随机排列方法检验方差变点

Variance Change Point Detection Using the Random Permutations Test

给出了随机排列的主要性质及证明。构造了随机排列检验方差变点的统计量。以GARCH(1,1)过程为例,模拟比较了随机排列方法与近似极限分布方法关于方差变点检验的临界值。应用随机排列方法检测人民币兑美元汇率的变点,并与惩罚对比函数方法作比较。模拟与实证结果均表明随机排列方法检验方差变点是灵活有效的。

In this paper,some main properties of the random permutations test are presented,and an random permutations test statistic for variance change point is constructed.Also,the performances of the random permutation test is discussed in terms of the critical values for change-point detection under GARCH(1,1) processes,by contrasting with the approximate limit distribution test.Finally,the paper apply the random permutations test and a penalized contrast methodology being a contrast side to detect Change-points in CNY data.Both simulation and empirical results show that the random permutations tests are flexible and effective tools to analyze the volatility of a data which population distribution is not known.