Regularized kriging as a generalization of simple, universal, and bayesian kriging

  • J. M. MatíasEmail author
  • W. González-Manteiga
Original Paper


In this article the properties of regularized kriging (RK) are studied. RK is obtained as a result of relaxing the universal kriging (UK) non-bias condition by using the support vectors methodology. More specifically, we demonstrate how RK is a continuum of solutions in function of the regularizing parameter, which includes as particular and extreme cases, simple kriging (SK) and UK, and as an intermediate case, Bayesian kriging (BK). Likewise, expressions are obtained for the mean, variance and mean squared error (MSE), as also the expression for the corresponding estimator of the coefficients of the mean. Finally, we investigate the relationship between RK and the support vector machines. By means of simulations we compare the MSE for RK with those for BK and UK, for different association models, for different levels of noise, and for differently sized mean coefficients. The RK results prove to be an improvement on the UK and BK results, and, moreover, these improvements are proportionally greater for greater levels of noise.


Bias-variance trade-off Cauchy class Kernel methods Matérn class Kriging Support vector machines 



The research of W. González-Manteiga was supported by the Spanish Ministry of Science and Technology under Project BFM2002-03213. The authors wish to express their thanks to the three anonymous referees whose comments and suggestions greatly improved this article.


  1. Christakos G (1992) Random field models in earth sciences. Academic, San DiegoGoogle Scholar
  2. Cressie N (1993) Statistics for spatial data. Wiley, New YorkGoogle Scholar
  3. De Oliveira V, Kedem B, Short DA (1997) Bayesian prediction of transformed gaussian random fields. J Am Stat Assoc 92:1422–1433CrossRefGoogle Scholar
  4. Diggle PJ, Tawn JA, Moyeed RA (1998) Model-based geostatistics. Appl Stat 47:299–350Google Scholar
  5. Girosi F, Jones M, Poggio T (1995) Regularization theory and neural networks architectures. Neural Comput 7:219–269CrossRefGoogle Scholar
  6. Gneiting T, Schlather M (2004) Stochastic models that separate fractal dimension and the Hurst effect. SIAM Rev 46:269–282CrossRefGoogle Scholar
  7. Gotway CA, Cressie N (1993) Improved multivariate prediction under a general linear model. J Multivariate Anal 45:56–72CrossRefGoogle Scholar
  8. Handcock MS, Stein ML (1993) A Bayesian analysis of kriging. Technometrics 35:403–410CrossRefGoogle Scholar
  9. Hoerl AE, Kennard RW (1970) Ridge regression: biased estimation for nonorthogonal problems. Technometrics 12:55–67CrossRefGoogle Scholar
  10. Kitanidis PK (1986) Parameter uncertainty in estimation of spatial functions: Bayesian analysis. Water Resour Res 22:499–507CrossRefGoogle Scholar
  11. Le ND, Zidek JV (1992) Interpolation with uncertain covariances: a Bayesian alternative to Kriging. J Multivariate Anal 43:351–374CrossRefGoogle Scholar
  12. MacKay DJC (1998) Introduction to Gaussian processes. In: Bishop CM (ed) Neural networks and machine learning. Proceedings of the NATO advanced Study Institute on Generalization in Neural Networks and Machine Learning. Springer, Berlin Heidelberg New York, pp 133–165Google Scholar
  13. Matheron G (1973) The intrinsic random functions and their applications. Adv Appl Probab 5:439–468CrossRefGoogle Scholar
  14. Matías JM, González-Manteiga W (2003) Regularized kriging: the support vector method applied to kriging. In: Proceedings of the joint 13th international conference on artificial neural networks and 10th international conference on neural information processing, Istambul. Springer, Berlin Heidelberg New York, LCNS 2714, pp 209–216Google Scholar
  15. Matías JM, Vaamonde A, Taboada J, González-Manteiga W (2004a) Comparison of kriging and neural networks with application to the exploitation of a slate mine. Math Geol 36: 485–508CrossRefGoogle Scholar
  16. Matías JM, Vaamonde A, Taboada J, González-Manteiga W (2004b) Support vector machines and gradient boosting for graphical estimation of a slate deposit. J Stochast Environ Res Risk Assess 18:309–323CrossRefGoogle Scholar
  17. Matías JM, González-Manteiga W, Francisco-Fernández M, Taboada J, Ordóñez C (2004c) Variography for model selection in local polynomial regression with spatial data. In: Proceedings of the conference on computational and mathematical methods on science and engineering, CMMSE-2004, Uppsala, pp 226–235Google Scholar
  18. Omre H (1987) Bayesian kriging—merging observations and qualified guesses in kriging. Math Geol 19:25–39CrossRefGoogle Scholar
  19. Omre H, Halvorsen KB (1989) The Bayesian bridge between simple and universal kriging. Math Geol 21:767–786CrossRefGoogle Scholar
  20. Schölkopf B, Smola AJ (2002) Learning with Kernels. MIT Press, CambridgeGoogle Scholar
  21. Smola A, Frieß J, Scholköpf B (1999) Semiparametric support vector and linear programming machines. In: Kearns MS, Solla SA, Cohn DA (eds) Advances in neural information processing systems. MIT Press, Cambridge, pp 585–591Google Scholar
  22. Stein ML (1999) Interpolation of spatial data. Springer, Berlin Heidelberg New YorkGoogle Scholar
  23. Vapnik V (1998) Statistical learning theory. Wiley, New YorkGoogle Scholar
  24. Williams CKI (1998) Prediction with gaussian processes: from linear regression to linear prediction and beyond. In: Jordan MI (ed) Learning and inference in graphical models. Kluwer, Dordrecht, 599–621Google Scholar

Copyright information

© Springer-Verlag 2005

Authors and Affiliations

  1. 1.Department of StatisticsUniversity of VigoVigoSpain
  2. 2.Department of StatisticsUniversity of Santiago de CompostelaSantiago de CompostelaSpain

Personalised recommendations