Jama Eigenvectors

["Faiyaz Doctor" <fdocto@essex.ac.uk>]

I'm a research student and I'm trying to use Jama to calculate eigenvectors from cluster covariance matrices in order to correctly project the fuzzy membership functions from the clusters.

 

 In comparing jama with jampack and matlab for a specific examples I'm unsure why I'm getting different results

 

1

 

given a 2x2 matrix

 

-28   24

-40   34

 

then the eigenvectors given by matlab and Jampack are:

 

-0.6247         -0.6000

 

-0.7809         -8.0000

 

while Jama gives me:

 

-0.8       0.60000..

25.000    -20.0000..

 

why is this the case??

 

Also given another example matrix:

 

0   -6     -1

6    2     -16

-5   20    -10

 

I'm getting different results between Jama and Matlab and JamPack, why??

 

I'd appreciate any input on this,

 

Thanks,

 

Faiyaz