Accuracy of singular vectors obtained by projection-based SVD methods

The standard approach to computing an approximate SVD of a large-scale matrix is to project it onto lower-dimensional trial subspaces from both sides, compute the SVD of the small projected matrix, and project it back to the original space. This results in a low-rank approximate SVD to the original matrix, and we can then obtain approximate left and right singular subspaces by extracting subsets from the approximate SVD. In this work we assess the quality of the extraction process in terms of the accuracy of the approximate singular subspaces, measured by the angle between the exact and extracted subspaces (relative to the angle between the exact and trial subspaces). The main message is that the extracted approximate subspaces are optimal usually to within a modest constant.