|Table of Contents|

Modified VC-Dimension-based Wavelet-denoising Method of Non-stationary Signals

《南京理工大学学报》(自然科学版)[ISSN:1005-9830/CN:32-1397/N]

Issue:
2009年05期
Page:
648-652
Research Field:
Publishing date:

Info

Title:
Modified VC-Dimension-based Wavelet-denoising Method of Non-stationary Signals
Author(s):
ZHANG Deng-feng1WANG Zhi-quan2ZHANG Wei1
1.School of Mechanical Engineering;2.School of Automation,NUST,Nanjing 210094,China
Keywords:
signal denoising Vapnik-Chervonenkis dimension wavelet multi-resolution analysis function estimation structural risk
PACS:
TN911.7
DOI:
-
Abstract:
By using the orthogonal wavelet multi-resolution analysis technology,the filter-denoising problem is investigated for non-stationary signals with non-over-sampling and finite samples.Based on the structural risk minimization(SRM) principle and Vapnik-Chervonenkis dimension(VC-dimension) theory of statistical learning theory,a modified VC-dimension based wavelet-denoising approach is proposed.The approach can ensure the minimum actual risk of denoised signals in the view of function estimation,overcoming the drawbacks of application of traditional wavelet-denoising approaches.The simulation results from denoising the characteristic non-stationary signals show that,compared with several existing denoising approaches,the approach proposed here can improve the denoising performance and realize the optimized tradeoff between the denoising performance and the model complexity.

References:

[ 1]  CherkasskyV, KiltsS. Myopotentialdenoising ofECG signalsusingwavelet thresholdingmethods[ J]. Neural Networks, 2001, 14( 8): 1129- 1137.
[ 2] CherkasskyV, ShaoX. Signal estimationanddenois ing using VCtheory [ J]. Neural Networks, 2001, 14( 1): 37-52.
[ 3] VapnikV N. Statistical learning theory[M]. New York: WileyPress, 1998.
[ 4] 彭玉华. 小波变换与工程应用[M]. 北京: 科学出版社, 1999. 36-51.
[ 5] BriceA, DonohoD, GaoHY. Wavelet analysis[ J]. IEEESpectrum, 1996, 33(10): 26- 35.
[ 6] VapnikVN. Anoverviewof statistical learning theory [ J]. IEEETransactions onNeural Networks, 1999, 10( 5): 988- 999.
[ 7] VapnikVN. 统计学习理论的本质[M]. 张学工, 译. 北京: 清华大学出版社, 2000. 52- 65.
[ 8] CherkasskyV, ShaoX, Mulier F, et a.l Model com plexity control for regression using VCgeneralization bounds[ J]. IEEETransactions onNeural Networks, 1999, 10( 5): 1075- 1089.
[ 9] DonohoDL. Denoisingby soft thresholding[ J]. IEEE Transactions on Information Theory, 1995, 41( 3): 613-627.
[ 10] 胡昌华,张军波, 夏军,等. 基于MATLAB的系统分析与设计( 小波分析[M]. 西安:西安电子科技大学出版社, 1999. 109- 119.
[ 11] WangWJ, McFaddenPD. Applicationofwavelets to gearbox vibrationsignals for fault detection[ J]. Inter national Journal of Sound and Vibration, 1996, 192( 5): 927-939.

Memo

Memo:
-
Last Update: 2012-11-19