[R-sig-Geo] Kest (spatstat) and "r" values

[Marcelino de la Cruz]

Maybe the information related to r you are looking for is in the Warning 
section:

"Warnings

The estimator of K(r) is approximately unbiased for each fixed r. Bias 
increases with r and depends on the window geometry. For a rectangular 
window it is prudent to restrict the r values to a maximum of 1/4 of the 
smaller side length of the rectangle".

Maybe I'm wrong but I think that  Kest computest K(r) by default for r 
up to 1/4 of the shorter side of the *enclosing rectangle* of your window.

HTH,

Marcelino



El 22/10/2014 a las 20:01, Silvia Cordero-Sancho escribió:
> Hello,
>
> I am employing the Ripley's K function to evaluate my data and have some
> concerns regarding the maximum distance reported.
>
> My data was collected in 7 different sites. At each site we recorded the
> location x-y of the events. The 7 sites were had *"similar conditions"
> *but their
> extent vary, and in addition, not all sites presented the same number of
> events (number of events per site vary from 92 to up to 400 ).
>
> Each site was evaluated independently, meaning that Site-1 had their on set
> of events and own window of observation (W). I used the *ripras* function
> to define *W, *(for all sites, W is an irregular shaped window).
>
> I have run two first order exploratory  tools
> (quadrat.test, clarkevans.test). The overall results is that my
> observations are not random and show aggregation tendencies for all sites.
>
>
> I applied the function* Kest* to all my seven sites to explore the second
> order properties, I did not provide a *r* value (as recommended in the help
> section) and here is the "curious" outcome. Indifferently of the area of W
> or the number  of events per site, the maximum r-values reported for all
> sites is very similar (between 5000 m to  6140 m). I am trying to
> understand the reason behind these results. Is this an overall
> characteristic of my pattern OR is a result of the algorithm that spatstat
> employs to define r (e.g. it has a maximum threshold)
>
> Here my code:
>
> K2<-Kest(site2,correction="best"); where site2 is mydata
> max(K2$r) = 5,260 m
>
> Here the maximum distance values (in meters) for the other sites
>
> max(K1$r) = 6,140
> max(K3$r) = 5,148
> max(K4$r) = 5,016
> max(K5$r) = 5,194
> max(K6$r) = 5,055
> max(K3$r) = 5,317
>
> PS: I used the extent of the observation window for site 2 to create a
> random set of x-y locations (n=182, same than the observed number of events
> in this site). I run the test and the maximum distance was the same (5,260
> m)
>
> I cannot find information how r is defined in the function Kest
>
> If any one any comment I will appreciate
>
> Thank you!
>
> Silvia Cordero-Sancho
>
> 	[[alternative HTML version deleted]]
>
> _______________________________________________
> R-sig-Geo mailing list
> R-sig-Geo at r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>
>