上海大学学报(自然科学版) ›› 2011, Vol. 17 ›› Issue (6): 740-745.

• 数理化科学 • 上一篇    下一篇

相依序列方差变点的非参数统计分析

 邹美红, 程建强, 何幼桦   

  1. (上海大学 理学院,上海 200444)
  • 收稿日期:2010-05-19 出版日期:2011-12-28 发布日期:2011-12-28
  • 通讯作者: 何幼桦(1960~),男,副教授,博士,研究方向为数理统计. Email: heyouhua@shu.edu.cn
  • 基金资助:

    国家自然科学基金资助项目(11071158);上海市重点学科建设资助项目(S30104)

Nonparametric Statistical Analysis of Structural Change Point in Volatility Models for Dependent Time Series

  1. (College of Sciences, Shanghai University, Shanghai 200444, China)
  • Received:2010-05-19 Online:2011-12-28 Published:2011-12-28

摘要: 基于严平稳β-混合过程,建立回归函数和条件方差函数形式均未知情况下的自回归异方差模型方差变点的估计方法;并给出变点检验统计量渐近正态性的证明,由此得到方差变点的检验方法;最后,通过数值模拟,展示估计方法的有效性.

关键词: &beta, 方差变点, 非参数异方差模型, -混合过程, 渐近正态性, 局部线性估计

Abstract: Based on a stationary β-mixed process, this paper proposes a method for nonparametric estimation of structural change point in volatility models with unknown regression and conditional variance functions. Asymptotic normality of test statistic is proven, and a corresponding test method presented. Effectiveness of the method is shown with simulations.

Key words: β-mixed process, asymptotic normality, change point in volatility, locally linear estimation, nonparametric heteroscedastic model

中图分类号: