Presentation 1997/6/27
Effect of the updating block size on ICA algorithms
Allan Kardec Barros, Hani Yehia, Noboru Ohnishi,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) Independent component analysis (ICA) is a new kind of algorithm used for separating mixed sources. Here we study the influence of the updating block size on the convergence of ICA algorithms, for stationary and non-stationary data. We found that ICA algorithms with a small enough number of observations: a) do not necessarily converge to a solution which cancels the mixing effect; b) depend on the initial estimated parameters; c) separate more than one Gaussian source.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) blind separation / independent component analysis / batch update / Gaussian signals
Paper # CAS97-24
Date of Issue

Conference Information
Committee DSP
Conference Date 1997/6/27(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 Digital Signal Processing (DSP)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Effect of the updating block size on ICA algorithms
Sub Title (in English)
Keyword(1) blind separation
Keyword(2) independent component analysis
Keyword(3) batch update
Keyword(4) Gaussian signals
1st Author's Name Allan Kardec Barros
1st Author's Affiliation Graduate School of Engineering, Nagoya University()
2nd Author's Name Hani Yehia
2nd Author's Affiliation ATR Human Information Lab.
3rd Author's Name Noboru Ohnishi
3rd Author's Affiliation Graduate School of Engineering, Nagoya University:RIKEN BMC Research Center
Date 1997/6/27
Paper # CAS97-24
Volume (vol) vol.97
Number (no) 141
Page pp.pp.-
#Pages 4
Date of Issue