ASYMPTOTIC THEORY FOR KRIGING WITH ESTIMATED PARAMETERS AND ITS APPLICATION TO NETWORK DESIGN ( PRELIMINARY VERSION )

A second-order expansion is established for predictive distributions in Gaussian processes with estimated covariances. Particular focus is on estimating quantiles of the predictive distribution and their subsequent application to prediction intervals. Two basic approaches are considered, (a) a “plug-in” approach using the restricted maximum likelihood estimate of the covariance parameters, (b) a Bayesian approach using general priors. Calculations of “coverage probability bias” show that the Bayesian approach is superior in the tails of the predictive distributions, regardless of the prior. However they also imply the existence of a “matching prior” for which the second-order coverage probability bias vanishes. Previously suggested frequentist corrections do not have this property, but we use our results to suggest a new frequentist approach that does. We also compute the expected length of a Bayesian prediction interval, suggesting that this might be used as a design criterion combining recent “estimative” and “predictive” approaches to network design. A surprising parallel emerges with the recent two-stage estimative-predictive approach of Zhu and Stein.

[1]  N. L. Johnson,et al.  Multivariate Analysis , 1958, Nature.

[2]  Malay Ghosh,et al.  Some remarks on noninformative priors , 1995 .

[3]  J. Zidek,et al.  Multivariate spatial interpolation and exposure to air pollutants , 1994 .

[4]  L. Tierney,et al.  Accurate Approximations for Posterior Moments and Marginal Densities , 1986 .

[5]  Daniel R. Jeske,et al.  Mean Squared Error of Estimation or Prediction under a General Linear Model , 1992 .

[6]  D. Harville.  Bayesian inference for variance components using only error contrasts , 1974 .

[7]  M. Stein,et al.  Spatial sampling design for prediction with estimated parameters , 2006 .

[8]  K. Mardia,et al.  Maximum likelihood estimation of models for residual covariance in spatial regression , 1984 .

[9]  N. Cressie,et al.  Mean squared prediction error in the spatial linear model with estimated covariance parameters , 1992 .

[10]  V. D. Oliveira,et al.  Objective Bayesian Analysis of SpatiallyCorrelated , 2000 .

[11]  Markus Abt.  Estimating the Prediction Mean Squared Error in Gaussian Stochastic Processes with Exponential Correlation Structure , 1999 .

[12]  J. Zidek,et al.  Interpolation with uncertain spatial covariances: a Bayesian alternative to Kriging , 1992 .

[13]  George Casella,et al.  Implementing matching priors for frequentist inference , 2003 .

[14]  M. Stein,et al.  A Bayesian analysis of kriging , 1993 .

[15]  David R. Cox,et al.  Prediction and asymptotics , 1996 .

[16]  P. van den Driessche,et al.  Asymptotic expansion of integrals occurring in linear wave theory , 1972, Bulletin of the Australian Mathematical Society.

[17]  David R. Cox,et al.  Prediction Intervals and Empirical Bayes Confidence Intervals , 1975, Journal of Applied Probability.

[18]  J. Berger,et al.  Objective Bayesian Analysis of Spatially Correlated Data , 2001 .