四川大学学报（自然科学版）2024，Vol.61Issue(2)：29-35,7.DOI:10.19907/j.0490-6756.2024.021004
平稳自回归模型的调整经验似然置信域
Adjusted empirical likelihood confidence regions for stationary autoregressive models
摘要
Abstract
Autoregressive models are widely used in diverse fields such as economics and information sci-ence.The statistical inference problem of autoregressive models is an important research topic in statistics.This paper aims at the adjusted empirical likelihood methods for autoregressive models.Based on the empiri-cal likelihood ratio statistic introduced by Chuang and Chan by combining the conditional least squares estima-tion of autoregressive models and the score function,we construct an adjusted empirical likelihood statistic for those cases that the empirical likelihood method may not have a solution.Then we prove that the limit distri-bution of the adjusted empirical likelihood ratio statistic is chi-square distribution,which means that one can construct the adjusted empirical likelihood confidence regions(intervals)for the model parameters.Finally,performance of the adjusted empirical likelihood method and the empirical likelihood method is compared by some numerical simulations.It is shown that the interval estimation obtained by the adjusted empirical likeli-hood method has better coverage.关键词
调整经验似然/平稳自回归模型/置信区间Key words
Adjusted empirical likelihood/Smooth autoregressive model/Confidence interval分类
数理科学引用本文复制引用
侯梦哲,秦永松,张正家..平稳自回归模型的调整经验似然置信域[J].四川大学学报（自然科学版）,2024,61(2):29-35,7.基金项目
国家自然科学基金(12061017,12161009) （12061017,12161009）
广西高校中青年教师科研基础能力提升项目(2022KY0049) （2022KY0049）
