We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

math.ST

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Mathematics > Statistics Theory

Title: On optimality of empirical risk minimization in linear aggregation

Abstract: In the first part of this paper, we show that the small-ball condition, recently introduced by Mendelson (2015), may behave poorly for important classes of localized functions such as wavelets, piecewise polynomials or trigonometric polynomials, in particular leading to suboptimal estimates of the rate of convergence of ERM for the linear aggregation problem. In a second part, we recover optimal rates of covergence for the excess risk of ERM when the dictionary is made of trigonometric functions. Considering the bounded case, we derive the concentration of the excess risk around a single point, which is an information far more precise than the rate of convergence. In the general setting of a L2 noise, we finally refine the small ball argument by rightly selecting the directions we are looking at, in such a way that we obtain optimal rates of aggregation for the Fourier dictionary.
Comments: 29 pages
Subjects: Statistics Theory (math.ST)
Cite as: arXiv:1605.03433 [math.ST]
  (or arXiv:1605.03433v3 [math.ST] for this version)

Submission history

From: Adrien Saumard [view email]
[v1] Wed, 11 May 2016 13:25:39 GMT (29kb)
[v2] Tue, 7 Jun 2016 11:37:26 GMT (29kb)
[v3] Tue, 18 Oct 2016 09:35:34 GMT (27kb)

Link back to: arXiv, form interface, contact.