Sparse symmetric rank-revealing decompositions |
|
| Abstract | This thesis investigates the possibility of a sparse symmetric rank-revealing VSV decomposition. Rank-revealing decompositions of both general and symmetric rank-deficient matrices are described, as well as the numerical tools used in these. The cause of fill in the VSV decomposition is investigated, and numerical experiments with possible sparsity optimizations are discussed. Drop-tolerances in the estimated singular vectors are described as a tool in obtaining a sparse decomposition. |
| Keywords | Sparse VSV decomposition, norm and condition estimation, ordering schemes, sparse factorizations and drop-tolerances |
| Type | Master's thesis [Academic thesis] |
| Year | 2002 |
| Publisher | Informatics and Mathematical Modelling, Technical University of Denmark, DTU |
| Address | Richard Petersens Plads, Building 321, DK-2800 Kgs. Lyngby |
| Series | IMM-THESIS-2002-73 |
| Note | Supervisor: Per Christian Hansen |
| Electronic version(s) | [zip] |
| BibTeX data | [bibtex] |
| IMM Group(s) | Scientific Computing |