build.clusters.spatial: Builds clusters with a given maximum size using a k-means...

In HiDimMaxStable: Inference on High Dimensional Max-Stable Distributions

Description

Builds clusters from the spatial locations of sites using a k-means clustering, to get a partition (of the sites) whose block sizes are at most 5 so that the partition-composite likelihood for observations of a max-stable process at the sites can be computed in a moderate time.

Usage

build.clusters.spatial(xy,max.size=5,plot=FALSE)

Arguments

xy	a matrix giving the coordinates of each location. Each row corresponds to one location.
max.size	an integer giving the maximum size of the blocks of the returned partition.
plot	Whether you also whant a plot or not.

Value

the return value is a vector of integers giving the index of the set for each location.

Examples

n.site<-10
xy<-matrix(runif(2 * n.site, 0, 0.5), ncol = 2)

param<-c(0.5,1.5)
n.obs<-20
library(SpatialExtremes)
data<-t(rmaxstab(n.obs, xy, "whitmat",
                 nugget = 0, range = param[1], smooth = param[2]))
cl<-build.clusters.spatial(xy)

d<-maxstable.l.clusters(data,clusters=cl,
                        params=param,
                        category="normal",
                        spatial=list(sites=xy,family=spatialWhittleMatern))

HiDimMaxStable documentation built on May 29, 2017, 6:20 p.m.