Any covariance matrix is positive semi-definite: A proof

Let X=(X1,X2,,Xn)Rn be a vector of random variables. The covariance matrix Σ of X is a square (n×n) matrix whose elements are covariances between the components of X. That is,

Σij=Cov(Xi,Xj)

where Cov(Xi,Xj) is the covariance between Xi and Xj, i,j=1,2,,n

Cov(Xi,Xj)=E[(XiE[Xi])(XjE[Xj])].

Here, E[] indicates the expectation value of a random variable. Any covariance matrix has the following properties:

  1. Symmetric. That is, Σ=Σ.
  2. Positive semi-definite. That is,vRn,vΣv0.

See also: Positive definite matrix (Wolfram MathWorld)

The symmetry is obvious from the definition of the covariance matrix. 

Now, let us prove that the covariance matrix is positive semi-definite. First, note that the covariance matrix can be expressed as
Σ=Rnρ(x)(xμ)(xμ)dx
where ρ(x) is the density of X, and
μ=E[X]=Rnρ(x)xdxRn
is the expectation value of X

Let vRn be an arbitrary vector. We have
vΣv=Rnρ(x)v(xμ)(xμ)vdx=Rnρ(x)[(xμ)v][(xμ)v]dx=Rnρ(x)[(xμ)v]2dx0
because the integrand is non-negative.

Remark (2022-10-14): In the initial version of this post, I used the eigenvalue decomposition of the covariance matrix to prove the positive semi-definiteness. But that is not only unnecessary but overly complicated. The above proof is simpler and more direct.

Comments

Popular posts from this blog

Birth process

Branching processes: Mean and variance

Informal introduction to formal logic