606 Spatial Point Patterns: Methodology and Applications with R
or ‘dividing hyp othesis’ [17 9, pp. 51–52] that separates spatial clustering from sp a tial regularity, as
sketched in Figure 7.1.
For multitype point patterns, th e analogous reference model is Complete Spatial Randomne ss
and Independ ence (CSRI). In some studies, the key scientific question is whether the data could
have been generated by CSRI. However, CSRI does not serve quite the same role as a ‘dividing
hypothesis’. For multitype patterns there are at least two natural choice s for a dividing hypothesis:
random labelling (MPP2): the type labe ls are randomly allo cated to the points;
independent components (MPP4): the poin ts of different types are indep endent.
If both of these ...