Theoretical properties of tests for spatial clustering of count data |
| |
Authors: | Inkyung Jung Martin Kulldorff |
| |
Institution: | 1. Department of Epidemiology and Biostatistics University of Texas Health Science Center San Antonio, TX 78229, USA;2. Department of Ambulatory Care and Prevention Harvard Medical School and Harvard Pilgrim Health Care Boston, MA 02215, USA |
| |
Abstract: | Testing for spatial clustering of count data is an important problem in spatial data analysis. Several procedures have been proposed to this end but despite their extensive use, studies of their fundamental theoretical properties are almost non‐existent. The authors suggest two conditions that any reasonable test for spatial clustering should satisfy. The latter are based on the notion that the null hypothesis should be rejected almost surely as the amount of spatial clustering tends to infinity. The authors show that the chisquared test and the Potthoff—Whittinghill V have both properties but that other classical tests do not. |
| |
Keywords: | Cluster analysis clustering global clustering tests spatial statistics |
|