I'am trying to understand Singular Value Decomposition(SVD) fully.
I know that when we do PCA, we usually do centering(subtracting mean) to build a covariance matrix.
Regardless above, I want to understand the meaning of SVD of matrix of size (m,n) without centering.
In my gut feeling, this will also quite catch the principle component of matrix without centering, but not sure for this is correct.
Is there any intuition about SVD of raw (m,n) data matrix?