
Spacing 295
which is more convenient for studying summary functions. For example, the K-functio n and G-
function of a stationar y point proce ss are
K(r) = E
!u
n(X ∩()b(u,r))
G(r) = P
!u
{d(u, X) ≤ r}.
The statement at the end of the last section implies that for a Poisson point process, the reduced
Palm distribution is identical to the usual distribution of the process.
8.13 FAQs
• A plot of the G-function for my d ata suggests the p attern is regular, but the plot of th e K-func tion
suggests clustering. Which is right?
The results from G and K are not contradictory, because a summa ry function does not completely
characterise the point process. The