摘要
基于误差是φ-混合序列,研究一阶适度爆炸自回归模型回归系数最小二乘(least squares,LS)估计问题。在误差满足弱条件E(u1)=0,E(|u1|4)<∞和φ(n)=O(n-8)时,获得回归系数最小二乘估计的极限分布——柯西分布。通过数据模拟得到的模拟结果与理论结果一致。作为应用,利用一阶适度爆炸模型和自回归系数区间估计,研究英伟达股票2013—2023年的股价增长过程。
The least squares(LS)estimator is studied for the first-order and mildly explosive autoregression withφ-mixing errors.Under some weak conditions of E(u1)=0,E(|u1|4)<∞andφ(n)=O(n-8),the limit distribution of standard Cauchy distribution is obtained for the LS estimator.Some simulations are given,which agree with the theory results.As an application,the first-order mildly explosive model and autoregressive coefficient interval estimation are used to study the growth process of share price for the NVIDIA corporation common stock from 2013 to 2023.
作者
段体城
许成凯
鲁紫婕
贾沛焱
谢素雅
杨文志
DUAN Ticheng;XU Chengkai;LU Zijie;JIA Peiyan;XIE Suya;YANG Wenzhi(School of Big Data and Statistics,Anhui University,Hefei 230601,Anhui,China;Wendian College,Anhui University,Hefei 230601,Anhui,China)
出处
《山东大学学报(理学版)》
北大核心
2025年第3期77-87,共11页
Journal of Shandong University(Natural Science)
基金
安徽省自然科学基金资助项目(2008085MA14)
安徽大学大学生创新创业训练项目(202310357126)。
关键词
适度爆炸自回归
最小二乘估计
柯西分布
混合序列
mildly explosive autoregression
least squares estimator
Cauchy distribution
mixing sequence
