Abstract
This paper proposes a two-stage correlated non-linear shrinkage estimation methodology for spatial random processes. A block hard thresholding design, based on Shannon’s entropy, is formulated in the first stage. The thresholding design is adaptive to each resolution level, because it depends on the empirical distribution function of the mutual information ratios between empirical wavelet blocks and the random variables of interest, at each scale. In the second stage, a global correlated (inter- and intra-scale) shrinkage is applied to approximate the values of interest of the underlying spatial process. Additionally, a simulation study is developed, in the Gaussian context, to analyze the sensitivity, measured by empirical stochastic ordering, of the entropy-based block hard thresholding stage in relation to the parameters characterizing local variability (fractality) and dependence range of the spatial process of interest, the noise level, and the design of the region of interest.
Similar content being viewed by others
Notes
A Riesz basis of a Hilbert space H is a basis satisfying:
-
(i)
there exist constants \(C^{\prime }>C>0\) such that, for every sequence of scalars \({\left\{ c_{m}:m\in \mathbb{N}\right\} },\)
$$ C\sum_{m}\left| c_{m}\right| ^{2}\leq \left\| \sum_{m}c_{m}\phi_{m}\right\| _{H}^{2}\leq C^{\prime }\sum_{m}\left| c_{m}\right| ^{2},\quad \hbox {and} $$(6) -
(ii)
the vector space of finite sums \(\sum_{m}c_{m}\phi _{m}\) (on which (6) is tested) is dense in H.
-
(i)
References
Abramovich F, Sapatinas T, Silverman BW (1998) Wavelet thresholding via a Bayesian approach. J R Stat Soc B 60:725–749
Amiri M, Azimifar Z, Fieguth P (2007) Correlated non-linear wavelet shrinkage. In: Proceedings of the international conference on image processing (ICIP 2008), 12–15 Oct 2008, San Diego, pp 2348–2351
Angulo JM, Ruiz-Medina MD (1999) Multiresolution approximation to the stochastic inverse problem. Adv Appl Probab 31:1039–1057
Angulo JM, Bueso MC, Alonso FJ (2000) A study on sampling design for optimal prediction of space-time stochastic processes. Stoch Environ Res Risk Assess 14:412–427
Angulo JM, Ruiz-Medina MD, Alonso FJ, Bueso MC (2005) Generalized approaches to spatial sampling design. Environmetrics 16:523–534
Anh VV, Angulo JM, Ruiz-Medina MD, Tieng Q (1998) Long-range dependence and second-order intermittency of two dimensional turbulence. Environ Model Softw 13:233–238
Autin F (2008) Maxisets for μ-thresholding rules. Test 17:332–349
Bueso MC, Angulo JM, Alonso FJ (1998) A state-space-model approach to optimal spatial sampling design based on entropy. Environ Ecol Stat 5:29–44
Chipman HA, Kolaczyk ED, Mcculloch RE (1997) Adaptive Bayesian wavelet shrinkage. J Am Stat Assoc 92:1413–1421
Christakos G (2000) Modern spatiotemporal geostatistics. Oxford University Press, New York
Cressie N, Huang HC (1999) Classes of nonseparable, spatio-temporal stationary covariance functions. J Am Stat Assoc 352:3651–3685
Crouse M, Nowak R, Baraniuk R (1998) Wavelet-based statistical signal processing using hidden Markov models. IEEE Trans Signal Process 46:886–902
Daubechies I (1992) Ten lectures on wavelets. CBMS conference series in applied mathematics, vol 61. SIAM, Philadelphia
De Boor C, De Vore RA, Ron A (1993) On the construction of multivariate (pre) wavelets. Constr Approx 9:123–166
De Iaco S, Myers D, Posa D (2001) Space-time analysis using a general product-sum model. Stat Probab Lett 52:21–28
De Iaco S, Myers D, Posa D (2002) Nonseparable space-time covariance models: some parametric families. Math Geol 34:23–41
Donoho DL, Johnstone IM (1994a) Ideal spatial adaptation via wavelet shrinkage. Biometrika 81:425–455
Donoho DL, Johnstone IM (1994b) Minimax risk over l p -balls for l q -error. Probab Theory Relat Fields 99:277–303
Donoho DL, Johnstone IM (1995) Adapting to unknown smoothness via wavelet shrinkage. J Am Stat Assoc 90:1200–1224
Donoho DL, Johnstone IM (1998) Minimax estimation via wavelet shrinkage. Ann Stat 26:879–921
Gneiting T (2002) Stationary covariance functions for space-time data. J Am Stat Assoc 97:590–600
Guttorp P, Le ND, Sampson PD, Zidek JV (1993) Using entropy in the redesign of an environmental monitoring network. In: Patil GP, Rao CR (eds) Multivariate environmental statistics. Elsevier, Amsterdam, pp 175–202
Ivanov AV, Leonenko NN (2004) Asymptotic theory of nonlinear regression with long-range dependence. Math Methods Stat 13:153–178
Johnstone IM, Bernard W, Silverman B (2005) Empirical Bayes selection of wavelet thresholds. Ann Stat 33:1700–1752
Kerkyacharian PC, Picard D (2000) Thresholding algorithms, maxisets and well concentrated bases. Test 9:283–345
Kerkyacharian PC, Picard D (2002) Minimax or maxisets. Bernoulli 8:219–253
Kiriakidis PC, Journel AG (1999) Geostatistical space-time models: a review. Math Geol 31:651–684
Le ND, Zidek JV (2006) Statistical analysis of environmental space-time processes. Springer, New York
Liu J, Moulin P (2001) Complexity-regularized image denoising. IEEE Trans Image Process 10:841–851
Mallat SG (1989) Multiresolution approximations and wavelet orthonormal bases of \({L^{2}(\mathbb{R})}.\) Trans Am Math Soc 315:69–87
Meyer Y (1992) Wavelets and operators. Cambridge University Press, Cambridge
Romberg J, Choi H, Baraniuk R (2001) Bayesian tree-structured imaged modeling using wavelet-domain hidden Markov models. IEEE Trans Image Process 10:1056–1068
Ruiz-Medina MD, Angulo JM (2002) Spatio-temporal filtering using wavelets. Stoch Environ Res Risk Assess 16:241–266
Ruiz-Medina MD, Angulo JM (2007) Functional estimation of spatiotemporal heterogeneities. Environmetrics 18:775–792
Ruiz-Medina MD, Alonso FJ, Angulo JM, Bueso MC (2003) Functional stochastic modeling and prediction of spatio-temporal processes. J Geophys Res Atmos 108(D24):9003–9023. doi:10.1029/2003JD003416
Ruiz-Medina MD, Angulo JM, Anh VV (2008) Multifractality in space-time statistical models. Stoch Environ Res Risk Assess 22(S1):81–86
Stein M (2003) Space-time covariance functions. J Am Stat Assoc 100:310–321
Vidakovic B (1998a) Statistical modeling by wavelets. Wiley series in probability and statistics. Wiley, New York
Vidakovic B (1998b) Wavelet-based nonparametric Bayes methods. Practical nonparametric and semiparametric Bayesian statistics. Lecture Notes in Statistics, vol 133. Springer, New York, pp 133–155
Wikle CK, Cressie N (1999) A dimension-reduction approach to space-time Kalman filtering. Biometrika 86:815–829
Acknowledgments
The authors would like to thank the referees for their constructive comments and suggestions to improve the original manuscript. This work has been supported in part by projects MTM2009-13250 and MTM2009-13393 of the SGPI, and P08-FQM-3834 and P09-FQM-5052 of the Andalusian CICE, Spain.
Author information
Authors and Affiliations
Corresponding author
Appendix
Appendix
It well known that an orthonormal basis of wavelet functions of \({L^{2}({\mathbb{R}}^{d})},\) the space of square integrable functions defined on \({{\mathbb{R}}^{d}, \,d\in \mathbb{N}},\) provides a multiresolution analysis of this space. Thus, it allows the description of a L 2-signal at different resolution levels, from the macro-scale to the micro-scale, that is, from its global characteristics and asymptotic properties to its local variability properties (see, for example, De Boor et al. 1993; Daubechies 1992; Mallat 1989; Meyer 1992).
A multiresolution approximation of \({L^{2}( {\mathbb{R}}^{d}) }\) is defined as an increasing sequence \({\left\{ V_{j}:j\in \mathbb{Z}\right\} }\) of closed linear subspaces of \({L^{2}({\mathbb{R}}^{d}) }\) satisfying the following properties:
-
(i)
\(\bigcap_{-\infty }^{\infty }V_{j}=\left\{0\right\}, \bigcup_{-\infty }^{\infty }V_{j}\) is dense in \({L^{2}( {\mathbb{R}}^{d})};\)
-
(ii)
\(f\left({\mathbf x}\right) \in V_{0}\) iff \(f\left( 2{\mathbf x}\right) \in V_{j+1},\) for all \({{\mathbf x}\in {\mathbb{R}}^{d}}\) and \({j\in \mathbb{Z}};\)
-
(iii)
\(f\left({\mathbf x}\right) \in V_{0}\) iff \(f\left( {\mathbf x}-{\mathbf k}\right) \in V_{0},\) for all \({{\mathbf x}\in{\mathbb{R}}^{d}}\) and \({{\mathbf k}\in {\mathbb{Z}}^{d}},\) and
-
(iv)
there exists a function \(g\left( \cdot \right) \in V_{0}\) such that the sequence \({\left\{ g\left( {\mathbf x}- {\mathbf k}\right): {\mathbf k}\in {\mathbb{Z}}^{d}\right\} }\) is a Riesz basis of V 0.
Different approaches have been adopted in the construction of an orthonormal basis of wavelet functions associated with a multiresolution analysis of \({L^{2}({\mathbb{R}}^{d})}.\) In particular, from a Riesz basisFootnote 1 of V 0, an orthonormal basis of wavelet functions can be constructed using the spectral theory of \(L^{2}\left( [0,2\pi )^{d}\right)\)-functions (see, for example, Meyer 1992). Formally, the following decomposition of \({L^{2}({\mathbb{R}}^{d})}\) is obtained from an orthonormal basis of wavelet functions:
Thus, for a function \({f\in L^{2}({\mathbb{R}}^{d})},\) its projection on an orthonormal scaling basis of V 0 provides the draft of f, and its projections on the orthonormal wavelet bases of the spaces \({W_{j},\, j\in \mathbb{N}},\) generated by translations and contractions of the basic wavelet functions, reproduce local variability properties of f at different resolution levels \({j\in \mathbb{N}}.\) Note that, for every \({j \in \mathbb{N}},\)
Rights and permissions
About this article
Cite this article
Angulo, J.M., Madrid, A.E. & Ruiz-Medina, M.D. Entropy-based correlated shrinkage of spatial random processes. Stoch Environ Res Risk Assess 25, 389–402 (2011). https://doi.org/10.1007/s00477-010-0450-8
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00477-010-0450-8