[R] Need help creating spatial correlation for MC simulation

Roger Bivand Roger.Bivand at nhh.no
Tue Sep 23 12:57:48 CEST 2008

jjh21 <jjharden <at> gmail.com> writes:

> Thank you for the input.
> Which command in the spatstat package am I looking for? The documentation is
> unclear to me.
> milton ruser wrote:
> > 

> > I think that on spatial stat you will find several ways of simulate
> > spatial
> > pattern that (point or line) that may be what you are looking for. Case
> > not,

> > 
> > On Wed, Sep 17, 2008 at 7:36 PM, jjh21 <jjharden <at> gmail.com> wrote:
> > 
> >>
> >> I want to create a dataset in R with spatial correlation (i.e.
> >> clustering) 
> >> built in for a linear regression analysis. Any tips on how to do this?

The original question was pretty opaque too. I think that the response from 
both the people who replied was based on your use of the word "clustering", 
which is most often used in spatial statistics with reference to the 
analysis of point patterns.

Since that possibly isn't the case, perhaps you should reformulate your 
question, and consider posting on the R-sig-geo list, after reviewing 
the Spatial Task View on your nearest CRAN mirror. My guess would be that
you are thinking of local indicators of spatial association using areal
data, in which case there are some functions in the spdep package, but
until we know what your actual needs are, it is hard to say. You could,
for example, be looking for some of the tests in the DCluster package,
but we cannot tell.

Roger Bivand

More information about the R-help mailing list