James V. Zidek

Optimal monitoring network designs

The selection of a monitoring network is formulated as a decision problem whose solutions would then be optimal. The theory is applied where the underlying field has a multivariate normal probability structure.

Interpolation with uncertain spatial covariances: a Bayesian alternative to Kriging

In this paper a Bayesian alternative to Kriging is developed. The latter is an important tool in geostatistics. But aspects of environmetrics make it less suitable as a tool for interpolating spatial random fields which are observed successively over time. The theory presented here permits temporal (and spatial) modeling to be done in a convenient and flexible way. At the same time model misspecifications, if any, can be corrected by additional data if and when it becomes available, and past data may be used in a systematic way to fit model parameters. Finally, uncertainty about model parameters is represented in the (posterior) distributions, so unrealistically small credible regions for the interpolants are avoided. The theory is based on the multivariate normal and related distributions, but because of the hierarchical prior models adopted, the results would seem somewhat robust with respect to the choice of these distributions and associated hyperparameters.

A Comparison of n Estimators for the Binomial Distribution

Abstract Estimating the number, n, of trials, given a sequence of independent success counts obtained by replicating the n-trial experiment is less studied and a considerably harder problem than estimating p for the binomial distribution. Both the method of moments and the maximum likelihood estimators of n are considered. When p is unknown and small, while n is large, both estimators become highly unstable. Stabilized versions of these estimators are proposed, and their performance is described in terms of results obtained by simulation.

Designing and integrating composite networks for monitoring multivariate Gaussian pollution fields

Networks of ambient monitoring stations are used to monitor environmental pollution fields such as those for acid rain and air pollution. Such stations provide regular measurements of pollutant concentrations. The networks are established for a variety of purposes at various times so often several stations measuring different subsets of pollutant concentrations can be found in compact geographical regions. The problem of statistically combining these disparate information sources into a single ‘network’ then arises. Capitalizing on the efficiencies so achieved can then lead to the secondary problem of extending this network. The subject of this paper is a set of 31 air pollution monitoring stations in southern Ontario. Each of these regularly measures a particular subset of ionic sulphate, sulphite, nitrite and ozone. However, this subset varies from station to station. For example only two stations measure all four. Some measure just one. We describe a Bayesian framework for integrating the measurements of these stations to yield a spatial predictive distribution for unmonitored sites and unmeasured concentrations at existing stations. Furthermore we show how this network can be extended by using an entropy maximization criterion. The methods assume that the multivariate response field being measured has a joint Gaussian distribution conditional on its mean and covariance function. A conjugate prior is used for these parameters, some of its hyperparameters being fitted empirically.

Simultaneous Estimation of the Means of Independent Poisson Laws

Abstract A procedure is derived for simultaneously estimating the means of Poisson populations using independent samples. The total loss is the sum of component, standardized squared error losses. The derived procedure is minimax, generalized Bayes, admissible relative to a large class, and better than using the sample means as estimators of the corresponding population means. The latter procedure is compared with the derived procedure, both numerically and asymptotically.

Bayesian Multivariate Spatial Interpolation with Data Missing by Design

SUMMARY In a network of sg 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 su 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.

CAUSALITY, MEASUREMENT ERROR AND MULTICOLLINEARITY IN EPIDEMIOLOGY

SUMMARY This paper demonstrates that measurement error can conspire with multicollinearity among explanatory variables to mislead an investigator. A causal variable measured with error may be overlooked and its significance transferred to a surrogate. The latter’s significance can then be entirely spurious, in that controlling it will not predictably change the response variable. In epidemiological research, such a response may be a health outcome. The phenomenon we study is demonstrated through simulation experiments involving nonlinear regression models. Also, the paper presents the measurement error problem in a theoretical setting. The paper concludes by echoing the familiar warning against making conclusions about causality from a multiple regression analysis, in this case because of the phenomenon presented in the paper.

Asymptotic properties of maximum weighted likelihood estimators

The relevance weighted likelihood method was introduced by Hu and Zidek (Technical Report No. 161, Department of Statistics, The University of British Columbia, Vancouver, BC, Canada, 1995) to formally embrace a variety of statistical procedures for trading bias for precision. Their approach combines all relevant information through a weighted version of the likelihood function. The present paper is concerned with the asymptotic properties of a class of maximum weighted likelihood estimators that contains those considered by Hu and Zidek (Technical Report No. 161, Department of Statistics, The University of British Columbia, Vancouver, BC, Canada, 1995, in: Ahmed, S.E. Reid, N. (Eds.), Empirical Bayes and Likelihood Inference, Springer, New York, 2001, p. 211). Our results complement those of Hu (Can. J. Stat. 25 (1997) 45). In particular, we invoke a different asymptotic paradigm than that in Hu (Can. J. Stat. 25 (1997) 45). Moreover, our adaptive weights are allowed to depend on the data.

Bayesian Spatial Prediction of Random Space-Time Fields With Application to Mapping PM2.5 Exposure

This article presents a multivariate spatial prediction methodology in a Bayesian framework. The method is especially suited for use in environmetrics, where vector-valued responses are observed at a small set of ambient monitoring stations “(gauged sites)” at successive time points. However, the stations may have varying start-up times so that the data have a “staircase” pattern (“monotone” pattern in the terminology of Rubin and Shaffer). The lowest step corresponds to the newest station in the monitoring network. We base our approach on a hierarchical Bayes prior involving a Gaussian generalized inverted Wishart model. For given hyperparameters, we derive the predictive distribution for currently gauged sites at times before their start-up when no measurements were taken. The resulting predictive distribution is a matric t distribution with appropriate covariance parameters and degrees of freedom. We estimate the hyperparameters using the method of moments (MOM) as an easy-to-implement alternative to the more complex EM algorithm. The MOM in particular gives exact parameter estimates and involves less cumbersome calculations than the EM algorithm. Finally, we obtain the predictive distribution for unmeasured responses at “ungauged” sites. The results obtained here allow us to pool the data from different sites that measure different pollutants and also to treat cases where the observed data monitoring stations have a monotonic “staircaserldquo; structure. We demonstrate the use of this methodology by mapping PM2.5 fields for Philadelphia during the period of May 1992 to September 1993. Large amounts of data missing by design make this application particularly challenging. We give empirical evidence that the method performs well.

Modeling Nonstationary Processes Through Dimension Expansion

In this article, we propose a novel approach to modeling nonstationary spatial fields. The proposed method works by expanding the geographic plane over which these processes evolve into higher-dimensional spaces, transforming and clarifying complex patterns in the physical plane. By combining aspects of multidimensional scaling, group lasso, and latent variable models, a dimensionally sparse projection is found in which the originally nonstationary field exhibits stationarity. Following a comparison with existing methods in a simulated environment, dimension expansion is studied on a classic test-bed dataset historically used to study nonstationary models. Following this, we explore the use of dimension expansion in modeling air pollution in the United Kingdom, a process known to be strongly influenced by rural/urban effects, amongst others, which gives rise to a nonstationary field.