AMOEBA (version 1.1)

AMOEBA: A Multidirectional Optimum Ecotope-Based Algorithm

Description

A function to calculate spatial clusters using the Getis-Ord local statistic (Ord and Getis, 1995). It searches irregular clusters (ecotopes) on a map through boundaries or grid of it.

Usage

AMOEBA(outc, neig, power = 1, cpu = 1)

Arguments

outc

Numeric vector with the study outcome

neig

Neighbours object from "spdep" package

power

Number of times to apply AMOEBA

cpu

Total number of cpu to run faster the algorithm

Value

Return a vector with the classification of the outcome. The classification is ordered from lowest (low risk) to high (high risk). Also, it has as much $3 ^ power$ levels of risk.

Acknowledgments

This work is included within the project "A longitudinal multilevel analysis of socioeconomic disparities in cardiovascular diseases: questioning past evidence with new methodological approaches" supported by a grant from The Swedish Research Council (#D054740, PI: Juan Merlo).

Details

outc must be the same lenght as the total number of units in neig.

power must be a positive integer. This value tell us how many times the AMOEBA is applied. When you apply the algorithm one time, the algorithm is called AMOEBA (Aldstadt and Getis, 2006; Duque et al., 2011), but when you apply more than one time, the algorithm is called AMOEBA+.

cpu must be a positive integer. Also, be careful and no put a cpu bigger than you have in your computer because it still will work but it probably will take long time. This value helps the function with his computational time, because it is used in the parallelization of the algorithm (Widener et al, 2012).

References

Ord, J. K. and Getis, A. (1995). Local spatial autocorrelation statistics: Distributional issues and application.

Aldstadt, J. and Getis, A. (2006). Using AMOEBA to create a spatial weights matrix and identify spatial clusters.

Duque, J. C., Aldstadt, J., Velasquez, E., Franco, J. L., and Betancourt, A. (2011). A computationally efficient method for delineating irregularly shaped spatial clusters.

Widener, M. J., Crago, N. C., and Aldstadt, J. (2012). Developing a parallel computational implementation of AMOEBA.

Examples

####################
##### Shiny App
# Visit: http://amoeba-spatial-cluster.shinyapps.io/AMOEBA/

####################
##### R App
require(AMOEBA)

### Columbus OH spatial analysis data set (spdep package)
data(columbus)
require(maptools)
map<-readShapePoly(system.file('etc/shapes/columbus.shp',package='spdep')[1])

### Application of AMOEBA
res<-AMOEBA(columbus$CRIME,col.gal.nb,1,1)

### Plot
color<-rev(rainbow(length(table(res)),start=0,end=2/6))
plot(map,col=color[as.factor(res)])
title('Clustering of crimes at Columbus (OH)')
names<-c('\nLow','\nMedium','\nHigh')
require(plotrix)
color.legend(map@bbox[1,1],map@bbox[2,1]-0.2,map@bbox[1,2],map@bbox[2,1]-0.4,names,color,align='rb')