I'm only interested in finding the largest singular value of an $m\times n$ real matrix $A$. I don't need the corresponding (left and right) singular vectors.
- Is there a way to do so without performing full SVD?
- Is there an analytical expression? If not, is there an analytical approximation?