I have one question regarding how to measure the correlation between minimum singular values of submatrices extracted from a large random matrix.
Given a m*n random matrix $\mathbf{A}$, the submatrices of it are generated by randomly selecting n out of m rows. Let $\sigma_{\mathrm{min},i}$ be the minimum singular value of submatrix $i$. Obviously, there is correlation between these minimum singular values. The question is how can we measure such correlation in simulation or analytic manner? Is there any difference if the matrix $\mathbf{A}$ is a real matrix or complex matrix?
Thanks a lot!!