ABSTRACT
The likelihood function plays a fundamental role in both classical and Bayesian approaches to statistical inference. When more ad hoc methods, such as those described in Chapter 7, are used instead this is often for pragmatic reasons. The most obvious of these is that likelihood-based methods for most spatial point process models are notoriously intractable. However, this difficulty has to some extent been alleviated by recent developments in Monte Carlo methods of inference, including but not restricted to Markov chain Monte Carlo methods (Gilks, Richardson and Spiegelhalter, 1996).
