Asymptotic Accuracy and Reliability of Two-sided Bayesian Credible Region Under Discrimination Two Compound Hypotheses

The paper examines the properties of Bayesian two-sided confidence intervals within the framework of a probabilistic model, when the output parameter \(\theta\) has a normal distribution, and observations also followes with a normal law with a mean equal to \(\theta\) . It is shown that the functions of d-posterior reliability and accuracy approach unity with increasing sample size. Asymptotic equalities are established for these functions in various statistical analysis schemes.