[R-sig-eco] xPOD statistic

Dixon, Philip M [STAT] pdixon at iastate.edu
Sun Jan 26 13:35:26 CET 2014


Alexandre,

I don't know any prewritten function to calculate the collection of xPOD statisitics for a multiptype point pattern.  However, it won't be difficult to code that yourself using functions in the spatstat library.

Brown et al. define A_{ij} for each pair of species as Int( log g_{ij}(r) dr, where g_{ij}(r) is the pair correlation function for species i and j at distance r.  That is numerically approximated by a sum of log \hat{g}_{ij}(r) computed at 10 (their suggestion) values of r.

spatstat provides the pcf() function to compute g(r) and the alltypes() function to compute g(r) for all combinations of pairs of species.  After that, it is just a matter of manipulating the function value array (see fasp()).

Best wishes,
Philip Dixon



More information about the R-sig-ecology mailing list