Please use this identifier to cite or link to this item: http://hdl.handle.net/10397/68344
PIRA download icon_1.1View/Download Full Text
Title: Thresholded spectral algorithms for sparse approximations
Authors: Guo, ZC
Xiang, DH
Guo, X 
Zhou, DX
Issue Date: May-2017
Source: Analysis and applications, May 2017, v. 15, no. 3, p. 433-455
Abstract: Spectral algorithms form a general framework that unifies many regularization schemes in learning theory. In this paper, we propose and analyze a class of thresholded spectral algorithms that are designed based on empirical features. Soft thresholding is adopted to achieve sparse approximations. Our analysis shows that without sparsity assumption of the regression function, the output functions of thresholded spectral algorithms are represented by empirical features with satisfactory sparsity, and the convergence rates are comparable to those of the classical spectral algorithms in the literature.
Keywords: Learning theory
Thresholded spectral algorithm
Sparsity
Learning rate
Publisher: World Scientific
Journal: Analysis and applications 
ISSN: 0219-5305
EISSN: 1793-6861
DOI: 10.1142/S0219530517500026
Rights: Electronic version of an article published as Analysis and Applications, vol. 15, no. 3, 2017, p. 433-455, https://doi.org/10.1142/S0219530517500026, © World Scientific Publishing Company, https://www.worldscientific.com/worldscinet/aa
Appears in Collections:Journal/Magazine Article

Files in This Item:
File Description SizeFormat 
empirical161222.pdfPre-Published version734.36 kBAdobe PDFView/Open
Open Access Information
Status open access
File Version Final Accepted Manuscript
Access
View full-text via PolyU eLinks SFX Query
Show full item record

Page views

366
Last Week
3
Last month
Citations as of Mar 24, 2024

Downloads

127
Citations as of Mar 24, 2024

SCOPUSTM   
Citations

45
Last Week
0
Last month
Citations as of Mar 28, 2024

WEB OF SCIENCETM
Citations

45
Last Week
0
Last month
Citations as of Mar 28, 2024

Google ScholarTM

Check

Altmetric


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.