Results are always returned with signs mismatched from the correct answers. They are not always opposite signs and I cannot find a pattern to the error..

Anyone know if there is a mode I have to set or something?

For example, eigVc([1,-1;2,4]) returns [-.707, .707; .447, -.894]....SHOULD be [-.707, .707; -.447, .894]

4/21/2010 1:15:01 AM

example above has row/column swapped: should be this

For example, eigVc([1,-1;2,4]) returns [-.707, .447; .707, -.894]....SHOULD be [-.707, -.447; .707, .894]

4/21/2010 2:04:17 AM

http://www.bluebit.gr/matrix-calculator/

maybe you're doing it wrong

4/21/2010 1:14:15 PM

This is what I did:

To get eigVec command

MATH > 4: Matrix > A:eigVec

At the command line:

> eigVec([1,-1;2,4])

Returns
-.447214 , -.707107
.894427 , .707107

or in the form you have
[ -.447214 , -.707107; .894427 , .707107]

[Edited on April 21, 2010 at 9:38 PM. Reason : spaces]

[Edited on April 21, 2010 at 9:39 PM. Reason : spacing]

4/21/2010 9:37:45 PM

Follow up on this:

Tested on TI-89, returns correct answer
Tested on TI-89 titanium, returns incorrect swapped sign answer given above

go fucking figure



[Edited on April 21, 2010 at 10:10 PM. Reason : d]

4/21/2010 10:03:00 PM

All of the eigenvectors you are getting are correct. They're just normalized so their vector lengths are 1.

[-.707 , .707] = -[.707 , -.707]

are just scalar multiples of each other by -1.

If you multiply by an appropriate scalar in each case you'll get integers in the entries:
[-1,1] and [-1,2]

The lesson here is that eigenvectors are only unique up to scalar multiplication. In other words, if:
A*X = lambda*X

Then:
A*(bX) = bA*X = b*lambda*X = lambda*(bX)

where A is the matrix, lambda is the eigenvalue, X is an eigenvector, and b is a given scalar.

4/22/2010 7:48:30 PM

^

thank you

4/22/2010 8:31:22 PM