Rogerson_2002 - Change Detection Thresholds: Alternative...

Info iconThis preview shows pages 1–5. Sign up to view the full content.

View Full Document Right Arrow Icon
Change Detection Thresholds: Alternative Statistical Approaches to Detecting Temporal Change in Spatial Patterns Peter A. Rogerson Daikwon Han Ikuho Yamada Department of Geography National Center for Geographic Information and Analysis University at Buffalo Buffalo, NY 14261 USA January, 2002
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Acknowledgements The authors are grateful to the National Center for Health Statistics (NCHS) for providing the Compressed Mortality File. The analyses, results, and interpretation are the responsibility of the authors and not the NCHS. The support of National Institutes of Health Grant 1R01 ES09816-01, National Science Foundation Award BCS-9905900, and National Cancer Institute Grant R01 CA92693-01 is gratefully acknowledged.
Background image of page 2
1. Introduction How do we detect change over time in spatial patterns? Specifically, suppose we have data for a set of regions ( i= 1 ,…,m ), in the form of observed and expected counts (e.g., for a disease), or more generally, in the form of z -scores (which may represent, e.g., a differenced image in an application to remote sensing). Build upon (a) spatial statistics designed to find geographic patterns at one point in time, and upon (b) aspatial monitoring methods. Outline - classification of alternative approaches - illustration using annual, county-level breast cancer mortality data (1968–1998)
Background image of page 3

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
1. Aspatial Methods 2. Spatial Methods I. Static tests ______________________________ A. Global/ Chi-square -Spatial chi-square General goodness-of-fit -Tango’s test -Moran’s I B. Detection - M test - spatial M test of clustering Fuchs and Kennett Rogerson (2001) (max. of local (1980) - spatial scan tests) - retrospective - max of local change-point Moran’s I methods _______________________________ II. Prospective Monitoring C. Individual Shewhart test spatial z -scores charts chi-square each year, using ARL D. Univariate - special case surveillance of Cusum of 2.D. spatial statistics (2.A or 2.B) - monitoring (e.g., Rogerson change-point 1997; Kulldorff statistics 2001) E. Multivariate special case spatial, multi- cusum of 2.E. variate monitoring
Background image of page 4
Image of page 5
This is the end of the preview. Sign up to access the rest of the document.

Page1 / 25

Rogerson_2002 - Change Detection Thresholds: Alternative...

This preview shows document pages 1 - 5. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online