基于Bootstrap方法的AR(p)模型方差变点的检验  

Testing for Variance Changes in AR(p) Models Based on Bootstrap Method

作　　者：金浩[1] 张思[1] 杨云锋[1] JIN Hao;ZHANG Si;YANG Yun-feng(School of Science,Xi＇an University of Science and Technology,Xi＇an,710054,China)

机构地区：[1]西安科技大学理学院

出　　处：《数学的实践与认识》2018年第16期23-30,共8页Mathematics in Practice and Theory

基　　金：国家自然科学基金(71473194);陕西省科技厅科技基金(2013KJXX-40,2017JM1042);陕西厅专项基金(2016JK1500)

摘　　要：研究了基于AR（p）模型结构变点的检验.原假设条件成立时,证明了残量平方累积和统计量的极限分布仍然是一个标准布朗桥的上确界,并利用bootstrap抽样方法以提高经验势函数值.数值模拟和实例分析都充分说明方法对相依过程结构变点检验的有效性.This paper considers the problem of testing for variance breaks in AR（p） models. It is shown that the limiting distribution of the conventional residual CUSUM of squares test （RCUSQ） statistic is the sup of a standard Brown bridge under null hypothesis. In order to improve the empirical size and power, the bootstrap method is adopted. Finally, numerical evidence and data example suggest that our asymptotic results perform well against a range of dependent processes.

关 键 词：累积和检验 BOOTSTRAP 布朗桥 结构变点 

分 类 号：O212.1[理学—概率论与数理统计]