Solve Least Squares Problem for Matrix A and B | Homework Equations

Click For Summary

Homework Help Overview

The problem involves finding the vector x that minimizes the expression ||Ax - b||2, where A is a given matrix and b is a vector. The context is within linear algebra, specifically dealing with least squares problems.

Discussion Character

  • Exploratory, Assumption checking, Mathematical reasoning

Approaches and Questions Raised

  • Participants discuss the invertibility of the matrix A^TA and the implications of A having linearly dependent columns. There are attempts to use row reduction to solve the matrix equation derived from the least squares formulation.

Discussion Status

Some participants have provided guidance on using row reduction to approach the problem, while others have noted the presence of free variables in the solution, indicating multiple solutions may exist. The discussion is ongoing, with various interpretations being explored.

Contextual Notes

There is a mention of the original poster encountering issues with the inverse of A^TA being zero, and the implications of this on the problem setup are being questioned. The participants are also considering the structure of the augmented matrix formed during row reduction.

chuy52506
Messages
77
Reaction score
0

Homework Statement


Let
A=
|2 -1 -1|
|-1 2 -1|
|-1 -1 2|
and
B=
|1|
|2|
|3|


Homework Equations



Find the x in which minimizes ||Ax-b||2


The Attempt at a Solution


I tried to solve it by using this formula (A**A)-1A**b=x but i get the inverse of A*A equal 0
 
Physics news on Phys.org
Greetings! Right, since A does not have linearly independent columns, [tex]A^TA[/tex] is not invertible. Call [tex]B = A^TA[/tex] and [tex]\vec{y} = A^T\vec{b}[/tex] and try using row reduction to solve the matrix equation [tex]B\vec{x} = \vec{y}[/tex] for [tex]\vec{x}[/tex].
 
When i try to solve it i get the last row in rref of B=A*A to be a row of 0's equal to 3=/
Is there any other way to solve this??
 
I got that

[tex]B = A^TA = \[ \left[ \begin{array}{ccc}<br /> 6 & -3 & -3 \\<br /> -3 & 6 & -3 \\<br /> -3 & -3 & 6 \end{array} \right]\][/tex]

and

[tex]\vec{y} = A^T\vec{b} = \[ \left[ \begin{array}{c}-3 \\<br /> 0 \\3 \end{array} \right]\][/tex].

Is this what you got? Then we can row reduce

[tex][ B\ \vec{y} ] = \[ \left[ \begin{array}{cccc}<br /> 6 & -3 & -3 & -3 \\<br /> -3 & 6 & -3 & 0 \\<br /> -3 & -3 & 6 & 3 \end{array} \right]\][/tex]

In fact, reducing this shows that there is a free variable, meaning that there is a whole line worth of solutions that give the best approximation.
 
thanks!
 

Similar threads

  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 2 ·
Replies
2
Views
1K
Replies
9
Views
2K
Replies
12
Views
2K
  • · Replies 1 ·
Replies
1
Views
1K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 11 ·
Replies
11
Views
3K
Replies
2
Views
2K
  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 25 ·
Replies
25
Views
4K