, , , ,

[SOLVED] CMPUT 466 566 Machine learning Python

$25

File Name: CMPUT_466_566_Machine_learning_Python.zip
File Size: 348.54 KB

5/5 - (1 vote)

CMPUT 466 566

Machine learning

Problem 1.

Let  be the linear hypothesis class.

Prove that for, any  in  their linear combination

is also in  where 

Problem 2.

Consider a two dimensional space ℝ2 . Determine whether the following sets are convex or not. Prove or disprove.

Problem 3.

Consider the function 

a) View x1as a variable and x2  as a constant. Determine whether f is convex in x1  and prove it.

b) View x2as a variable and x1  as a constant. Determine whether f is convex in x2  and prove it.

c)  View f: ℝ2  → ℝ as a function of the input vector (x1, x2) . Determine whether f is convex in (x1, x2) and prove it.

Hints: For a) and b), treat one variable as a constant, and calculate the second-order derivative of a single-variable function.

For c), calculate the Hessian matrix H first and choose a point, say, (0,0). You may use numpy in Python to calculate the eigenvalue

import numpy as np

from numpy import linalg as LA

H = np.array ( [ [11, 12], [21, 22]])  # your values here

eigenval, eigenvec = LA.eig(H)

Print eigenval. If any number is less than 0, then the function is not convex. Otherwise, it is convex. Eigenvalues may also be calculated manually.

The example shows that an element-wise convex function may not be jointly convex.

 

Shopping Cart

No products in the cart.

No products in the cart.

[SOLVED] CMPUT 466 566 Machine learning Python
$25