Hierarchical Modeling and Analysis for Spatial Data
- PDF / 76,148 Bytes
- 2 Pages / 441 x 666 pts Page_size
- 64 Downloads / 217 Views
Book Review
Hierarchical Modeling and Analysis for Spatial Data By Sudipto Banerjee, Bradley P. Carlin, and Alan E. Gelfand Chapman & Hall/CRC, 2004, xvii+452 p., $74.95 (US), ISBN 1-58488-410-X I have recently reviewed a number of books for this journal on the general subjects of spatial statistics and spatial data analysis. In this series I have focused on texts that have attempted to address these topics from a holistic point of view rather than from the perspective of one specific tool or methodological area. For years spatial statistics developed along at least two parallel paths, one that can be broadly characterized as geostatistics and the other representing just about everything else (e.g., point processes, geographic information systems, mapping techniques, and various geographic measures of spatial association). Many have come to believe, however, that there is much more to spatial analysis than a single technique can provide, and that solid analytical work requires an integration of many different tools and methods. Hierarchical Modeling and Analysis for Spatial Data, by Sudipto Banerjee, Brad Carlin, and Alan Gelfand, is another recent text that comes at spatial problem solving from several different perspectives. The operative words in the title of this text are “hierarchical modeling.” Without question, this is a book whose fundamental approach to spatial modeling stems from the recently popularized hierarchical Bayesian methods, or what, in some circles, is referred to as the “Markov chain Monte Carlo (MCMC) revolution in Bayesian computing.” Readers without a bent towards Bayesian statistics or who are not interested in expanding their knowledge of these tools will likely be turned off fairly quickly. I was personally attracted to the book because I wanted to learn more about how to apply the Gibbs sampler, the Metropolis algorithm, and similar tools in problems that I encounter; but I was also drawn to the authors’ integration of these concepts and ideas with those from geostatistics, spatial point processes, and spatial association. The book starts off with a very nice overview of various kinds of spatial data problems. The authors address three basic kinds of data (point-referenced data, Published online: 14 April 2007. 261 C 2007 International Association for Mathematical Geology 0882-8121/07/0200-0261/1
262
Book Review
areal data, and point pattern data) that they develop throughout the text. One of the most interesting features of the book is the early discussion of cartography and its role in spatial modeling—a topic that is hardly ever mentioned in other contemporary books on spatial statistics. Although the information is not carried much further in the text, it is evidence that these authors have thought beyond the usual boundaries of statistical methods. There is also a fairly standard review of geostatistics, focused mainly on variogram estimation/modeling. Readers who are familiar with the recent text Statistical Methods for Spatial Data Analysis by Oliver Schabenberger and Caro
Data Loading...
