- #1
vibe3
- 46
- 1
I am solving linear least squares problems with generalized Tikhonov regularization, minimizing the function:
[tex]
\chi^2 = || b - A x ||^2 + \lambda^2 || L x ||
[/tex]
where [itex]L[/itex] is a diagonal regularization matrix and [itex]\lambda[/itex] is the regularization parameter. I am solving this system using the singular value decomposition, and I use the normal trick to convert the system into "Tikhonov standard form":
[tex]
\tilde{A} = A L^{-1}, \tilde{x} = L x
[/tex]
so that
[tex]
\chi^2 = || b - \tilde{A} \tilde{x} ||^2 + \lambda^2 || \tilde{x} ||
[/tex]
This system can now be solved easily with the SVD of [itex]\tilde{A}[/itex]. My problem is that I want to compute the condition number of [itex]A[/itex] while solving the system, but I don't see any easy way to compute this since I calculate only the SVD of [itex]\tilde{A}[/itex]. It seems that there is no simple relationship between the singular values of [itex]A[/itex] and the singular values of [itex]\tilde{A}[/itex], even when [itex]L[/itex] is a diagonal matrix.
Does anyone know of a solution to this problem, without having to separately compute the SVD of [itex]A[/itex] itself?
[tex]
\chi^2 = || b - A x ||^2 + \lambda^2 || L x ||
[/tex]
where [itex]L[/itex] is a diagonal regularization matrix and [itex]\lambda[/itex] is the regularization parameter. I am solving this system using the singular value decomposition, and I use the normal trick to convert the system into "Tikhonov standard form":
[tex]
\tilde{A} = A L^{-1}, \tilde{x} = L x
[/tex]
so that
[tex]
\chi^2 = || b - \tilde{A} \tilde{x} ||^2 + \lambda^2 || \tilde{x} ||
[/tex]
This system can now be solved easily with the SVD of [itex]\tilde{A}[/itex]. My problem is that I want to compute the condition number of [itex]A[/itex] while solving the system, but I don't see any easy way to compute this since I calculate only the SVD of [itex]\tilde{A}[/itex]. It seems that there is no simple relationship between the singular values of [itex]A[/itex] and the singular values of [itex]\tilde{A}[/itex], even when [itex]L[/itex] is a diagonal matrix.
Does anyone know of a solution to this problem, without having to separately compute the SVD of [itex]A[/itex] itself?