[R] SPATSTAT: Minimum points for a Ripley K to be sensible?

Adrian.Baddeley at csiro.au Adrian.Baddeley at csiro.au
Mon Jul 2 14:16:17 CEST 2012

Sebastian Pucilowski <s.pucilowski at student.unimelb.edu.au> writes:

 > What are the minimum number of points in a point pattern before a
 > clustering analysis using a Ripley K function loses any meaning?

It depends what are your definition of `meaningful'.

The K-function doesn't become meaningless (undefined) until there are fewer than 2 points. 

But if your dataset contained only 10 points, for example, you could type
    plot(envelope(runifpoint(10), Kest, nsim=1000, nrank=25))
to see the pointwise 95% prediction intervals for the K-function (as grey shading)
from a Poisson process with a mean of 10 points. 

To be detectablly different from a Poisson process, a dataset of 10 points
would need a K-function that goes outside these intervals somewhere. 
So it would need to be an extremely clustered pattern. 

Try for different values of "10" and adjust to suit your definition

Adrian Baddeley

More information about the R-help mailing list