Learning out of leaders

Gérard Kerkyacharian, Mathilde Mougeot, Dominique Picard () and Karine Tribouley
Additional contact information
Gérard Kerkyacharian: Université Paris-Diderot, CNRS LPMA
Mathilde Mougeot: Université Paris Ouest Nanterre, MODALX
Dominique Picard: Université Paris-Diderot, CNRS LPMA
Karine Tribouley: Université Paris Ouest Nanterre, MODALX

A chapter in Multiscale, Nonlinear and Adaptive Approximation, 2009, pp 295-324 from Springer

Abstract: Abstract In this paper we investigate the problem of supervised learning. We are interested in universal procedures producing exponential bounds. One main purpose is to link this problem to a general approach on high dimensional linear models in statistics and to propose some tools resulting from a combination of inspirations: many of them coming from previous works of Wolfgang Dahmen and coauthors combined with regression and thresholding techniques. We present different types of algorithms initiated in Wolfgang Dahmen’s (and coauthors) work and provide a new algorithm: the LOL procedure. We prove that it has optimal exponential rate of convergence. We also study the practical behavior of the procedure: our simulation study confirms its very good properties.

Keywords: Reconstruction Error; Internal Coherence; Sparsity Level; Outer Leaf; Sparsity Condition (search for similar items in EconPapers)
Date: 2009
References: Add references at CitEc
Citations:

There are no downloads for this item, see the EconPapers FAQ for hints about obtaining it.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:spr:sprchp:978-3-642-03413-8_9

Ordering information: This item can be ordered from
http://www.springer.com/9783642034138

DOI: 10.1007/978-3-642-03413-8_9

Access Statistics for this chapter

More chapters in Springer Books from Springer
Bibliographic data for series maintained by Sonal Shukla () and Springer Nature Abstracting and Indexing ().