EXAMPLES OF MARKOV CHAIN SAMPLING FOR SIMPLE DISTRIBUTIONS

Almost all uses of the software in this package involve sampling from
a distribution using Markov chain methods, and then making Monte Carlo
estimates for the expectations of functions of state based on this
sample.  This is done, for example, when make predictions for test
cases based on the posterior distribution of a neural network model.

Some ways of doing Markov chain sampling are illustrated in the
examples of modeling with neural networks, Gaussian processes, etc.
If your main interest is in those models, you could start with those
examples, but the simpler examples in this section may be more helpful
in understanding the Markov chain methods.  These examples also
introduce the facilities of the 'dist' module, which are used when
sampling from Bayesian models defined using formulas for the prior and
likelihood, as illustrated by the examples in Ex-bayes.doc.  The
examples there also illustrate some additional aspects of Markov chain
sampling.

The commands used in these examples can also be found in command files
in the "ex-dist" directory.