Householder transformation
This entry describes the Householder transformation u=Hv, the most frequently used algorithm for performing QR decomposition. The key object here is the Householder matrix H, a symmetric
and orthogonal matrix
of the form
H=I-2xxT, |
where I is the identity matrix and we have used any normalized vector x with ||x||22=xTx=1.
The Householder transformation zeroes the last m-1 elements of a column vector below the first element:
[v1v2⋮vm]→[c0⋮0]withc=±||v||2=±(m∑i=1v2i)1/2 |
One can verify that
x=f[v1-cv2⋮vm]withf=1√2c(c-v1) |
fulfils xTx=1 and that with H=I-2xxT one obtains the vector [c0⋯0]T.
To perform the decomposition of the m×n matrix A=QR (with m≥n) we construct an m×m matrix H(1) to change the m-1 elements of the first column to zero. Similarly, an m-1×m-1 matrix G(2) will change the m-2 elements of the second column to zero. With G(2) we produce the m×m matrix
H(2)=[10⋯00⋮G(2)0]. |
After n such orthogonal transformations (n-1 times in the case that m=n), we let
R=H(n)⋯H(2)H(1)A. |
R is upper triangular and the orthogonal matrix Q becomes
Q=H(1)H(2)⋯H(n). |
In practice the H(i) are never explicitly computed.
References
-
•
Originally from The Data Analysis Briefbook (http://rkb.home.cern.ch/rkb/titleA.htmlhttp://rkb.home.cern.ch/rkb/titleA.html)
Title | Householder transformation |
---|---|
Canonical name | HouseholderTransformation |
Date of creation | 2013-03-22 12:06:07 |
Last modified on | 2013-03-22 12:06:07 |
Owner | mathcam (2727) |
Last modified by | mathcam (2727) |
Numerical id | 10 |
Author | mathcam (2727) |
Entry type | Algorithm |
Classification | msc 15A57 |
Classification | msc 65F25 |
Synonym | Householder reflection |
Synonym | Householder matrix |
Related topic | GramSchmidtOrthogonalization |