Introduction Matrix Action Perpframes, Aligners and Hangers Stretchers Coordinates Projections SVD Matrix Subspaces Linear Systems, Pseudo-Inverse Condition Number Matrix Norm, Rank One Data Compression Noise Filtering
Todd Will
|
Fundamental
Equations Orthonormal Bases |
Subspaces |
Bonus
Facts
Exercises |
Matrix SubspacesIn the last section you battled to show that every matrix A could be written as A=(hanger)(stretcher)(aligner).What's the pay off?
Fundamental equations2x2 ExampleHere is an SVD of a 2 x 2 matrix :Now watch what the matrix A does to an ellipse lined up on the perpframe . The matrix stretches the ellipse and transfers it from the
perpframe to the
perpframe.
If you look carefully at the "during" plot you'll see that A
sends
to
and
to .
More generally A,
You can verify this by computing what each factor of A does. Check what each factor of A does to :
(The aligner matrix brings to the y-axis.) (The stretcher matrix stretches the y-axis by a factor of .) (The hanger rotates the y-axis to the direction of .)
Theorem:If you build a 2x2 matrix A=(hanger)(stretcher)(aligner) using
These two equations are the fundamental equations
of matrices of the form (hanger)(stretcher)(aligner).
In words they say that A
Proof: Outlined above--follow the action of each factor of A on .
3 x 3 matrices.The following plot shows the perpframe in blue and the perpframe in red.See what the matrix does to the surface lined up on . Just as in the 2x2 case, the matrix A obeys the fundamental equations: In fact, matrices of the form (hanger)(stretcher)(aligner) always obey these fundamental equations, regardless of their dimension.
Orthonormal bases for fundamental subspaces.Suppose the SVD for a matrix is .This decomposition presents orthonormal bases for
Proofs:The Column space of A, Col[A], is the span of the columns of A.
(The column way to multiply A v.)
(Since the a's are a basis.)
(Linearity)
(Fundamental Equations)
. This shows shows that Col[A] = span. Since are perpendicular unit vectors they are an orthonormal basis for Col[A].
The nullspace of A, N[A], is the set of vector that A sends to the zero
vector.
Suppose .
Then
(Linearity and Fundamental Equations.)
. Now,
tells you that
and so .
This tells you that any vector in N[A] is a linear combination of and , i.e. . On the other hand, ,
so .
The row space of A is the span of the rows of A, which is the same as the column space of . So finding a basis for the row space of A is the same as finding a basis
for the columns space of .
Here's a look at :
Now you have written as (hanger)(stretcher)(aligner),
i.e., you have an SVD of .
So from what you did above you know
CommentsThe proofs above easily generalize to show that
Bonus Facts.Definitions:The rank of A is usually defined as the dimension of the column space.Since the columns of the hanger matrix corresponding to non-zero singular
values form a basis for the column space, you know that the rank of A is
equal to the number of non-zero singular values.
Similarly, since the nullity of A is the dimension of the null space, the number of zero singular values equals the nullity of A.
Theorem: rank[A]+nullity[A]Proof:rank[A]+nullity[A] =(number of non-zero singular values)+(number of zero singular values) =(total number of singular values) =n.
Theorem:Proof:Rank[A] =(number on non-zero singular values of A) (number of non-zero singular values of ) = . (1) Since the stretcher matrix for
is the transpose of the stretcher matrix for A, the singular values
for A and
are identical.
Exercises1. Go with the 3x4 matrix . Based on the SVD of A:
give orthonormal bases for a. the column space of A b. the row space of A c. the nullspace of A d. the nullspace of .
2. Let . Here is an SVD for A.
Based on the SVD, give an orthonormal basis for the subspace spanned by 3. Let . Here is an SVD for A.
Based on the SVD of A and the fundamental equations, compute a. b. If you check your answer by using , you may not get agreement since the SVD for A has been rounded to two decimals. |