[R] Comparing bivariate kernel density estimates

[Timothy S. Hare]

Dear R-Users,

Using R, is there method for comparing two sets of kernel density surfaces? 
I need to implement either a goodness-of-fit test such as 
sm.density.compare [sm], but for bivariate kernel density surfaces, or find 
a means for calculating matrix comparison statistics for asymmetrical matrices.

Background:
I am investigating the territorial behavior of baboons in a closed habitat. 
My data are composed of point spatial coordinates marked with the length of 
occupation in seconds for each baboon observed (N=17). I want to estimate 
the spatial areas of habitual use and compare between individuals to assess 
wether the estimated territories are derived from the same distribution or 
not. I initially thought to use cross-K point pattern analysis, but I could 
not find a way to use the continuous variable attached to each data point. 
I also thought to use the method for comparing kernel density estimates 
describe by Fotheringham, Brunsdon and Charlton in "Quantitative Geography: 
Perspectives on Spatial Analysis", but I cannot find a way to implement the 
calculation of the variance for the comparison of the pairs of matrices.

I appreciate any help in this matter.

Thank You,
Timothy


Timothy S. Hare
Mellon Postdoctoral Fellow and Lecturer
Anthropology Department
Grinnell College                                (641) 269-3966
204 Goodnow Hall                                fax (641) 269-4330
Grinnell, IA 50112-1690                 haret at grinnell.edu

"If a geographer is not fascinated by maps to the extent of always needing 
to be surrounded by them, then that is a clue that he or she [or zie] has 
chosen the wrong profession." -- Carl Sauer
-------------- next part --------------

---