摘要
本文研究强混合样本下随机设计情形线性模型的经验似然推断,将分块技术应用到经验似然方法中,证明了线性模型的参数β的对数经验似然比统计量的渐近分布为卡方分布,由此构造了强混合样本下β的经验似然置信区间.在有限样本情况下给出数值模拟结果.
In this paper,we study the blockwise empirical likelihood(EL)method to construct the confidence regions for the regression vectorβin a linear model with random designs under strong mixing samples.It is shown that the blockwise EL ratio statistic forβis asymptoticallyχ^(2) distributed,which are used to construct the EL-based confidence region forβwith random designs under strong mixing samples.Results of a simulation study on the finite sample performance of the confidence region are reported.
作者
黄玉
秦永松
HUANG Yu;QIN Yongsong(Nanning University,Nanning 530200,China;Guangxi Normal University,Guilin 541004,China)
出处
《应用数学》
CSCD
北大核心
2022年第2期261-271,共11页
Mathematica Applicata
基金
Supported by the National Natural Science Foundation of China(11901124)
the Basic Ability Promotion Project for Young and Middle-age Teachers of Guangxi(2020KY64007)
the Professor Nurturing Engineering Project of Nanning University(2021JSGC07)。
关键词
线性模型
随机设计
分块经验似然
强混合
置信区间
Linear model
Random design
Blockwise empirical likelihood
Strong mixing
Confidence region
