Advertisement

The A-optimal regularization parameter in uniform Tykhonov-Phillips regularization — α-weighted BLE-

  • Jianqing Cai
  • Erik W. Grafarend
  • Burkhard Schaffrin
Part of the International Association of Geodesy Symposia book series (IAG SYMPOSIA, volume 127)

Abstract

Numerical tests have documented that the estimate \(\hat \xi \) of type BLUUE of the parameter vector ξ within a linear Gauss-Markov model {=E{y}, Σ y = D{y}} is not robust against outliers in the stochastic observation vector y. It is for this reason that we give up the postulate of unbiasedness, but keeping the set-up of a linear estimation \(\hat \xi = Ly\) of homogeneous type. Grafarend and Schaffrin (1993) as well as Schaffrin (2000) have systematically derived the best linear estimators of type homBLE (Best homogeneously Linear Estimation), S-homBLE and α-homBLE of the fixed effects ξ, which turn out to enhance the best linear uniformly unbiased estimator of type BLUUE, but suffer from the effect being biased. Here the regularization parameter in uniform Tykhonov-Phillips regularization (α-weighted BLE) is determined by minimizing the trace of the Mean Square Error matrix MSE α,s {\(\hat \xi \)} (A-optimal design) in the general case. In lieu of a case study, both model and estimators are tested and analyzed with numerical results computed from simulated direct observations of a random tensor of type strain rate.

Keywords

A-optimal design Tykhonov-Phillips regularization α-homBLE α-HAPS ridge estimator biased estimation Mean Square Error 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. Allen, D.M. (1971): Mean square error of prediction as a criterion for selecting variables, Technometrics 13 (1971) 469–481CrossRefGoogle Scholar
  2. Allen, D.M. (1974): The relationship between variable selection and data argumentation and a method for prediction, Technometrics 16 (1974) 125–127CrossRefGoogle Scholar
  3. Arslan, O. and Billor, B. (2000): Robust Liu estimator for regression based on an M-estimator, Journal of Applied Statistics 27 (2000) 39–47CrossRefGoogle Scholar
  4. Bapat, R.B. (1999): Linear Algebra and Linear Models, 2nd Edition, Springer, New York, Berlin, 1999Google Scholar
  5. Bouman, J. (1998): Quality of regularization methods,DEOS Report no. 98.2, Delft University Press 1998Google Scholar
  6. Chaturvedi, A. and Singh, S.P. (2000): Stein rule prediction of the composite target function in a general linear regression model, Statistical Papers 41 (2000) 359–367CrossRefGoogle Scholar
  7. D. R. Cox and N. Reid (2000): The Theory of the Design of Experiments, Chapman Hall/CRC, 2000Google Scholar
  8. Donatos, GS. and Michailidis, GC.(1990)• Small sample properties of ridge estimators with normal and non-normal disturbances, Commun. Statist.-Simula. 19 (1990) 935–950CrossRefGoogle Scholar
  9. Draper, R.N. and Van Nostrand, C.R. (1979): Ridge regression and James- Stein estimation: Review and comments, Technometrics 21 (1979) 451–466CrossRefGoogle Scholar
  10. Droge, B. (1993): On finite-sample properties of adaptative least squares regression estimates, Statistics 24 (1993) 181–203CrossRefGoogle Scholar
  11. Engels, J., et al. (1993): The geoid as an inverse problem to be regularized, Proceedings of the international conference “Inverse Problems: Principles and Applications in Geophysics, Technology, and Medicine” Anger, G et al. (eds). Akademie Verlag, Berlin 1993Google Scholar
  12. Engl, H. (1993): Regularization methods for the stable solution of inverse problems, Surveys on Math. for Industry 3 (1993) 371–143Google Scholar
  13. EL-Sayed, S.M. (1996): The sampling distribution of Ridge parameter estimator, The Egyptian Statistical Journal 40 (1996) 211–219Google Scholar
  14. Farebrother, R.W. (1975): The minimum mean square error linear estimator and ridge regression, Technometrics 17 (1975) 127–128CrossRefGoogle Scholar
  15. Farebrother, R.W. (1976): Further results on the mean square error of ridge regression, Journal of the Royal Statistical Society B.38 (1976) 248–250Google Scholar
  16. Farebrother, R.W. (1978): Partitioned ridge regression, Technometrics 20 (1978) 121–122Google Scholar
  17. Firinguetti, L. (1996): A simulation study of ridge regression in an incorrectly specified model, Estadistica 48 (1996) 115–154Google Scholar
  18. Firinguetti, L. and Rubio, H. (2000): A note on the moments of stochastic shrinkage parameters in ridge regression Commun. Statist.-Simula. 29 (3) (2000) 955–970CrossRefGoogle Scholar
  19. Gibbons, D.G(1981): A simulation study of some ridge estimators, Journal of the American Statistical Association 7 (1981) 131–139CrossRefGoogle Scholar
  20. Golub, G, Heath, M. and Wahba, G (1979): Generalized cross-validation as a methods for choosing a good ridge parameter, Technometrics 21 (1979) 215–223CrossRefGoogle Scholar
  21. Grafarend, E. and Schaffrin B. (1993): Ausgleichungsrechnung in linearen Modellen, Bibliographisches Institut Wissenschaftsverlag, Mannheim 1993Google Scholar
  22. Gui, Q. and Zhang, J. (1998a): Robust biased estimation and its applications in geodetic adjustments, Journal of Geodesy 72 (1998) 430–435CrossRefGoogle Scholar
  23. Gui, Q., Huang, W., and Zhang, J. (1998b): Robust universal ridge estimation, Arta geodetica et cartographica sinica 3 (1998)Google Scholar
  24. Gui, Q. and Liu, J. (2000): Biased estimation in Gauss-Markov model, AVN 3 (2000) 104–107Google Scholar
  25. Gui, Q., Yang, Y. and Guo, J. (2001): Biased estimation in Gauss-Markov model with Constraints, AVN 1 (2001) 28–30Google Scholar
  26. Gunst, R.F. and Mason R.L. (1977): Biased Estimation in Regression: An Evaluation Using Mean Squared Error, Journal of the American Statistical Association 72 (1977) 616–628CrossRefGoogle Scholar
  27. Gunst, R.F. and Mason, R.L. (1980): Regression Analysis and its Application, Marcel Dekker, New York 1980Google Scholar
  28. Hansen, P. (1992): Analysis of discrete ill-posed problems by means of the L-curve, SIAM Review 34 561–580CrossRefGoogle Scholar
  29. Hansen, P. (1993): The use of the L-curve in the regularization of discrete ill-posed problems, SIAM J. Sci. Comput. 14 (1993) 1487–1503Google Scholar
  30. Hanke, M. and Hansen, P. (1993): Regularization methods for large scale problems, Surveys on Math. for Industry 3 (1993) 253–315Google Scholar
  31. Hemmerle, W.J. (1975): An explicit solution for generalized ridge regression, Technometrics 17 (1975) 309–314CrossRefGoogle Scholar
  32. Hemmerle, W.J. and Brantle, T.F. (1978): Explicit and constrained generalized ridge estimation, Technometrics 20 (1978) 109–120CrossRefGoogle Scholar
  33. Hocking, R.R. (1976): The analysis and selection of vari- ables in linear regression, Biometrics 32 (1976) 1–49CrossRefGoogle Scholar
  34. Hoerl, A. E. and Kennard, R.W. (1970a): Ridge regression: Biased estimation for nonorthogonal problems, Technometrics 12 (1970) 55–67CrossRefGoogle Scholar
  35. Hoerl, A.E. and Kennard, R.W. (1970b): Ridge regression: Applications to nonorthogonal problems, Technometrics 12 (1970) 69–82CrossRefGoogle Scholar
  36. Hoerl, A.E., Kennard, R. W. and Baldwin, K.F. (1975): Ridge regression: Some simulations Analysis, Communications in Statistics 4 (1975) 105–123.CrossRefGoogle Scholar
  37. Hoerl, R. W. (1985): Ridge Analysis 25 Years Later, The American Statistician 39 (1985) 186–192Google Scholar
  38. Hoerl, R. W., Schuenemeyer, J. H. and Hoerl, A. E. (1985): A simulation of biased estimation and subset selection regression techniques, Technometrics 28 369–380CrossRefGoogle Scholar
  39. Ilk, K. H. (1986): On the regularization of ill-posed problems, Proc. Int. symp. Figure and Dynamics of the Earth, Moon and Planets, Prague 1986, 365–383Google Scholar
  40. Kacirattin, S. Sakailoglu, S. and Akdeniz, F. (1998): Mean squared error comparisons of the modified ridge regression estimator and the restricted ridge regression 27 (1998) 131–138Google Scholar
  41. Kakkuri J. and Chen R. (1992): On Horizontal Crustal Strain in Finland. Bull. Géod. 66 (1992) 12–20CrossRefGoogle Scholar
  42. Lawless, J.F. and Wang, P. (1976): A simulation study of ridge and other regression estimations. Communications in Statistics AS (1976) 307–322Google Scholar
  43. Liski, E. P., et al. (2001) Topic in Optimal Design, Springer, New York and Berlin, 2001Google Scholar
  44. Liu, K (1993): A new class of biased estimates in linear regression, Communications in Statistics A22 (1993) 393–402Google Scholar
  45. Louis, A. K. and Maass, P. (1990): A mollifier method for linear operator equation of the first kind, Inverse Problems, 6 (1990) 427–440 Google Scholar
  46. Lowerre, J. M. (1974): On the mean square error of parameter estimates for some biased estimators, Technometrics 16 (1974) 461–464CrossRefGoogle Scholar
  47. Mallows, C.L. (1973): Some comments on Cp, Technometrics 15 (1973) 661–675Google Scholar
  48. Markiewicz, A. (1996): Characterization of general ridgeestimators, Statistics Probability Letters 27 145–148CrossRefGoogle Scholar
  49. Marquardt, W.D. (1970): Generalized inverses, ridge regression, biased linear estimation, and nonlinear estimation, Technometrics 12 (1970) 591–612Google Scholar
  50. Marquardt, W.D. (1974): Discussion No.2, Technometrics 16 (1974) 189–192Google Scholar
  51. Marquardt, W.D. and Snee, R.D. (1975): Ridge regression in practice, The American Statistician 29 (1975) 3–20Google Scholar
  52. Mayer, L.S. and Willke, T.A. (1973): On biased estimation in linear models, Technometrics 15 (1973) 497–508CrossRefGoogle Scholar
  53. McDonald, GC. and Galameau, D.I. (1975): A Monte Carlo evaluation of some ridge-type estimators, Journal of the American Statistical Association 70 (1975) 407–416CrossRefGoogle Scholar
  54. Nomura, M. (1988): On exact small sample properties of ordinary ridge estimators, Economics letters 27 (1988) 251–255CrossRefGoogle Scholar
  55. Nomura, M. (1998): On exact small sample properties of the minimax generalized ridge regression estimators, Commun. Statist.-Simula. 17 (1988) 1213–1229Google Scholar
  56. Ohtani, K. (1986): On small’ sample properties of the almost unbiased generalized ridge estimator, Commun. Statist.-Simula. 15 (1986) 1571–1578Google Scholar
  57. Ohtani, K. (1998): An MSE comparison of the restricted Stein-rule and minimum mean squared error estimators in regression, Sociedad de Estadistica e Investigacion Operativa Test 7 (1998) 361–376Google Scholar
  58. Phillips, D.L. (1962): A technique for the numerical solution of certain integral equations of the first kind, J. Ass. Comput. Mach. 9 (1962) 84–96CrossRefGoogle Scholar
  59. Pazman, A. (1986): Foundations of Optimum Experimental Design, Kluwer Academic, Dordrecht, 1986Google Scholar
  60. Pukelsheim, F. (1993): Optimal Design of Experiments, Wiley, New York 1993Google Scholar
  61. Rao, C.R. (1975): Simultaneous estimation of parameters in different linear models and applications to biometric problems, Biometrics 31 (1975) 545–554CrossRefGoogle Scholar
  62. Rao, C.R. (1976): Estimation of parameters in a linear model (The 1975 Wald Memorial Lectures), The Annals of Statistics 4 (1976) 1023–1037CrossRefGoogle Scholar
  63. Schaffrin, B., Heidenreich, E. and Grafarend, E. (1977): A representation of the standard gravity field, Manuscipta geodaetica 2 (1977) 135–174Google Scholar
  64. Schaffrin, B. and Middel B.(1990): Robust predictors and an alternative iteration scheme for ill-posed problem, in Geophysical Data Inversion-Methods and Applications, A. Vogel, Ch. O. Ofoegbu, R. Gorenflo and Bj. Ursin (Eds.), Vieweg, Braunschweig (1990) 33–51CrossRefGoogle Scholar
  65. Schaffrin, B. (1995): A comparision of inverse techniques: regularization, weight estimation and homBLUP, IUGG General Association 1995Google Scholar
  66. Schaffrin, B (2000): Minimum mean square error adjustment, Part I: The empirical BLE and the repro-BLE for direct observations, Journal of the Geodetic Society of Japan 46 (2000) 21–30Google Scholar
  67. Shalabh (1998): Improved estimation in measurement error models through Stein rule procedure, Journal of Multivariate Analysis 67 (1998) 35–48CrossRefGoogle Scholar
  68. Smith, G and Campbell, F. (1980): A Critique of Some Ridge Regression Methods, Journal of the American Statistical Association 75 (1980) 74–81CrossRefGoogle Scholar
  69. Srivastava, V.K. et al. (1983): Some properties of the distribution of an operational ridge estimator, Metron 30 (1983) 227–237Google Scholar
  70. Theobald, C.M. (1974): Generalizations of mean square error applied to ridge regression, Journal of the Royal Statistical Society B.36 (1974) 103–106Google Scholar
  71. Tykhonov, A.N. (1963): The regularization of incorrectly posed problem, Soviet Math. Doklady 4 (1963) 1624–1627Google Scholar
  72. Tykhonov, A.N. et al. (1977): A variational method of regularization in the adjustment of free geodetic nets, Geodesy, Mapping and Photogrammetry 19 (1977) 127–139Google Scholar
  73. Tykhonov, A.N. and Arsenin, V.Y. (1977): Solutions of Ill-Posed Problems, Wiley, New York 1977Google Scholar
  74. Vinod, H.D. and Ullah, A. (1981): Recent advances in re- gression methods, Marcel Dekker, New York 1981Google Scholar
  75. Xu, P. (1992a): Determination of surface gravity anomalies using gradiometric observables, Geophys. J. Int. 110 (1992) 321–332Google Scholar
  76. Xu, P. (1992b): The value of minimum norm estimation of geopotential fields, Geophys.J.Int. 111 (1992) 170–178CrossRefGoogle Scholar
  77. Xu, P. (1998): Truncated SVD methods for discrete linear ill-posed problems, Geophys.J.Int. 135 (1998) 505–514CrossRefGoogle Scholar
  78. Xu, P. and Rummel, R(1994a): A simulation study of smoothness methods in recovery of regional gravity fields, Geophys.J.Int. 117 (1994) 472–486CrossRefGoogle Scholar
  79. Xu, P. and Rummel, R(1994b): Generalized ridge regression with applications in determination of potential fields, manuscripta geodaetica 20 (1994) 8–20Google Scholar
  80. Wang, Y. and Xiao, T. (2001): Fast realization algorithms for determining regularization parameters in linear inverse problems, Inverse Problems 17 (2001) 281–291CrossRefGoogle Scholar
  81. Wenchenko, E. (2000): Estimation of the signal-to-noise in the linear regression model, Statistical Papers 41 327–343CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2004

Authors and Affiliations

  • Jianqing Cai
    • 1
  • Erik W. Grafarend
    • 1
  • Burkhard Schaffrin
    • 2
  1. 1.Department of Geodesy and GeoInformaticsUniversity of StuttgartStuttgartGermany
  2. 2.Department of Civil and Environmental Engineering and Geodetic ScienceThe Ohio State UniversityColumbusUSA

Personalised recommendations