统计研究 ›› 2011, Vol. 28 ›› Issue (4): 99-105.

• 论文 • 上一篇    下一篇

正交多项式逼近下非线性趋势序列单位根检验

刘田 谈进   

  • 出版日期:2011-04-15 发布日期:2011-04-14

Unit Root Test of Series with Nonlinear Trend Based on Orthogonal Polynomial Approximation

Liu Tian & Tan Jin   

  • Online:2011-04-15 Published:2011-04-14

摘要: 传统单位根检验方法常常假设带有线性的确定性趋势,但如果趋势是非线性的,通常将因为检验功效大幅下降而导致检验失败。本文研究用正交多项式逼近非线性趋势,然后对残差进行单位根检验的方法。研究了用正交多项式进行趋势逼近的性质,推导了这种方法进行单位根检验时统计量的极限分布,提出了正交多项式最高阶数的确定方法,仿真研究了残差相关与不相关时的检验功效。结果表明,检验方法是有效的。

关键词: 农产品, 生存分析, 贸易关系, 持续时间, 正交多项式逼近, 非线性趋势, 单位根检验, 检验功效

Abstract: Conventional unit root test usually assumes the series with linear deterministic trend. But in case of nonlinear trend, the test power becomes impotent drastically, resulting in test failure. Orthogonal polynomial approximation is proposed to estimate the nonlinear deterministic trend and then the residual is used to test the unit root. This paper researches the properties of orthogonal polynomial approximation to any linear or nonlinear trend, and then deduces the limiting distribution of unit root test statistic based on orthogonal polynomial approximation and suggests how to decide the highest order of orthogonal polynomial. The test powers are simulated with residuals correlated or non-correlated. The results show that the proposed method works well.

Key words: Agriculture Products, Survival Analysis, Trade Relationship, Duration, Orthogonal Polynomial Approximation, Nonlinear Trend, Unit Root Test, Test Power