Hi all,
I'm using the spatstat package to calculate spatial clustering between
pairs of tree species within 20 meter radius neighborhoods, using the
Lcross.inhom function. The spatstat website says that different border
correction methods should give roughly the same results, unless the sample
size is too small or the window is too irregular (
http://www.spatstat.org/spatstat/explainKplot.html). I have a rectangular
window 400 x 500 m wide, so I think window irregularity shouldn't be a
problem. From plotting the L functions for each species pair, it looks like
the different correction methods agree when each species has roughly more
than 50 individuals, so I was excluding L20 values for species with fewer
individuals, but this seems pretty hand-wavy. Does anyone have any
recommendations for determining when the different correction methods are
'too different' or the sample size is too small?
Thanks for considering my question! Any help is much appreciated.
Best,
Megan Bartlett
[[alternative HTML version deleted]]
_______________________________________________
R-sig-Geo mailing list
[email protected]
https://stat.ethz.ch/mailman/listinfo/r-sig-geo
