Problem with SVD Decomposition


I am trying to diagonalize some matrix with the SVD method using TDecompSVD class.

I took the following 2x2 matrix :
A =
16 -35
6 -13

The eigen values are 2 and 1, and I try to solve it with TDecompSVD :

TMatrixD m(2,2);
m(0,0)=16; m(0,1)=-35; 
m(1,0)=6; m(1,1)=-13;
TDecompSVD svd(m);
TVectorD S = svd.GetSig();

I obtained :

Vector (2)  is as follows

     |        1  |
   0 |41.0609 
   1 |0.0487081 

The eigen values are incoherent…
I tried it with symmetric matrices and it worked for all of them.

What’s happenning ??


Confusing eigen values and singular values ?

Have a look at … omposition

specifically section “Relation to eigenvalue decomposition”

  • Eddy

You may be really right…
Thanks for helping !