- #1
jollage
- 63
- 0
Hi all,
Suppose I have a matrix [itex] A_{N\times N}[/itex]. I compute its eigenmodes
[itex] A V = V \Lambda[/itex].
[itex] V, \Lambda[/itex] are eigenvectors and eigenvalues of size [itex] N\times N [/itex]. The eigenvalues are descending.
Now I cut off several eigenmodes (the ones having small value), it becomes
[itex] A V_{N \times n} = V_{N \times n} \Lambda_{n \times n}[/itex].
What I want is to reconstruct [itex]A[/itex] from [itex] V_{N \times n}, \Lambda_{n \times n}[/itex].
It seems that in Matlab, if I just modify the above equation
[itex] A = V_{N \times n} \Lambda_{n \times n}V^{-1}_{N \times n}[/itex],
it will not work.
So my question is: how can I reliably reconstruct the original matrix by using its reduced eigenmodes? Thanks!
Suppose I have a matrix [itex] A_{N\times N}[/itex]. I compute its eigenmodes
[itex] A V = V \Lambda[/itex].
[itex] V, \Lambda[/itex] are eigenvectors and eigenvalues of size [itex] N\times N [/itex]. The eigenvalues are descending.
Now I cut off several eigenmodes (the ones having small value), it becomes
[itex] A V_{N \times n} = V_{N \times n} \Lambda_{n \times n}[/itex].
What I want is to reconstruct [itex]A[/itex] from [itex] V_{N \times n}, \Lambda_{n \times n}[/itex].
It seems that in Matlab, if I just modify the above equation
[itex] A = V_{N \times n} \Lambda_{n \times n}V^{-1}_{N \times n}[/itex],
it will not work.
So my question is: how can I reliably reconstruct the original matrix by using its reduced eigenmodes? Thanks!
Last edited: