[R-sig-Geo] (no subject)

[Adrian.Baddeley at csiro.au]

Mudrak, Erika [EEOBS] writes:

> Hello r-sig-geo,
> I have a question regarding envelopes for the pair correlation function in spatstat.

Questions about the spatstat package should be sent to the package authors. 

> The envelopes near r = 0 suddenly stretch to high values.   The envelopes also dip slightly at the maximum radius displayed. 
> Can someone tell me why this happens, and what I should do to avoid it?  Should I set xlim to be slightly above zero?

This is an artefact of the standard estimation algorithm, which gives a weight of 1/r to pairs of points at distance r.
It tends to happen when the point pattern is clustered. See help(pcf.ppp) for a description and literature references. 

Depending on what you want to do with the pair correlation, you can simply restrict the x limits using xlim (for plotting) or ginterval (for envelope-based tests like dclf.test)
or you could plot the log of the pair correlation e.g. plot(pcf(X), log(.) ~ r).

There exist other, non-standard estimation algorithms which will be added to spatstat sometime this year.

Adrian Baddeley
                                                    Prof Adrian Baddeley FAA
School of Earth & Environment                          |      CSIRO Mathematics, Informatics & Statistics
University of Western Australia                         |      Leeuwin Centre
35 Stirling Hwy, Crawley WA 6009, Australia    |      65 Brockway Rd, Floreat WA 6014, Australia
                                                        skype adrian.baddeley