deepdream of
          a sidewalk
Show Question
Math and science::Algebra

Spectral Theorem

Spectral theorem

Every real symmetric matrix S can be expressed as

S=QΛQT

There are two arguments on the reverse that motivate this statement.


Both of the below arguments don't cover the existance of the eigenvectors.

Motivation 1: eigenvectors are in the column space and row space

Consider two eigenvector-value pairs (v1,λ1) and (v2,λ2).

Firstly, observe that for an eigenvector-value pair (v1,λ1), the defining statement:

Sv1=λ1v1

also gives us the property: (SαI)v1=(λ1α)v1

Which implies that v1 is in the column space of SαI as long as λ1α. When λ1=α, the eigenvector is in the null space of SαI. Being in the null space implies that the eigenvector is orthogonal to all rows of SαI.

When v1 is in the null space of Sλ1I, v2 is in the column space, and as S is symmetric, v2 is also in the row space of Sλ1I. So this leads us to conclude that v1 is orthogonal to v2.

This argument extends to more eigenvectors, and we can conclude that a symmetric matrix can be diagonalized by orthogonal eigenvectors.

Motivation 2: SVD of a symmetric matrix

If we assume that symmetric matrix S can be decomposed via SVD as:

S=UΣVT

where U and V are orthonormal matrices with the same dimensionality as S.

Then:

ST=VΣUT

Equating these two expressions, we get:

UΣVT=VΣUT

TODO: how to justify this next jump?

This equality suggests that U=V. And SVD is instead an eigenvector decomposition:

S=UΣUT