Testing differences between case and control point patterns using nearest neighbour distances and bootstrapping

KA Henry, LM Burge, D Nguyen

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

1 Scopus citations

Abstract

This paper proposes a method for comparing point patterns in case and control data by using nearest neighbour distances, bootstrapping and the Wilcoxon Signed-Rank Test. Case-control comparisons are frequently used in medical geography and epidemiology to examine the patterns of disease and infer transmission of infectious pathogens. Our strategy addresses the problem of handling spatial analysis when the numbers of cases differ from controls. Differences in sample sizes can affect density of points and therefore bias nearest neighbour distances. To demonstrate this method we created a control set of 250 points and two sets of cases 125 points each. Bootstrapping the control data and comparing each run statistically to the cases can provide confidence intervals and estimate the risk of erroneously rejecting the null hypothesis. We follow with a case study of tuberculosis. The spatial distributions of different bacterial strains were compared and the nearest neighbour distances were analyzed as a surrogate for possible transmission of tuberculosis. The method may be useful to epidemiologists, geologists, biologists, geographers and ecologists for evaluating differences between the spatial structures of points.

Original languageAmerican English
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
EditorsVipin Kumar, Marina L. Gavrilova, Chih Jeng Kenneth Tan, Pierre L’Ecuyer, Chih Jeng Kenneth Tan
PublisherSpringer Verlag
Pages33-42
Number of pages10
ISBN (Print)3540401563
DOIs
StatePublished - 2003

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume2669
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Fingerprint

Dive into the research topics of 'Testing differences between case and control point patterns using nearest neighbour distances and bootstrapping'. Together they form a unique fingerprint.

Cite this