Abstract Starting with the Sterne's confidence regions, there have been a string of related attempts to improve on the two-sided Clopper–Pearson bounds for binomial inference. That work is brought to fruition using Walley's imprecise posterior likelihood as the basis for a new test statistic. The results are expressed as a consonant confidence structure, from which a new set of exact confidence intervals for binomial inference can be efficiently computed.

中文翻译：

---

使用沃利的不精确后验概率作为检验统计量得出的二项式推理的新两侧置信区间

摘要 从 Sterne 的置信区域开始，已经有一系列相关的尝试来改进二项式推理的两侧 Clopper-Pearson 界限。使用沃利的不精确后验可能性作为新检验统计量的基础，这项工作取得了成果。结果表示为辅音置信结构，从中可以有效地计算出一组新的二项式推理的精确置信区间。