[R-sig-Geo] LSCV hlim

Maren Huck M.Huck at derby.ac.uk
Thu Sep 4 11:23:02 CEST 2014

Dear All,

I think I have a problem understanding the meaning of hlim when calculating the kernel bandwidth using LSCV. As I understood the help page, the LSCV would be searched within the limits of "ranging from hlim[1]*href to hlim[2]*href". So if the href is for example 23.7, I would expect that under the default setting where hlim is defined to be c(0.1,1.5) my LSCV-h would be estimated to be somewhere between 2.37 and 47.4.
With my dataset (sorry, not attached, since I am not sure how to), I get 3.7. So far so good, although this gives a rediculously small home-range.
If I now play around with the limits I can get very different values, e.g.,
> udLSCV<- kernelUD(E500sp, h="LSCV", hlim = c(0.001, 10))
> udLSCV at h$h        # gives you the value of h
[1] 4.811839

> udLSCV<- kernelUD(E500sp, h="LSCV", ,hlim = c(0.001, 100))
> udLSCV at h$h        # gives you the value of h
[1] 23.96653

> udLSCV<- kernelUD(E500sp, h="LSCV", ,hlim = c(0.001, 1000))
> udLSCV at h$h        # gives you the value of h
[1] 239.4542

If LSCV "forces" me to play around with these limits to get a home-range that "looks" sensible, what is the point of using a supposedly more "objective" estimator, rather than using some rule-of-thumb or user-defined method (e.g., Schuler, K. L., G. M. Schroeder, J. A. Jenks, and J. G. Kie. 2013. Ad-hoc smoothing parameter performance in kernel estimates of GPS-derived home ranges. Wildlife Biology. In press.)?
Or is there also an "objective" way on how to decide on reasonable limits for LSCV to get a reasonable home-range?

Thanks in advance for any suggestions,

Maren Huck
Dep. of Biological and Forensic Sciences
University of Derby
Kedleston Road
Derby DE22 1GB

The University of Derby has a published policy regarding email and reserves the right to monitor email traffic. If you believe this was sent to you in error, please select unsubscribe.

Unsubscribe and Security information contact:   infosec at derby.ac.uk
For all FOI requests please contact:   foi at derby.ac.uk
All other Contacts are at http://www.derby.ac.uk/its/contacts/

More information about the R-sig-Geo mailing list