JAMA Eigenvectors Inconsistent with those of Matlab


Hi
 
Firstly I would like to say what a great tool JAMA is - thanks.
 
I'm using the JAMA package as part of of a multivariant data tool. Whilst performing some tests on the program, I noticed a difference between the eigenvectors given by JAMA and Matlab 5.2 for the following test matrix:
 
                                1     0     4
                                0     2     0
                                3     1    -3
 
The eigenvalues are the same, however, JAMA gave the following eigenvectors:
 
                                0.8944    -0.5590    -0.5714
                                0             0             1
                                0.4472     0.8385    -0.1429
 
whilst Matlab gave these:
 
                                0.8944    -0.5547    -0.4924
                                0             0             0.8616
                                0.4472     0.8321    -0.1231
 
Any thoughts or advise would be welcome.
 
Thanking you in advance
 
Russell Hill
 
 
 


Date Index | Thread Index | Problems or questions? Contact list-master@nist.gov