ABSTRACT
From Chapter 5 it should be clear that the only thing that truly distinguishes frequentist and subjective approaches to measuring uncertainty is how we assign the initial probabilities to elementary events. We argued that, since there must always be some subjective assumptions in the initial assignment of probabilities, it was inevitable that all probability was subjective. What this means is that the probability assigned to an uncertain event A is always conditional on a context K, which you can think of as some set of knowledge and assumptions. It is this central role of conditional probability that lies at the heart of the Bayesian approach described in this chapter. Fortunately, there is actually very little extra to introduce by way of theory.
