getHotspots: find spatial clusters using supervised learning methods

In TreeHotspots: Hotspot Detection using Classification Trees

Description

Here we consider one fixed rotation of the data and

seek hot spots using a classification tree

Usage

getHotspots(formula = violent ~ X + Y, rotX, NullClass = "0", 


    minsize = 200, minArea = 20, maxArea = 250, ORfilter = list(OR = TRUE, 


        OR1 = 1.8, OR2 = 0.1), TreeAlgorithm = c("rpart", "ctree")[1], 


    verbose = 1, ...)

Arguments

formula	A formula expression. The left-hand-side (response) should be either a numerical vector when a regression tree will be fitted or a factor, when a classification tree is produced. The right-hand-side should be a series of numeric or factor variables separated by +; there should be no interaction terms. Both . and - are allowed: regression trees can have offset terms.
rotX	data, possibly already rotated
NullClass	if y is a factor, this is the category used for the background
minsize	minimum number of points inside a cluster
minArea	minimum area of a cluster, units: lat*lon
maxArea	maximum area of a cluster, units: lat*lon
ORfilter	filter on minimum and maximum odds ratios (OR)
TreeAlgorithm	which tree algorithm to choose
verbose	level of verbosity
...	further arguments to tree

Value

identified clusters (if any)

Author(s)

Markus Loecher

Examples

#examples to come


data("drugCrimes", envir = environment())


drugCrimes$MATCH = factor(drugCrimes$MATCH)


spot1 = getHotspots(MATCH ~ X+Y,drugCrimes, minArea=20/10^6,maxArea=250/10^6, 


                    ORfilter=list(OR=FALSE,OR1=0.8,OR2=0.1))


suppressWarnings(suppressMessages(library("PBSmapping")))


PBSmapping::plotPolys(spot1[1:5,],density=NULL,xlim=range(drugCrimes$X),ylim=range(drugCrimes$Y),


          border="blue",lwd=2)


ranRows=sample(1:nrow(drugCrimes), 5000)


points(Y~X,data=drugCrimes[ranRows,],col=RgoogleMaps::AddAlpha(4-as.numeric(MATCH)),pch=20,cex=0.6)

TreeHotspots documentation built on May 2, 2019, 5:17 p.m.