
Cluster a nd Cox Models 459
Figure 12.9. Log-Gaussian Cox Pro c ess. Le ft: random drivin g intensity function Λ(u) such
that logΛ(u) is a Gaussian random field with mean
µ
= 4 and exponentia l covariance C
0
(r) =
0.2exp(−10r). Right: realised point process X.
The var and scale default to 1, but this default would rarely be sensible.
If the RandomFields package is not available, the user could with a bit of effort simulate a log-
Gaussian Cox process f rom scratch using the constructive procedure described in Section 12.2.5.
However, this possibility only allows for generating log-Gaussian Cox processes with an exponen-
tial covariance structur ...