journal6 ›› 2011, Vol. 32 ›› Issue (3): 22-25.

• 数学 • 上一篇    下一篇

基于Wild Bootstrap 非参数方法的AR模型线性检验

  

  1. (铜陵学院经济贸易系,安徽 铜陵 244061)
  • 出版日期:2011-05-25 发布日期:2012-04-08
  • 作者简介:王敬勇(1978-),男,安徽淮北人,铜陵学院经济贸易系讲师,博士,主要从事计量经济学研究.

Wild Bootstrap Nonparametric Kernel Test for Linearity in AR Regression Models

  1. (Economics and Trade Department of Tongling University,Tongling 244061,Anhui China)
  • Online:2011-05-25 Published:2012-04-08

摘要:介绍了AR模型线性的非参数核检验统计量,并对检验统计量进行了检验水平和检验势的Wild Bootstrap模拟.模拟结果显示,非参数核检验统计量的Wild Bootstrap检验稳健性和可靠性都比渐近检验高,TAR与双线性AR模型比其他3种AR的检验势高.

关键词: Wild Bootstrap, AR模型, 检验水平, 检验势, 非线性

Abstract: This article considers Wild Bootstrap nonparametric kernel test for linearity in AR regression models,and examines finite sample performance:test size and test power.Simulation results show that test reliability and robustness of Wild Bootstrap is higher than asymptotic method,and test power of TAR and bilinear models is better than the others AR models.

Key words: Wild Bootstrap, AR model, test size, test power, nonlinearity

公众号 电子书橱 超星期刊 手机浏览 在线QQ