计算机科学 ›› 2014, Vol. 41 ›› Issue (6): 63-68.doi: 10.11896/j.issn.1002-137X.2014.06.013
李繁,金明录,刘继
LI Fan,JIN Ming-lu and LIU Ji
摘要： 对奇异值分解(SVD)求解最小平方估计的问题进行了研究。提出迭代式分割与合并的算法(IDMSVD),目的是解决奇异值分解在估计参数时非常耗费内存空间的问题。基于IDMSVD提出了并行IDMSVD算法,并使用GPU实现之。实验结果显示,IDMSVD可以有效地解决SVD求最小平方解耗费运行时间与内存空间的问题,并行IDMSVD算法可进一步改善IDMSVD的运行时间。
| [1] Montgomery D C,Peck E A,Vining G G．Introduction to linear regression analysis(4th ed)[M]．Hoboken,N.J.,USA:Wiley-Interscience,2006:68-113 [2] Myers R H,Montgomery D C,Vining G G,et al.Generalizedlinear models:with applications in engineering and the sciences(2nd ed)[M]．Hoboken,N.J.,USA:Wiley-Interscience,2010:217-339 [3] Lee S-J,Ouyang C-S．A neuro-fuzzy system modeling with self-constructing rule generation and hybrid SVD-based learning [J]．IEEE Transactions on Fuzzy Systems,2003,11(3):341-363 [4] Foster L V．Solving rank-deficient and ill-posed problems using UTV and QR factorizations [J]．SIAM Journal on Matrix Ana-lysis and Applications,2003,6(2):682-600 [5] Moler C B．Numerical computing with matlab[M]．Philadelphia,PA,USA:Society for Industrial Mathematics,2004:71-165 [6] Hari V．Accelerating the SVD block-jacobi method [J]．Computing,2006,6(1):27-63 [7] Yamamoto Y,Fukaya T,Uneyama T,et al.Accelerating the singular value decomposition of rectangular matrices with the CSX600and the integrable SVD[J]．Lecture Notes in Computer Science,2007,6(7):340-346 [8] Kondaa T,Nakamura Y．A new algorithm for singular value decomposition and its parallelization[J]．Parallel Computing,2009,36(6):331-344 [9] Beˇka M,Oka G,Vajteric M,et al．On iterative QR pre-processing in the parallel block-jacobi SVD algorithm[J]．Parallel Computing,2009,6(6):297-307 [10] Ltaief H,Kurzak J,Dongarra J．Parallel two-sided matrix reduction to band bidiagonal form on multicore architectures[J]．IEEE Transactions on Parallel and Distributed Systems,2010,1(4):417-423 |
| No related articles found! |
|
||
首页