Abstract
The truncated singular value decomposition may be used to find the solution of linear discrete ill-posed problems in conjunction with Tikhonov regularization and requires the estimation of a regularization parameter that balances between the sizes of the fit to data function and the regularization term. The unbiased predictive risk estimator is one suggested method for finding the regularization parameter when the noise in the measurements is normally distributed with known variance. In this paper we provide an algorithm using the unbiased predictive risk estimator that automatically finds both the regularization parameter and the number of terms to use from the singular value decomposition. Underlying the algorithm is a new result that proves that the regularization parameter converges with the number of terms from the singular value decomposition. For the analysis it is sufficient to assume that the discrete Picard condition is satisfied for exact data and that noise completely contaminates the measured data coefficients for a sufficiently large number of terms, dependent on both the noise level and the degree of ill-posedness of the system. A lower bound for the regularization parameter is provided leading to a computationally efficient algorithm. Supporting results are compared with those obtained using the method of generalized cross validation. Simulations for two-dimensional examples verify the theoretical analysis and the effectiveness of the algorithm for increasing noise levels, and demonstrate that the relative reconstruction errors obtained using the truncated singular value decomposition are less than those obtained using the singular value decomposition.
Similar content being viewed by others
Notes
A box plot is a visual representation of summary statistics for a given sample. Horizontal lines of each plotted box represent the \(75\%\), \(50\%\) (median), and \(25\%\) quantiles, with outliers plotted as individual crosses or points.
References
Abascal, J.-F.P.J., Arridge, S.R., Bayford, R.H., Holder, D.S.: Comparison of methods for optimal choice of the regularization parameter for linear electrical impedance tomography of brain function. Physiol. Meas. 29, 1319–1334 (2008)
Aster, R.C., Borchers, B., Thurber, C.H.: Parameter Estimation and Inverse Problems, 2nd edn. Elsevier, Amsterdam (2013)
Bakushinskii, A.B.: Remarks on choosing a regularization parameter using the quasi-optimality and ratio criterion. USSR Comput. Math. Math. Phys. 24(4), 181182 (1984)
Bauer, F., Lukas, M.A.: Comparing parameter choice methods for regularization of ill-posed problems. Math. Comput. Simul. 81, 1795–1841 (2011)
Björck, A.: Numerical Methods for Least Squares Problems. Society for Industrial and Applied Mathematics, Philadelphia (1996)
Chung, J., Nagy, J.G., O’Leary, D.P.: A weighted GCV method for Lanczos hybrid regularization. Electron. Trans. Numer. Anal. 28, 149–167 (2008)
Drineas, P., Mahoney, M.W.: RandNLA: randomized numerical linear algebra. Commun. ACM 59, 80–90 (2016)
Drineas, P., Mahoney, M.W., Muthukrishnan, S., Sarlós, T.: Faster least squares approximation. Numer. Math. 117, 219–249 (2011)
Fenu, C., Reichel, L., Rodrigues, G., Sadok, H.: GCV for Tikhonov regularization by partial SVD. BIT Numer. Math. 57, 1019–1039 (2017)
Gazzola, S., Hansen, P.C., Nagy, J.G.: IR tools: a MATLAB package of iterative regularization methods and large-scale test problems. Numer. Algorithms. https://doi.org/10.1007/s11075-018-0570-7 (2018)
Golub, G.H., Heath, M., Wahba, G.: Generalized cross-validation as a method for choosing a good ridge parameter. Technometrics 21, 215–223 (1979)
Golub, G.H., van Loan, C.F.: Matrix Computations, 3rd edn. Johns Hopkins Press, Baltimore (1996)
Gower, R.M., Richtárik, P.: Randomized iterative methods for linear systems. SIAM J. Matrix Anal. Appl. 36, 1660–1690 (2015)
Hämäläinen, K., Harhanen, L., Kallonen, A., Kujanpää, A., Niemi, E., Siltanen, S.: Tomographic X-ray data of a walnut. arXiv:1502.04064 (2015)
Hämarik, U., Palm, R., Raus, T.: On minimizationstrategies for choice of the regularization parameter in ill-posed problems. Numer. Funct. Anal. Optim. 30, 924–950 (2009)
Hämarik, U., Palm, R., Raus, T.: A family of rules for parameter choice in Tikhonov regularization of ill-posed problems with inexact noise level. J. Comput. Appl. Math. 236, 2146–2157 (2012)
Hanke, M., Hansen, P.C.: Regularization methods for large-scale problems. Surv. Math. Ind. 3, 253–315 (1993)
Hansen, J.K., Hogue, J.D., Sander, G.K., Renaut, R.A., Popat, S.C.: Non-negatively constrained least squares and parameter choice by the residual periodogram for the inversion of electrochemical impedance spectroscopy data. J. Comput. Appl. Math. 278, 52–74 (2015)
Hansen, P.C.: The discrete Picard condition for discrete ill-posed problems. BIT Numer. Math. 30, 658–672 (1990)
Hansen, P.C.: Regularization tools—a Matlab package for analysis and solution of discrete ill-posed problems. Numer. Algorithms 46, 189–194 (1994)
Hansen, P.C.: Rank-Deficient and Discrete Ill-Posed Problems. Society for Industrial and Applied Mathematics, Philadelphia (1998)
Hansen, P.C.: Discrete Inverse Problems. Society for Industrial and Applied Mathematics, Philadelphia (2010)
Hofmann, B.: Regularization for Applied Inverse and Ill-posed Problems: A Numerical Approach, Teubner-Texte zur Mathematik. Teubner, B.G., Berlin (1986)
Levin, E., Meltzer, A.Y.: Estimation of the regularization parameter in linear discrete ill-posed problems using the Picard parameter. SIAM J. Sci. Comput. 39, A2741–A2762 (2017)
Lin, Y., Wohlberg, B., Guo, H.: UPRE method for total variation parameter selection. Signal Process. 90, 2546–2551 (2010)
Mahoney, M.W.: Randomized algorithms for matrices and data. Found. Trends® Mach. Learn. 3, 123–224 (2011)
Mead, J.L., Renaut, R.A.: A Newton root-finding algorithm for estimating the regularization parameter for solving ill-conditioned least squares problems. Inverse Probl. 25, 025002 (2009)
Meng, X., Saunders, M.A., Mahoney, M.W.: LSRN: a parallel iterative solver for strongly over- or underdetermined systems. SIAM J. Sci. Comput. 36, C95–C118 (2014)
Morozov, V.A.: On the solution of functional equations by the method of regularization. Sov. Math. Dokl. 7, 414–417 (1966)
Renaut, R.A., Horst, M., Wang, Y., Cochran, D., Hansen, J.: Efficient estimation of regularization parameters via downsampling and the singular value expansion. BIT Numer. Math. 57, 499–529 (2017)
Renaut, R.A., Vatankhah, S., Ardestani, V.E.: Hybrid and iteratively reweighted regularization by unbiased predictive risk and weighted GCV for projected systems. SIAM J. Sci. Comput. 39, B221–B243 (2017)
Rokhlin, V., Tygert, M.: A fast randomized algorithm for overdetermined linear least-squares regression. Proc. Natl. Acad. Sci. 105, 13212–13217 (2008)
Stein, C.M.: Estimation of the mean of a multivariate normal distribution. Ann. Stat. 9, 1135–1151 (1981)
Taroudaki, V., O’Leary, D.P.: Near-optimal spectral filtering and error estimation for solving ill-posed problems. SIAM J. Sci. Comput. 37, A2947–A2968 (2015)
Toma, A., Sixou, B., Peyrin, F.: Iterative choice of the optimal regularization parameter in tv image restoration. Inverse Probl. Imaging 9, 1171 (2015)
Vatankhah, S., Ardestani, V.E., Renaut, R.A.: Automatic estimation of the regularization parameter in 2D focusing gravity inversion: application of the method to the Safo manganese mine in the northwest of Iran. J. Geophys. Eng. 11, 045001 (2014)
Vatankhah, S., Ardestani, V.E., Renaut, R.A.: Application of the \(\chi ^2\) principle and unbiased predictive risk estimator for determining the regularization parameter in 3-D focusing gravity inversion. Geophys. J. Int. 200, 265–277 (2015)
Vatankhah, S., Renaut, R.A., Ardestani, V.E.: 3-D projected \(\ell _1\) inversion of gravity data using truncated unbiased predictive risk estimator for regularization parameter estimation. Geophys. J. Int. 210, 1872–1887 (2017)
Vatankhah, S., Renaut, R.A., Ardestani, V.E.: A fast algorithm for regularized focused 3-D inversion of gravity data using the randomized SVD. Geophysics (2018)
Vatankhah, S., Renaut, R.A., Ardestani, V.E.: Total variation regularization of the 3-D gravity inverse problem using a randomized generalized singular value decomposition. Geophys. J. Int. 213, 695–705 (2018)
Vogel, C.: Computational Methods for Inverse Problems. Society for Industrial and Applied Mathematics, Philadelphia (2002)
Author information
Authors and Affiliations
Corresponding author
Additional information
Communicated by Lothar Reichel.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rosemary Renaut acknowledges the support of NSF Grant DMS 1418377: “Novel Regularization for Joint Inversion of Nonlinear Problems”.
Rights and permissions
About this article
Cite this article
Renaut, R.A., Helmstetter, A.W. & Vatankhah, S. Unbiased predictive risk estimation of the Tikhonov regularization parameter: convergence with increasing rank approximations of the singular value decomposition. Bit Numer Math 59, 1031–1061 (2019). https://doi.org/10.1007/s10543-019-00762-7
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10543-019-00762-7