ABSTRACT
We now consider situations in which a different attitude is taken towards 0 ; namely, it is regarded as the realized value of a random variable 0 , with p.d.f. /© (0), called the prior distribution. Moreover this p.d.f. is, for the most part, regarded as known independently of the data under analysis, although in Section 10.7 we deal with socalled empirical Bayes procedures, where /© (0) is regarded as unknown and to be estimated from relevant data.
