If $\Sigma$ is a real, symmetric, positive semidefinite matrix (a covariance matrix), then we can construct $M = \Sigma - \operatorname{diag}(\Sigma)$, where we essentially take the covariance matrix and set the diagonals to zero.

Is this matrix studied in any context, perhaps with a special name? For instance, one could ask whether there are relationships between eigendecomposition of $\Sigma$ and $M$. Obviously $M$ is still real, symmetric but no longer p.s.d.

The relationship between the graph Laplacian and adjacency matrix is somewhat similar, but the rows in $\Sigma$ don't have to sum to zero as they do for the Laplacian.


Your Answer

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

Browse other questions tagged or ask your own question.