Set-theoretic adaptive filtering based on data-driven sparsification

In this article, we propose a fast and efficient algorithm named the adaptive parallel Krylov‐metric projection algorithm. The proposed algorithm is derived from the variable‐metric adaptive projected subgradient method, which has recently been presented as a unified analytic tool for various adapti...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:International journal of adaptive control and signal processing Ročník 25; číslo 8; s. 707 - 722
Hlavní autoři: Yukawa, Masahiro, Yamada, Isao
Médium: Journal Article
Jazyk:angličtina
Vydáno: Chichester, UK John Wiley & Sons, Ltd 01.08.2011
Témata:
Adaptive algorithms
Adaptive filters
Algorithms
Convergence
Mathematical analysis
Optimization
Projection
set-theoretic adaptive filtering
subgradient projection
variable-metric projection
ISSN:0890-6327, 1099-1115, 1099-1115
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:In this article, we propose a fast and efficient algorithm named the adaptive parallel Krylov‐metric projection algorithm. The proposed algorithm is derived from the variable‐metric adaptive projected subgradient method, which has recently been presented as a unified analytic tool for various adaptive filtering algorithms. The proposed algorithm features parallel projection—in a variable‐metric sense—onto multiple closed convex sets containing the optimal filter with high probability. The metric is designed based on (i) sparsification by means of a certain data‐dependent Krylov subspace and (ii) maximal use of the obtained sparse structure for fast convergence. The numerical examples show the advantages of the proposed algorithm over the existing ones in stationary/nonstationary environments. Copyright © 2011 John Wiley & Sons, Ltd.
Bibliografie:istex:EE42B9D713B83EFBC438DBF2267B718E50087AD3
ark:/67375/WNG-6HP97T1Q-K
ArticleID:ACS1237
ObjectType-Article-2
SourceType-Scholarly Journals-1
ObjectType-Feature-1
content type line 23
ISSN:0890-6327
1099-1115
1099-1115
DOI:10.1002/acs.1237