spseg: Integrated Functions for Spatial Segregation Analysis Spatial...
In spatialkernel: Non-Parametric Estimation of Spatial Segregation in a Multivariate Point Process

Description Usage Arguments Details Value Note Author(s) See Also

Integrated Functions for Spatial Segregation Analysis Spatial segregation analysis to be performed by a single function and presentations by associated plot functions.

Run the spatial segregation analysis

spseg for spatstat objects

## S3 method for class 'matrix'
spseg(pts, marks, h, opt = 2, ntest = 100, poly = NULL,
  delta = min(apply(apply(pts, 2, range), 2, diff))/100, proc = TRUE, ...)

plotcv(obj, ...)

plotphat(obj, types = unique(obj$marks), sup = TRUE,
  col = risk.colors(10), breaks = seq(0, 1, length = length(col) + 1), ...)

plotmc(obj, types = unique(obj$marks), quan = c(0.05, 0.95), sup = FALSE,
  col = risk.colors(10), breaks = seq(0, 1, length = length(col) + 1), ...)

spseg(pts, ...)

## S3 method for class 'ppp'
spseg(pts, h, opt, ...)

`pts`	an object that contains the points. This could be a two-column matrix or a ppp object from spatstat.
`marks`	numeric/character vector of the types of the point in the data.
`h`	numeric vector of the kernel smoothing bandwidth at which to calculate the cross-validated log-likelihood function.
`opt`	integer, 1 to select bandwidth; 2 to calculate type-specific probabilities; and 3 to do the Monte Carlo segregation test.
`ntest`	integer with default 100, number of simulations for the Monte Carlo test.
`poly`	matrix containing the `x,y`-coordinates of the polygonal boundary of the data.
`delta`	spacing distance of grid points at which to calculate the estimated type-specific probabilities for `image` plot.
`proc`	logical with default `TRUE` to print the processing message.
`...`	other arguments concerning `plot` and `points`
`obj`	list of the returning value of `spseg`.
`types`	numeric/character types of the marks of data points to plot the estimated type-specific probabilities, default to plot all types.
`sup`	logical with default `FALSE`, if `TRUE` to superimpose data points on the estimated type-specific probability surface.
`col`	list of colors such as that generated by `risk.colors`.
`breaks`	a set of breakpoints for the `col`: must give one more breakpoint than colour.
`quan`	numeric, the pointwise significance levels to add contours to `image` plot of the estimated type-specific probability surface, with default of `c(0.05, 0.95)`.

spseg implements a complete spatial segregation analysis by selecting bandwidth, calculating the type-specific probabilities, and then carrying out the Monte Carlo test of spatial segregation and pointwise significance. Some plot functions are also provided here so that users can easily present the results.

These functions are provided only for the convenience of users. Users can instead use individual functions to implement the analysis step by step and plot the diagrams as they wish.

Examples of how to use spseg and present results using plot functions are presented in spatialkernel-package.

This is the details of the S3 generic method

Does spseg for marked ppp objects

A list with components

hcv: bandwidth selected by the cross-validated log-likelihood function.
gridx,gridy: x, y coordinate vectors at which the grid points are generated at which to calculate the type-specific probabilities and pointwise segregation test p-value.
p: estimated type-specific probabilities at grid points generated by vectors gridx, gridy.
pvalue: p-value of the Monte Carlo spatial segregation test.
stpvalue: pointwise p-value of the Monte Carlo spatial segregation test.
...: copy of pts, marks, h, opt.

spseg results

an spseg object

Setting h to a unique value may force spseg to skip the selecting bandwidth step, go straight to calculate the type-specific probabilities and then test the spatial segregation with this fixed value of bandwidth.

Barry Rowlingson

cvloglk, phat, mcseg.test, pinpoly, risk.colors, and metre