Scan Statistics and Applications pp 303-322
Spatial Scan Statistics: Models, Calculations, and Applications
A common problem in spatial statistics is whether a set of points are randomly distributed or if they show signs of clusters or clustering. When the locations of clusters are of interest, it is natural to use a spatial scan statistic.
Different spatial scan statistics have been proposed. These are discussed and presented in a general framework that incorporates two-dimensional scan statistics on the plane or on a sphere, as well as three-dimensional scan statistics in space or in space—time. Computational issues are then looked at, presenting efficient algorithms that can be used for different scan statistics in connection with Monte Carlo-based hypothesis testing. It is shown that the computational requirements are reasonable even for very large data sets. Which scan statistic to use will depend on the application at hand, which is discussed in terms of past as well as possible future practical applications in areas such as epidemiology, medical imaging, astronomy, archaeology, urban and regional planning, and reconnaissance.
Keywords and phrasesSpatial statistics geography spatial clusters space—time clusters maximum likelihood likelihood ratio test
Unable to display preview. Download preview PDF.
- 3.Alm, S. E. (1998). On the distribution of scan statistics for Poisson processes in two and three dimensionsExtremes(to appear).Google Scholar
- 4.Alt, K. W. and Vach, W. (1991). The reconstruction of `genetic kinship’ in prehistoric burial complexes-problems and statistics, InClassification Data Analysis and Knowledge Organization (Eds.H. H. Bock and P. Ihm), Berlin: Springer-Verlag.Google Scholar
- 5.Anderson, N. H. and Titterington, D. M. (1997). Some methods for investigating spatial clustering with epidemiological applicationsJournal of the Royal Statistical Society Series A 16087–105.Google Scholar
- 6.Besag, J. and Newell, J. (1991). The detection of clusters in rare diseasesJournal of the Royal Statistical Society Series A 154143–155.Google Scholar
- 14.Kulldorff, M., Athas, W. F., Feuer, E. J., Miller, B. A. and Key, C. R. (1998). Evaluating cluster alarms: A space-time scan statistic and brain cancer in Los AlamosAmerican Journal of Public Health(submitted).Google Scholar
- 17.Kulldorff, M. and Williams, G. (1997).SaTScan v 1.0 Software for the Space and Space-Time Scan StatisticsBethesda, MD: National Cancer Institute.Google Scholar
- 18.Lawson, A. (1997). Cluster modeling of disease incidence via MCMC methodsJournal of Statistical Planning and Inference(submitted).Google Scholar
- 20.Mánsson, M. (1996). On Clustering of Random Points in the Plain and in SpacePh.D. ThesisDepartment of Mathematics, Chalmers University of Technology and Gothenburg University, Gothenburg.Google Scholar
- 21.Mantel, N. (1967). The detection of disease clustering and a generalized regression approachCancer Research 27209–220.Google Scholar
- 22.Miller, B. A., Gloeckler Ries, L. Y., Hankey, B. F., Kosary, C. L., Harras, A., Devesa, S. S. and Edwards, B. K. (1993).SEER Cancer Statistics Review 1973–1990Bethesda, MD: National Cancer Institute.Google Scholar
- 27.Priebe, C. (1998). A spatial scan statistic for stochastic scan partitionsJournal of the American Statistical Association(to appear).Google Scholar
- 30.Turnbull, B., Iwano, E. J., Burnett, W. S., Howe, H. L. and Clark, L. C. (1990). Monitoring for clusters of disease: Application to leukemia incidence in Upstate New YorkAmerican Journal of Epidemiology 132S136—S143.Google Scholar
- 31.Wallenstein, S., Gould, M. S. and Kleinman, M. (1989a). Use of the scan statistic to detect time-space clusteringAmerican Journal of Epidemiology 1301057–1064.Google Scholar
- 36.Walsh, S. J. and Fenster, J. R. (1997). Geographical clustering of mortality from systemic sclerosis in the Southeastern United States, 1981–90Journal of Rheumatology(to appear).Google Scholar