•   Home
  • Proceedings
  • International Conference on Information, Electronic and Comp...

International Conference on Information, Electronic and Computer Science (ICIECS 2010 E-BOOK)

Zibo,China,11.26-11.28,2010

ISBN: 978-1-935068-42-6 Scientific Research Publishing, USA

E-Book 2224pp Pub. Date: November 2010

Category: Computer Science & Communications

Price: $360

Title: The Preprocessing Technique of Independent Component Analysis: On the Approach of Source Signals AR Modeling
Source: International Conference on Information, Electronic and Computer Science (ICIECS 2010 E-BOOK) (pp 1600-1604)
Author(s): Shuxin Zheng, Faculty of information science and engineering, Ningbo University, Ningbo, P.R China
Weiqiang Zhang, Faculty of information science and engineering, Ningbo University, Ningbo, P.R China
Abstract: Independent component analysis (ICA) is a statistical and computational technique for revealing hidden factors with statistical independence and to meet the non-Gaussian. Before applying an ICA algorithm, it is usually very useful to do some preprocessing, e.g., whitening is performed before applying the source separation algorithm because of that sources are often not white. However, the preprocessing techniques just like whitening, centering, factor analysis[1][2]and so on, that make the problem of ICA estimation simpler and better conditioned[3].Since the AR models describe fairly well many natural stochastic processes, e.g., speech signals, they are used in many applications. This paper aims at the preprocessing technique of ICA and especially focus on the performance of ICA based on the source signals AR modeling. It also presents the structure model of ICA based on the source signals preprocessing. Several simulation experimental results demonstrate that the source signals AR modeling can improve the performance of ICA algorithm on numerical stability.
Proceedings by Subjects
Resources
Products
Contact us
book@scirp.org
WhatsApp +86 18163351462(WhatsApp)
Click here to send a message to me 1243940697
Book Publishing WeChat
Follow SCIRP
Free SCIRP Newsletters
Copyright © 2006-2024 Scientific Research Publishing Inc. All Rights Reserved.
Top