xxT a, where x ∈ Rn , a ∈ Rn .
6. Let f : R2 → R, g : R2 → R2 , g(x,y) = (x2 y,x − y) and h = f ◦ g = f(g(x,y)). Find h北 |x=1,y=2 if f北 |x=2,y= −1 = 3 and |x=2,y= −1 = −2. (Hint: use the chain rule)
2022-11-12