Abstract
This paper addresses, via thresholding, the estimation of a possibly sparse signal observed subject to Gaussian noise. Conceptually, the optimal threshold for such problems depends upon the strength of the underlying signal. We propose two new methods that aim to adapt to potential local variation in this signal strength and select a variable threshold accordingly. Our methods are based upon an empirical Bayes approach with a smoothly variable mixing weight chosen via either spline or kernel based marginal maximum likelihood regression. We demonstrate the excellent performance of our methods in both one and two-dimensional estimation when compared to various alternative techniques. In addition, we consider the application to wavelet denoising where reconstruction quality is significantly improved with local adaptivity.
Similar content being viewed by others
References
Abramovich, F., Benjamini, Y., Donoho, D.L., Johnstone, I.M.: Adapting to unknown sparsity by controlling the false discovery rate. Ann. Stat. 34, 584–653 (2006)
Baraud, Y., Giraud, C., Huet, S.: Gaussian model selection with an unknown variance. Ann. Stat. (2008, to appear)
Beran, R.: Hybrid shrinkage estimators using penalty bases for the ordinal one-way layout. Ann. Stat. 32, 2532–2558 (2004)
Breiman, L., Friedman, J.H., Olshen, R.A., Stone, C.J.: Classification and Regression Trees. Wadsworth, Belmont (1984)
Broyden, C.G.: A new double rank minimization algorithm. Not. Am. Math. Soc. 16, 670 (1969)
Burman, P.: A comparative study of ordinary cross-validation, v-fold cross-validation and the repeated learning-testing methods. Biometrika 76, 503–514 (1989)
Cai, T.T.: On block thresholding in wavelet regression: Adaptivity, block size, and threshold level. Stat. Sin. 12, 1241–1273 (2002)
Cai, T.T., Silverman, B.W.: Incorporating information on neighbouring coefficients into wavelet estimation. Sankhyā Ser. B 63, 127–148 (2001)
De Boor, C.: A Practical Guide to Splines. Springer, New York (1994)
Donoho, D.L., Johnstone, I.M.: Ideal spatial adaptation by wavelet shrinkage. Biometrika 81, 425–455 (1994)
Donoho, D.L., Johnstone, I.M.: Adapting to unknown smoothness via wavelet shrinkage. J. Am. Stat. Assoc. 90, 1200–1224 (1995)
Fletcher, R.: A new approach to variable metric algorithms. Int. Bibliogr. Inf. Doc. 13, 317–322 (1970)
Goldfarb, D.: A family of variable metric methods derived by variational means. Math. Comp. 24, 23–26 (1970)
Green, P.J., Silverman, B.W.: Nonparametric Regression and Generalized Linear Models: A Roughness Penalty Approach. Chapman and Hall, London (1994)
Heaton, T.: Wavelets and accompanying applications. Ph.D. thesis, University of Bristol (2005)
Johnstone, I.M., Silverman, B.W.: Needles and straw in haystacks: Empirical Bayes estimates of possible sparse sequences. Ann. Stat. 32, 1594–1649 (2004)
Johnstone, I.M., Silverman, B.W.: Ebayesthresh: R programs for empirical Bayes thresholding. J. Stat. Softw. 12, 1–38 (2005a)
Johnstone, I.M., Silverman, B.W.: Empirical Bayes selection of wavelet thresholds. Ann. Stat. 33, 1700–1752 (2005b)
Mallat, S.G.: Multiresolution approximations and wavelet orthonormal bases of l 2(ℝ). Trans. Am. Math. Soc. 315, 69–87 (1989)
Midcourse Space Experiment: (1996–1997) NASA/IPAC infrared science archive. http://irsa.ipac.caltech.edu/index.html
Nason, G.P.: Wavelet shrinkage using cross-validation. J. R. Stat. Soc. B 58, 463–479 (1996)
Shanno, D.F.: Conditioning of quasi-newton methods for function minimization. Math. Comp. 24, 647–656 (1970)
Stone, M.: Cross-validatory choice and assessment of statistical predictions (with discussion). J. R. Stat. Soc. B 36, 114–147 (1974)
Wand, M.P., Jones, M.C.: Kernel Smoothing. Chapman and Hall, London (1995)
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Heaton, T.J. Adaptive thresholding of sequences with locally variable strength. Stat Comput 19, 57–71 (2009). https://doi.org/10.1007/s11222-008-9071-1
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11222-008-9071-1