- #1
alyflex
- 9
- 0
I have stumbled upon a problem which I have so far been unable to solve.
I we consider a general set of linear equations:
[tex]Ax=b[/tex],
I know the the system is inconsistent which makes least square method the logical choice.
So the mission is to minimize [tex]||Ax-b||[/tex]
And the usual way I do this is by setting Ax=p, where p is the projection of b onto Ax.
By isolating:
[tex]x=(A^TA)^{-1}A^T \cdot p
=(A^TA)^{-1} \cdot A^T \cdot b[/tex]
However the product [tex]A^T*A[/tex] is also singular, and thus I am unable to do this.
I pretty sure there is a very simple way to do this, but when i look in my old algebra book I see no solution to the problem.
Anyone know the way?
I we consider a general set of linear equations:
[tex]Ax=b[/tex],
I know the the system is inconsistent which makes least square method the logical choice.
So the mission is to minimize [tex]||Ax-b||[/tex]
And the usual way I do this is by setting Ax=p, where p is the projection of b onto Ax.
By isolating:
[tex]x=(A^TA)^{-1}A^T \cdot p
=(A^TA)^{-1} \cdot A^T \cdot b[/tex]
However the product [tex]A^T*A[/tex] is also singular, and thus I am unable to do this.
I pretty sure there is a very simple way to do this, but when i look in my old algebra book I see no solution to the problem.
Anyone know the way?