Abstract
The major difficulty arising in statistics of multi-variable functions is “the curse of dimensionality”: the rates of accuracy in estimation and separation rates in detection problems behave poorly when the number of variables increases. This difficulty arises for most popular functional classes such as Sobolev or Hölder balls.
In this paper we consider functional classes of a new type, first introduced by Sloan and Woźniakowski in 1998. We consider balls F σ,s in a “Sloan—Woźniakowski” or “weighted Sobolev” space characterized by two parameters: σ > 0 is a “smoothness” parameter, and s > 0 determines the weight sequence which describes “importance” of the variables. Previously Kuo and Sloan [18] used the spaces of similar structure to address the problem of numerical integration.
For the classes F σ,s we show that in the white Gaussian noise model, the separation rates in detection are similar to those for one-variable functions of smoothness σ* = min(s,σ) regardless of the original problem dimension; thus the curse of dimensionality is “lifted”. Similar results hold for the estimation problem.
The studies are based on known results for estimation and detection problems for ellipsoids. Using these results, the asymptotics in the problems are determined by asymptotics of “distribution of coefficients” of ellipsoids. The key point of the paper is the study of these asymptotics for the balls F σ,s .
Similar content being viewed by others
References
E. N. Belitser and B. Y. Levit, “On Minimax Filtering over Ellipsoids”, Math. Methods Statist. 4, 259–273 (1995).
R. Bellman, Dynamic Programming (Princeton Univ. Press, Princeton—New York, 1957).
L. D. Brown and M. G. Low, “Asymptotic Equivalence of Nonparametric Regression and White Noise”, Ann. Statist. 24, 2384–2398 (1996).
T. Gai and M. G. Low, “Nonparametric Estimation over Shrinking Neighborhoods: Superefficiency and Adaptation”, Ann. Statist. 33, 184–213 (2002).
D. L. Donoho and I. M. Johnstone, “Minimax Estimation via Wavelet Shrinkage”, Ann. Statist. 26, 668–701 (1998).
D. L. Donoho, “High-Dimensional Data Analysis: The Curses and Blessings of Dimensionality”, in Lecture on the AMS Conference “Mathematical Challenges of the 21st Century”, Los Angeles, August 6–11, 2000. http://www-stat.stanford.edu/ donoho/Lectures/AMS2000/AMS2000.html
S. V. Efromovich, Nonparametric Curve Estimation (Springer, New York, 1999).
M. S. Ermakov, “Minimax Detection of a Signal in a Gaussian White Noise”, Theory Probab. Appl. 35, 667–679 (1990).
I. A. Ibragimov and R. Z. Khasminskii, “One Problem of Statistical Estimation in a White Gaussian Noise”, Soviet Math. Dokl. 236(4), 333–337 (1977).
I. A. Ibragimov and R. Z. Khasminskii, “Some Estimation Problems on Infinite Dimensional Gaussian White Noise”, in Festschrift for Lucien Le Cam. Research papers in Probab. and Statist. (Springer, New York, 1997), pp. 275–296.
Yu. I. Ingster, “Minimax Nonparametric Detection of Signals in White Gaussian Noise”, Problems Inform. Transmission 18, 130–140 (1982).
Yu. I. Ingster, “Asymptotically Minimax Testing of Nonparametric Hypotheses”, in Proc. 4th Vilnius Conference on Probab. Theory and Math. Statist. (VNU Science Press, Vilnius, 1987), Vol. 1, pp. 553–573.
Yu. I. Ingster, “Asymptotically Minimax Hypothesis Testing for Nonparametric Alternatives. I, II, III”, Math. Methods Statist. 2, 85–114, 171–189, 249–268 (1993).
Yu. I. Ingster and I. A. Suslina, Nonparametric Goodness-of-Fit Testing under Gaussian Model, in Lectures Notes in Statist. (Springer, New York, 2002), Vol. 169.
Yu. I. Ingster and I. A. Suslina, “On Estimation and Detection of Smooth Functions of Many Variables”, Math. Methods Statist. 14(3), 299–331 (2005).
Yu. I. Ingster and I. A. Suslina, “On Estimation and Detection of Infinite-Variable Function”, Zapiski Nauchn. Seminar. POMI 328 (2006) [J. Math. Sci. 139 (3), 6548–6561 (2006)].
G. Kerkyacharian, O. Lepski, and D. Picard, “Nonlinear Estimation and Anisotropic Multi-Index Denoising”, Probab. Theory and Rel. Fields 121, 137–170 (2001).
F. Y. Kuo and J. H. Sloan, “Lifting the Curse of Dimensionality”, Notices of AMS 52(11), 1320–1329 (2005).
Y. Lin, “Tensor Product Space ANOVA Model”, Ann. Statist. 28(3), 734–755 (2000).
M. Nussbaum, “Asymptotic Equivalence of Density Estimation and Gaussian White Noise”, Ann. Statist. 24, 2399–2430 (1996).
V. V. Petrov, Sums of Independent Random Variables (Springer, New York, 1975).
M. S. Pinsker, “Optimal Filtration of Square-Integrable Signals in Gaussian Noise”, Problems Inform. Transmission 16(2), 120–133 (1980).
A. V. Skorohod, Integration in Hilbert Spaces (Springer, Berlin—New York, 1974).
I. H. Sloan and H. Woźniakowski, “When are Quasi-Monte Carlo Algorithms Efficient for High-Dimensional Integrals?” J. Complexity 14, 1–33 (1998).
Ch. Stone, “Additive Regression and Other Nonparamtric Models”, Ann. Statist. 13(2), 689–705 (1985).
H. Woźniakowski, “Tractability of Multivariate Problems for Weighted Spaces of Functions”, in Approximation and Probability (Banach Center Publications, Inst. of Math., Polish Acad. of Sci., Warszawa, 2006), Vol. 72.
Yu. I. Ingster and Th. Sapatinas, Minimax Goodness-of-Fit Gesting in Multivariate Nonparametric Regression, Rechnical Report TR-22-2007 (Univ. of Cyprus, 2007). http://www.mas.ucy.ac.cy/english/technical reports eng207.htm
Author information
Authors and Affiliations
Corresponding author
About this article
Cite this article
Ingster, Y., Suslina, I. Estimation and detection of high-variable functions from Sloan—Woźniakowski space. Math. Meth. Stat. 16, 318–353 (2007). https://doi.org/10.3103/S1066530707040035
Received:
Accepted:
Issue Date:
DOI: https://doi.org/10.3103/S1066530707040035
Key words
- estimation of multi-variable functions
- detection of multi-variable functions
- minimax estimation and hypothesis testing
- separation rates
- weighted tensor-product Sobolev spaces