[R] Cross K Ripley's function and "spatio-temporal interaction power"

[Rolf Turner]

On 13/07/11 00:57, ruocco wrote:
> Dear All,
> I have a collections of spatial data. I have to analyze pairs of these 
> point patterns to test their spatial interaction. I was moving towards 
> the cross K Ripley's function. The problem, however, are the following:
> 1) What is the best way to get a single real value that represents the 
> interaction "power"?
> 2) How to obtain a value that even allows me to rank the pairwise 
> point patterns according to their interaction "power"?
>
> PS: I have to perform the same analysis for temporal interaction and 
> spatio-temporal interaction.
>
> Thanks in advance for your help
>
> Best Regards
>
> Massimiliano Ruocco

I don't believe that there is any (possible) answer to your question.  I 
would be
happy to be corrected in this regard.

What do you mean by ``interaction "power" ''?

As far as I know, if the temporal direction is to be treated differently 
from spatial
dimensions, then there is no appropriate analogue of the K function 
available.  So
you  cannot attack ``temporal interaction and spatio-temporal 
interaction'' in this
way.

I believe that you need to formulate your goals much more clearly and 
precisely.
If your goals can be so formulated then it may be possible to develop 
some techniques
to achieve them.  However my guess is that such development would 
constitute a
major research undertaking.

     cheers,

             Rolf Turner