could anyone please help me with the following question?

Let $\Sigma$ be the covariance matrix of an $n$-dimesnional Gaussian distribution. The generic requirement for $\Sigma$ is to be symmetric and positive definite. However, what's happening when $\Sigma$ is diagonal with equal entries, i.e., $\Sigma=\text{diag}\{\sigma^2,\dots,\sigma^2\}$. I mean, how is it called? Is the "isotropic covariance matrix" a proper name for such a covariance matrix? If so, is the Gaussian distribution called an "isotropic Gaussian distribution"? If not, then what? Anyway, do the "isotropic {Gaussian distribution, covariance matrix}" exist?

Thanks in advance!


I've never heard it being called an isotropic covariance matrix, but that doesn't necessarily mean much. But since it's a Gaussian distribution, if the covariance matrix is diagonal then the variables are all independent. So then you can say, for example, that

The elements of $\mathbf{y}$ are independent with means $\boldsymbol{\mu}$ and equal variance $\sigma^2$.

| cite | improve this answer | |
  • $\begingroup$ Yeah, I know that, but I would like to know if there's something like that notation. Thanks anyway! $\endgroup$ – nullgeppetto Nov 22 '13 at 13:17

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.