Skip to main content
SpringerLink
Log in

Iterative Density Estimation from Contaminated Observations

  • Original Article
  • Published:
Metrika Aims and scope Submit manuscript

Abstract

We introduce an iterative procedure for estimating the unknown density of a random variable X from n independent copies of Y=X+ɛ, where ɛ is normally distributed measurement error independent of X. Mean integrated squared error convergence rates are studied over function classes arising from Fourier conditions. Minimax rates are derived for these classes. It is found that the sequence of estimators defined by the iterative procedure attains the optimal rates. In addition, it is shown that the sequence of estimators converges exponentially fast to an estimator within the class of deconvoluting kernel density estimators. The iterative scheme shows how, in practice, density estimation from indirect observations may be performed by simply correcting an appropriate ordinary density estimator. This allows to assess the effect that the perturbation due to contamination by ɛ has on the density to be estimated. We also suggest a method to select the smoothing parameter required by the iterative approach and, utilizing this method, perform a simulation study.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Similar content being viewed by others

References

  • Artiles LM (2001) Adaptive minimax estimation in classes of smooth functions. Ph D Thesis, University of Utrecht, Netherlands

    Google Scholar 

  • Blum T, Susarla V (1977) Estimation of a mixing distribution function. Ann Probab 5:200–209

    Article  MATH  MathSciNet  Google Scholar 

  • Carroll RJ, Ruppert D, Stefanski LA (1995) Measurement error in nonlinear models. Chapman Hall, Great Britain

    MATH  Google Scholar 

  • Carroll RJ, Hall P (2003) Low-order approximations in deconvolution and regression with errors in variables. Technical Report, Department of Statistics, Texas A & M University

  • Cator EA (2001) Deconvolution with arbitrary smooth kernels. Stat Prob Lett 54:205–214

    Article  MATH  MathSciNet  Google Scholar 

  • Davis KB (1977) Mean integrated square error properties of density estimates. Ann Statist 5:530–535

    Article  MATH  MathSciNet  Google Scholar 

  • Diggle PJ, Hall P (1993) A Fourier-approach to nonparametric deconvolution of a density estimate. J R Stat Soc B 55:523–531

    MATH  MathSciNet  Google Scholar 

  • Efromovich S (1997) Density estimation in the case of supersmooth measurement error. J Amer Stat Assoc 92:526–535

    Article  MATH  MathSciNet  Google Scholar 

  • Fan J (1991a) Asymptotic normality for deconvoluting kernel density estimators. Sankhya A 53:97–110

    MATH  Google Scholar 

  • Fan J (1991b) On the optimal rates of convergence for non-parametric deconvolution problems. Ann Stat 19:1257–1272

    Article  MATH  Google Scholar 

  • Fan J (1992) Deconvolution with supersmooth distributions. Canad J Stat 20:155–169

    Article  MATH  Google Scholar 

  • Fienup JR (1978) Reconstruction of an object from the modulus of its Fourier transform. Optics Lett 3:27–29

    Article  Google Scholar 

  • Goldenshluger A (1999) On pointwise adaptive nonparametric deconvolution. Bernoulli 5:907–925

    Article  MATH  MathSciNet  Google Scholar 

  • Hayes MH, Lim JS, Oppenheim AV (1980) Phase-only signal restoration. In: Proceeding of 1980 int conf acoustics, speech, and signal processing, pp. 437–440

  • Hesse CH (1995a) Deconvolving a density from partially contaminated observations. J Multivariate Anal 55:246–260

    Article  MATH  MathSciNet  Google Scholar 

  • Hesse CH (1995b) How many “good” observations do you need for “fast” density deconvolution from supersmooth errors. Sankhya A 58:491–506, 10:343–373

    Google Scholar 

  • Hesse CH (1999) Data-driven deconvolution. J Nonparam Stat 10:343–373

    Article  MATH  MathSciNet  Google Scholar 

  • Ibragimov IA (2001) Estimation of analytic functions. State of the Art in Probability and Statistics, Festschrift for W. R. van Zwet (M. de Gunst, C. Klaassen, A. van der Vaart, eds.) IMS Lecture Notes, 36:359–383

  • Ibragimov IA, Khasminskii R (1981) Statistical Estimation: Asymptotic Theory. Springer, Heidelberg New York

    MATH  Google Scholar 

  • Kerkyacharian G, Picard D (1992) Density estimation in Besov spaces. Stat and Probab Lett 13:15–24

    Article  MATH  MathSciNet  Google Scholar 

  • Landau HJ, Miranker WL (1961) The recovery of distorted band-limited signals. J Math Anal Appl 2:97–104

    Article  MATH  MathSciNet  Google Scholar 

  • Liu MC, Taylor RC (1989) A consistent nonparametric density estimator for the deconvolution problem. Can J Stat 17:427–438

    Article  MATH  MathSciNet  Google Scholar 

  • Masry E, Rice JA (1992) Gaussian deconvolution via differentiation. Can J Stat 20:9–21

    Article  MATH  MathSciNet  Google Scholar 

  • Neumann MH (1997) On the effect of estimating the error density in nonparametric deconvolution. J Nonparam Stat 7:307–330

    Article  MATH  Google Scholar 

  • Papoulis A (1975) A new algorithm in spectral analysis and band-limited extra-polation. IEEE Trans Circuits Syst CAS-22:735–742

    Article  MathSciNet  Google Scholar 

  • Pensky M, Vidakovic B (1999) Adaptive wavelet estimator for nonparametric density deconvolution. Ann Stat 27:2033–2053

    Article  MATH  MathSciNet  Google Scholar 

  • Sabri MS, Steenaart W (1978) An approach to bandlimited signal extrapolation: The extrapolation matrix. IEEE Trans Circuits Syst CAS-25:74–78

    Article  Google Scholar 

  • Schafer RW, Mersereau RM, Richards MA (1981) Constrained iterative restoration algorithms. Proc IEEE 89:432–450

    Article  Google Scholar 

  • Silverman BW (1986) Density estimation for statistics and data analysis. Chapman and Hall, London

    MATH  Google Scholar 

  • Stefanski LA, Carroll RJ (1990) Deconvoluting kernel density estimators. Statistics 21:169–184

    Article  MATH  MathSciNet  Google Scholar 

  • Stefanski LA (1990) Rates of convergence of some estimators in a class of deconvolution problems. Stat Prob Lett 9:229–235

    Article  MATH  MathSciNet  Google Scholar 

  • Stefanski LA (2000) Measurement error models. J Amer Stat Assoc 95:1353–1358

    Article  MATH  MathSciNet  Google Scholar 

  • Thurigen D, Spiegelman D, Blettner M, Heurer C, Brenner H (2000) Measurement error correction using validation data: a review of methods and their applicability in case-control studies. Stat Meth Med Res 9:447–474

    Article  Google Scholar 

  • Van Es B, Uh HW (2000) Multi-bandwidth kernel estimators for nonparametric deconvolution problems: asymptotics and finite sample performance. J Nonparam Stat 13:107–128

    Article  MATH  MathSciNet  Google Scholar 

  • Watson GS, Leadbetter MR (1963) On the estimation of the probability density. I Ann Math Stat 34:480–491

    Article  MathSciNet  MATH  Google Scholar 

  • Yuan M, Chen JQ (2002) Deconvolving multidimensional density from partially contaminated observations. J Stat Plann Inf 104:147–160

    Article  MATH  MathSciNet  Google Scholar 

Download references

Author information

Authors and Affiliations

  1. FachbereichMathematik, Universität Stuttgart, Pfaffenwaldring 57, 70569, Stuttgart, Germany

    Christian H. Hesse

Authors
  1. Christian H. Hesse
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Christian H. Hesse.

Rights and permissions

Reprints and permissions

About this article

Cite this article

Hesse, C.H. Iterative Density Estimation from Contaminated Observations. Metrika 64, 151–165 (2006). https://doi.org/10.1007/s00184-006-0041-3

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s00184-006-0041-3

Keywords

AMS Mathematics Subject Classification (1991)

Search

Navigation