## 2 MATLAB implementations:1. C. G. Baker's algorithm*, which was also implemented by C. G. Baker.http://www.math.fsu.edu/~cbaker/IncPACK/ *C. G. Baker. "A block incremental algorithm for computing dominant singular subspaces". 2004 2. M. Brand's algorithm+, which was implemented by David Wingate. http://www.mit.edu/~wingated/resources.html (Search for the section Incremental SVD updates)+Matthew Brand, "Fast low-rank modifications of the thin singular value decomposition". ## A C/C++ implementation:This version is based on the algorithm by Gu*. According to this post, the author is Attila Axt.http://sourceforge.net/projects/isvd/ Quoted from the official website: `Description ` `Incremental singular value decomposition (SVD) routines for the Gnu Scientific Library (GSL). It outperforms the standard batch-SVD methods in running/windowed SVD applications.` * M. Gu, Studies in numerical linear algebra, Ph.D. thesis, Yale University, New Haven, CT, USA (1993) (ACM Portal). I cannot find the original thesis, but I found the following related publications: - M. Gu, S. C. Eisenstat, Downdating the singular value
decomposition, Research Report, YALEU/DCS/RR-939, Dept. of Computer
Science, Yale University, New Haven, CT, 1993, May (link).
- M. Gu, S. C. Eisenstat, A stable and fast algorithm for updating the singular value decomposition, Research Report, YALEU/DCS/RR-966, Dept. of Computer Science, Yale University, New Haven, CT, 1993, June
- Ming Gu, Stanley C. Eisenstat, A stable and efficient algorithm for the rank-one modification of the symmetric eigenproblem, SIAM J. Matrix Anal. Appl.,
**15**(1994), 1266–1276 [ISI] [ZentralblattMath] [MathRev]
## Notes when compiling isvdFirst, gsl is required. Second, when configuring isvd, use the following command s.t. configure can find gsl: env CPPFLAGS="-I</gsl/to/path>/include" LDFLAGS="-L</gsl/to/path>/lib" ./configure CPPFLAGS=-I</gsl/to/path>/include CFLAGS=-I</gsl/to/path>/include LDFLAGS=-L</gsl/to/path>/lib |

ResearchBlog >