针对二项分布比例参数的大样本置信区间,Agresti与Coull提出一种改进的Agresti -Coul置信区间。文章模拟研究Agresti-Coul置信区间的表现,证明它包含二项分布比例参数的大样本置信区间,同时也发现它只能在成功次数不低于5时可用。对于二项分布比例参数点估计为0%的情形,文章给出了一种替代估计形式。%Agresti & Coull proposed an improvement for Binomial proportion confidence interval under large sample ,called Agresti-Coul confidence interval .In this paper we conduct a detailed simulation study on Agresti -Coul confidence interval performances ,and prove that it contains previous Binomial proportion confidence interval with discovering that it is available only when the number of success in Binomial trial is not less than 5 .When the point estimator of the proportion parameter is 0% ,we present a replacement estimator for p confidence interval .
展开▼