I would like to add:
* If you perform a pointwise test, the value of "r" at which you conduct
the test much be chosen a priori --- before collecting, or at least
before "looking at" the data. Otherwise the associated p-value is
meaningless. It is *very* unlikely that you had an a priori value of r
in mind!
* A test based on the global envelope will not have very much power.
* My guess is that the dclf.test() route is your best bet.
cheers,
Rolf Turner
On 14/01/14 05:09, Marcelino de la Cruz wrote:
Yes, it is possible and very easy. How do you extract your p-value
depends on wether you are making a pointwise or a global test. See the
help page of envelope(). You can also try a maximum absolute deviation
test with dclf.test().
Cheers,
Marcelino
El 13/01/2014 16:52, Francesco Carrer escribió:
Hi,
I have a distribution of artifact within an archaeological surface
(dataset: ID, DIMENSION, X, Y), and I need to verify which is the
degree of
aggregation of these artifacts at different scales. I applied the L
Function (Lest in spatstat), and plotted the resulting observed values of
L(r) against the highest and lowest simulated values of L(r). Is it
possible to extract a p-value that assess that the aggregation of my data
is significantly higher (or lower) than the simulate values?
_______________________________________________
R-sig-Geo mailing list
R-sig-Geo@r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-geo
