ABSTRACT
Classical methods of inference for spatial point processes are hampered by the intractability of the likelihood function for most models of interest. To some extent, this difficulty has been alleviated by the development of Monte Carlo methods for calculating approximate likelihoods, and we shall discuss the resulting methods of inference in the next chapter. Nevertheless, the more ad hoc methods of inference described in this chapter, which operate by comparing theoretical and empirical summary descriptions, remain useful for at least two reasons. Firstly, they are computationally straightforward, and consequently useful for rapid exploration of a range of possible models. Secondly, they provide direct, graphical methods for assessing model fit.
