用传统方法得到的伽玛分布参数的置信区间显然不是最短，因而在这个意义上讲也不是最优的。
The confidence interval of parameter of Gama distribution, obtained by using the traditional method, is obviously not the shortest. So it is not the best one in this sense.
当参数限制在某一范围内并服从一致的分布，且多余参数未知时，其贝叶斯置信区间有很高的置信概率。
The Bayesian credible intervals that arise when a parameter is given a uniform distribution over the restricted range and nuisance parameters are unknown have good frequentist coverage probabilities.
这说明在小样本下，研究参数的最短置信区间是必要的。
So it is necessary to research the shortest confidence interval of parameter for small samples.
应用推荐