PLACE:321 Snedecor Hall

SPEAKER:
Li Zhu,
University of Minnesota

TITLE:
Hierarchical Modeling of Spatio-temporally Misaligned Data

ABSTRACT:

Bayes and empirical Bayes methods have proven effective in smoothing crude maps of disease risk, eliminating the instability of estimates in low-population areas while maintaining overall geographic trends and patterns. Recent work applies the Bayesian hierarchical modeling approach to the analysis of areal data which are spatially misaligned, i.e., involving variables that are aggregated over differing sets of regional boundaries. This talk first extends the approach to the spatio-temporal case, so that misalignment can arise either within a given timepoint, or across timepoints (as when the regional boundaries themselves evolve over time).  We then extend beyond the case of strictly areal data, offering a fully model-based Bayesian approach to solving the modifiable area unit problem (MAUP).  Here one of the spatially-referenced variables is point process in nature, while another is available either only as areal summaries or as a point process at a different collection of reference points. Several real datasets relating ambient air pollution to respiratory system disease are used to illustrate the approach. Like many recent hierarchical Bayesian spatial applications, computing is implemented via carefully tailored Markov Chain Monte Carlo (MCMC) methods, with map summaries created using a geographic information system (GIS).
 
 

COFFEE: 10:45 a.m., 104 Snedecor Hall