當樣本數少時，若使用近似常態法所得之信賴區間往往估計結果不盡理想。因此考慮使用貝氏方法調整，希望可以獲得一個更精確之信賴區間。調整後的貝氏可靠區間不僅其覆蓋機率至少為(1-α)，期望長度也比較短，而且達到均質的改善。 When the sample size is small, the evaluating confidence interval is not good by using asymptotic normal method. We consider that improving confidence intervals by using the Bayesian method to get a best one. The improved Bayes credible interval has that (a) the coverage probability is at least (1-α) (b) the length of the interval is small (c) uniformly improve the confidence interval.
