Abstract
Purchase PDF (1265 K)
E-mail Article
Add to my Quick Links
Cited By in Scopus (1)
Purchase PDF (741 K)
doi:10.1016/j.csda.2006.11.039 
Copyright © 2007 Elsevier B.V. All rights reserved.
Line and point cluster models for spatial health data
References and further reading may be available for this article. To view references and further reading you must purchase this article.
Andrew B. Lawsona, 
, 
, Silvia Simeonb, Martin Kulldorffc, Annibale Biggerib and Corrado Magnanid
Received 22 November 2005;
Abstract
Spatial cluster modelling of small area disease incidence and mortality has previously focused on clusters where excess risk is distributed around fixed points, and the aim is the reconstruction of these points (cluster centers). Often there is a need to assess clusters of a different form, such as around roads or river systems. These clusters are often linear or can be approximated by combinations of several linear segments. In this paper the recovery of point and line clusters is considered jointly. An example application is given where both linear or point clustering could be present.
Keywords: Spatial; Clustering; Parametric; Modelling; Linear; Point
Article Outline
- 1. Introduction
- 2. Definitions of model components
- 2.1. Prior distribution for line segments
- 2.2. Prior distribution of cluster centers
- 2.3. Hyperprior distributions
- 3. Birth–death MCMC algorithms
- 4. Posterior ratios for the BD algorithm
- 5. Center sampler design
- 6. Sampler design for other parameters
- 7. Convergence
- 8. Example: malignant mesothelioma in Casale Monferrato
- 9. Simulation behavior
- 10. Discussion and Conclusions
- References
