| Author |
Comment/Response |
Tamas Huszak
|
 05/28/12 2:00pm
In Response To 'Re: Weighted least squares'
---------
Hi!
My code works so far without errors until this point. I'm only looking for a simple way to program the wighted LSQ problem.
So far I came up with:
Inverse[Transpose[A].W.A].Transpose[A].W.b
where: W is the weights matrix
but this does not produce a result, because my matrix is badly conditioned, so I tried
PseudoInverse instead of Inverse
but got the problem, that when using a simple IdentityMatrix for W, which is supposed to give the same result as LeastSquares[A,b], but it doesent.
Maybe my math is wrong, or the PseudoInverse cuts off too many values I don't know.
I've attached both the notebook and the data as well.
Thomas
Attachment: WLSQ.zip, URL: , |
|
