Adaptive filtering by minimizing error's entropy power

Badong Chen; Yu Zhu; Jinchun Hu; Ming Zhang

Adaptive filtering by minimizing error's entropy power

Badong Chen
, Yu Zhu
, Jinchun Hu
, Ming Zhang

Tsinghua University

Research output: Contribution to journal › Article › peer-review

Abstract

In this paper, we propose the minimum error's entropy power (MEEP) criterion as an alternative to the minimum error entropy (MEE) criterion in adaptive filtering. Some important properties of the MEEP criterion are presented. Under MEEP criterion, the stochastic gradient algorithm for FIR filter is derived, which is actually the stochastic information gradient (SIG) algorithm with variable step-size. Theoretic analysis and simulation results emphasize the better performance of this new algorithm.

Original language	English
Pages (from-to)	2511-2520
Number of pages	10
Journal	Journal of Information and Computational Science
Volume	5
Issue number	6
State	Published - Dec 2008
Externally published	Yes

Keywords

Adaptive filtering
Entropy power
Stochastic gradient algorithm

Cite this

@article{b8b76360457342e1b21936acfcb0decb,

title = "Adaptive filtering by minimizing error's entropy power",

abstract = "In this paper, we propose the minimum error's entropy power (MEEP) criterion as an alternative to the minimum error entropy (MEE) criterion in adaptive filtering. Some important properties of the MEEP criterion are presented. Under MEEP criterion, the stochastic gradient algorithm for FIR filter is derived, which is actually the stochastic information gradient (SIG) algorithm with variable step-size. Theoretic analysis and simulation results emphasize the better performance of this new algorithm.",

keywords = "Adaptive filtering, Entropy power, Stochastic gradient algorithm",

author = "Badong Chen and Yu Zhu and Jinchun Hu and Ming Zhang",

year = "2008",

month = dec,

language = "英语",

volume = "5",

pages = "2511--2520",

journal = "Journal of Information and Computational Science",

issn = "1548-7741",

publisher = "Binary Information Press",

number = "6",

}

TY - JOUR

T1 - Adaptive filtering by minimizing error's entropy power

AU - Chen, Badong

AU - Zhu, Yu

AU - Hu, Jinchun

AU - Zhang, Ming

PY - 2008/12

Y1 - 2008/12

N2 - In this paper, we propose the minimum error's entropy power (MEEP) criterion as an alternative to the minimum error entropy (MEE) criterion in adaptive filtering. Some important properties of the MEEP criterion are presented. Under MEEP criterion, the stochastic gradient algorithm for FIR filter is derived, which is actually the stochastic information gradient (SIG) algorithm with variable step-size. Theoretic analysis and simulation results emphasize the better performance of this new algorithm.

AB - In this paper, we propose the minimum error's entropy power (MEEP) criterion as an alternative to the minimum error entropy (MEE) criterion in adaptive filtering. Some important properties of the MEEP criterion are presented. Under MEEP criterion, the stochastic gradient algorithm for FIR filter is derived, which is actually the stochastic information gradient (SIG) algorithm with variable step-size. Theoretic analysis and simulation results emphasize the better performance of this new algorithm.

KW - Adaptive filtering

KW - Entropy power

KW - Stochastic gradient algorithm

UR - https://www.scopus.com/pages/publications/67649223493

M3 - 文章

AN - SCOPUS:67649223493

SN - 1548-7741

VL - 5

SP - 2511

EP - 2520

JO - Journal of Information and Computational Science

JF - Journal of Information and Computational Science

IS - 6

ER -

Adaptive filtering by minimizing error's entropy power

Abstract

Keywords

Other files and links

Fingerprint

Cite this