3. BAYESIAN ESTIMATION
For Bayesian estimation, we need prior distribution
for scale parameter β. Jeffreys (1946) proposed a
formal rule for obtaining a non-informative prior
as
where β is M-vector valued parameter and I(β) is
the Fisher’s information matrix of order M × M. In
the present study, we consider the prior distribution
of β to be
where is constant. The posterior distribution of β
is
Therefore, the posterior density function of βα
given α is Gamma