spatialsegregation-shannon {spatialsegregation}R Documentation

Spatial Shannon Index

Description

Compute the spatial and aspatial Shannon index for a given multitype point pattern.

Usage

  shannonF(pp, parvec = 1:20, graph_type = "knn", ...)
  shannon_index(pp, graph_type = "knn", graph_par = 4, spatial = TRUE, ...)

Arguments

pp Multitype point pattern (see package 'spatstat')
parvec Default 1:20. Vector for the neighbourhood defining graph, e.g. geometric graph with different r.
graph_type Default knn. Type of the neighbourhood graph. Accepts: knn, geometric, delauney, gabriel.
graph_par The single parameter to calculate the value in.
spatial If FALSE, return the classical aspatial index value.
... Further tuning parametes for the function segregationFun.

Details

Date: 2009-03-09
License: GPL v2 or later

The form of Shannon index is $H = 1 - E(o)/E(N)$, where $E(N)$ is the global entropy and $E(o)$ is the local entropy calculated as $E(o)= - \sum \pi\_\tau log(\pi\_\tau)$, where the sum is over the types of the pattern, and $\pi\_\tau$ is the expected relative frequency of type $\tau$ points in a neighbourhood of a typical point of the pattern.

The function shannonF is the main calculation function. Uses function segregationFun with parameter ftype=2.

The function shannon_index is a shortcut to get a single value for the pattern. Uses 4-nn graph by default.

Author(s)

Tuomas Rajala University of Jyvaskyla, Finland tarajala@maths.jyu.fi

References

Rajala, Illian: Graph-based description of mingling and segregation in multitype spatial point patterns. To appear 2009.

Reardon, O'sullivan: Measures of spatial segregation. Sociological methodology, 34:121-162, 2004.


[Package spatialsegregation version 1.93 Index]