An orthogonal matrix from two orthognal matrices!

5 vues (au cours des 30 derniers jours)
wissam
wissam le 12 Mar 2012
Dear All.. I have two orthogonal matrices, A and B, where
A is (256 by 256) elements, A*A'=I; and B is (256 by 256) elements, B*B'=I.
Let C is a matrix formed by a combination of A and B such that:
C is (256 by 256) and C(1:128,:)=A(1:128,:); C(129:256,:)=B(1:128,:);
Unfortunately, C is not orthogonal, i.e. C*C~=I; so, how can i do some changes on C to make it orthogonal, let's say C1 is the orthogonal matrix extracted from C. I need C1 to be in the closest form (such as mesh plots) to C?
is there any method for that? looking forward to get your help best regards sam..

Réponse acceptée

Jan
Jan le 12 Mar 2012
The question is not clear. Your matrix C is not guaranteed to have full rank. Then a procedure to "make it orthonormal" is not straight. I do not understand the description "closest form as mesh plots".
  1 commentaire
wissam
wissam le 13 Mar 2012
hi jan,
thanks for the reply..
I just have a matrix C and wanna make it an orthogonal one.. any procedure that i can use such as SVD decomposition to extract an orthogonal matrix from C?
okay, you ignore what i have said.. "closest form as mesh plots"..
best regards..

Connectez-vous pour commenter.

Plus de réponses (2)

G A
G A le 12 Mar 2012
will this help?
doc orth
  1 commentaire
wissam
wissam le 13 Mar 2012
hi GA,
thanks for your reply..
when i try to use: b=orth(C), this gives me a matrix b and its orthogonal but when i do mesh plot it, the plot is like random numbers not similar to C shape that what i want.. i want ot generate b is similar (in the closet plot)to C but orthogonal..
thanks..

Connectez-vous pour commenter.


Bjorn Gustavsson
Bjorn Gustavsson le 13 Mar 2012
It doesn't necessarily work that way that the matrix (with vectors) spanning the same space as the columns of your matrix C is similar to C in a spatial sense. You might be able to resort the columns of C1 to make it looking nicer (at least different). But this is the way it is. You might have to learn to look at the new base-vectors of C1 one by one (or a few at a time).

Catégories

En savoir plus sur Matrix Decomposition dans Help Center et File Exchange

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by