• DOI:10.1109/ISIT.2011.6033687
Corpus ID: 14081686

Selecting the rank of truncated SVD by maximum approximation capacity

@article{Frank2011SelectingTR,
  title={Selecting the rank of truncated SVD by maximum approximation capacity},
  author={M. Frank and J. Buhmann},
  journal={2011 IEEE International Symposium on Information Theory Proceedings},
  year={2011},
  pages={1036-1040}
}

M. Frank, J. Buhmann

Published 2011

Mathematics, Computer Science

2011 IEEE International Symposium on Information Theory Proceedings

Truncated Singular Value Decomposition (SVD) calculates the closest rank-k approximation of a given input matrix. Selecting the appropriate rank k defines a critical model order choice in most applications of SVD. To obtain a principled cut-off criterion for the spectrum, we convert the underlying optimization problem into a noisy channel coding problem. The optimal approximation capacity of this channel controls the appropriate strength of regularization to suppress noise. In simulation… CONTINUE READING

View on IEEE

arxiv.org

Save to Library

Create Alert

Cite

Launch Research Feed