Finding vectors orthogonal to the span of a matrix

In summary, To extend a reduced QR factorization to a complete one, you need to add m-n additional orthogonal vectors to Q. One method to find these vectors is by taking the nullspace of the transpose of A. Another possibility is to take the cross product of the first two columns if extending from a 3x2 to a 3x3. However, this may be time consuming for a general case. Another approach is to look at the null space of the orthogonal projector onto the range of A, also known as the pseudo inverse.
  • #1
Fractal20
74
1

Homework Statement


My linear algebra is rusty. So to go from a reduced QR factorization to a complete QR factorization (ie the factorization of an over determined matrix) one has to add m-n additional orthogonal vectors to Q. I am unsure on how to find these.

If it is extending a 3x2 to a 3x3 I know I can take the cross product of the first two columns etc. but I am unsure of a process to do this in a general case.

It seems like one possibility would be to look at the null space of the orthogonal projector onto the range of A (the pseudo inverse is it called)? But that seems time consuming. It there a quick method to find a vector orthogonal to a set of orthonormal vectors? Thanks y'all!***edit
Or, I just realized, guess I could find the nullspace of the transpose of A. Is there a particular approach anybody would suggest?

Homework Equations


The Attempt at a Solution

 
Physics news on Phys.org
  • #2
I think I may have realized the answer. I will still post this in case it helps somebody else. To find the additional orthogonal vectors to extend a reduced QR factorization to a complete one you can take the nullspace of the transpose of A.
 

Related to Finding vectors orthogonal to the span of a matrix

1. How do you determine the vectors orthogonal to the span of a matrix?

The vectors orthogonal to the span of a matrix can be found by using the Gram-Schmidt process. This involves taking the given vectors and using orthogonal projections to create new vectors that are orthogonal to the previous ones.

2. Why is it important to find vectors orthogonal to the span of a matrix?

Finding vectors orthogonal to the span of a matrix allows for a basis to be formed for the null space or solution space of the matrix. This information is useful in solving systems of linear equations and in understanding the properties of the matrix.

3. Can there be more than one set of vectors orthogonal to the span of a matrix?

Yes, there can be multiple sets of vectors that are orthogonal to the span of a matrix. This is because there are infinite ways to choose the initial vectors and perform the Gram-Schmidt process.

4. How does finding vectors orthogonal to the span of a matrix relate to linear independence?

If the vectors orthogonal to the span of a matrix are linearly independent, then they form a basis for the null space of the matrix. This means that the vectors in the original matrix are also linearly independent.

5. Is it possible for a matrix to have no vectors orthogonal to its span?

Yes, it is possible for a matrix to have no vectors orthogonal to its span. This can occur when the matrix has a rank equal to its number of columns, meaning that there are no free variables and the null space is empty.

Similar threads

Replies
1
Views
1K
Replies
11
Views
3K
Replies
1
Views
1K
Replies
3
Views
2K
Replies
29
Views
2K
Replies
2
Views
1K
Back
Top