Presentation 2007-12-13
Comparison of Eigenvalue Computation Speed for Time-Varying Large-Size Symmetric Matrices
Hiroyuki TAIRA, Kenichi KANATANI,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) We study iterative schemes for computing the largest eigenvalue and the corresponding eigenvector of a time-varying large-scale symmetric matrix, which frequently arises in many problems including computer vision. Here, we focus on the steepest descent, the conjugate gradient, the power method, and its acceleration. We first summarize their algorithms and then compare their computation time, using simulated data. We show how the computational efficiency of individual methods is related to the eigenvalue distribution of the matrix in question.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) large-scale matrix / eigenvalue problem / steepest descent / conjugate gradient method / power method / acceleration of convergence
Paper # PRMU2007-135
Date of Issue

Conference Information
Committee PRMU
Conference Date 2007/12/6(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Pattern Recognition and Media Understanding (PRMU)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Comparison of Eigenvalue Computation Speed for Time-Varying Large-Size Symmetric Matrices
Sub Title (in English)
Keyword(1) large-scale matrix
Keyword(2) eigenvalue problem
Keyword(3) steepest descent
Keyword(4) conjugate gradient method
Keyword(5) power method
Keyword(6) acceleration of convergence
1st Author's Name Hiroyuki TAIRA
1st Author's Affiliation Department of Computer Science, Okayama University()
2nd Author's Name Kenichi KANATANI
2nd Author's Affiliation Department of Computer Science, Okayama University
Date 2007-12-13
Paper # PRMU2007-135
Volume (vol) vol.107
Number (no) 384
Page pp.pp.-
#Pages 6
Date of Issue