Bayesian Multivariate Spatial Interpolation with Data Missing by Design

Authors: Le, Nhu D.¹; Sun, Weimin²; Zidek, James V.³

Source: Journal of the Royal Statistical Society: Series B (Statistical Methodology), Volume 59, Number 2, 1997 , pp. 501-510(10)

Publisher: Wiley-Blackwell

Buy & download fulltext article:

Price: $48.00 plus tax (Refund Policy)

Abstract:

In a network of s_g sites, responses like levels of airborne pollutant concentrations may be monitored over time. The sites need not all measure the same set of response items and unmeasured items are considered as data missing by design. We propose a hierarchical Bayesian approach to interpolate the levels of, say, k responses at s_u other locations called ungauged sites and also the unmeasured levels of the k responses at the gauged sites. Our method involves two steps. First, when all hyperparameters are assumed to be known, a predictive distribution is derived. In turn, an interpolator, its variance and a simultaneous interpolation region are obtained. In step two, we propose the use of an empirical Bayesian approach to estimate the hyperparameters through an EM algorithm. We base our theory on a linear Gaussian model and the relationship between a multivariate normal and matrix T-distribution. Our theory allows us to pool data from several existing networks that measure different subsets of response items for interpolation.

Keywords: Bayesian Interpolation; Co-Kriging; Matrix T-Distribution; Predictive Distribution; Spatial Interpolation

Document Type: Original article

DOI: http://dx.doi.org/10.1111/1467-9868.00081

Affiliations: 1: University of British Columbia, Vancouver, and BC Cancer Agency, Vancouver, CA, 2: Statistics Canada, Ottawa, CA, 3: University of British Columbia, Vancouver, CA

Publication date: 1997-01-01