基于有序抽样样本参数的置信区间

Confidence interval of parameters based on ranked set sampling

讨论了一种新的抽样方法即有序抽样方法下样本的信息矩阵及参数的极大似然估计的性质 ,把它和简单随机抽样样本下的信息矩阵进行了比较 ,并以此估计为基础给出了一种新的参数的置信区间的构造方法 ,表明在相同的置信水平下以此方法构造的参数的置信区间比传统方法下构造的参数的置信区间更短 ,因而也具有更高的相对精度 .

This paper presented a sampling method—ranked set sampling and the sample information matrix as well as the characters of parameters′ maximum likelihood estimation under this sample. It also compared it with those under simple random sample. The confidence interval constructed on this base was characterized by shorter length and higher relative efficiency compared to that constructed by the traditional method under the same confidence level.