关于股票数据波动性的多变点研究

View on mean multiple change point test of stockdata

目前对金融时间序列模型结构变化问题主要集中于对单变点的研究,但在许多情况下,金融数据可能出现多个变点,因此,对实际问题的研究需要检验多变点的存在,需要对变点的个数以及变点发生的时刻作出估计.本文讨论了股票数据均值的多变点检验问题.在原假设下给出统计量的极限分布及渐近临界值的解析表达式.并且在递归检验的过程中同时得到变点时刻与变点个数估计.最后用实例分析说明了方法的有效性.

Nowadays,the change of structure of financial time series model is centered on the research of single change-point.But financial data may have multiple change points,so in the research of practicalities,we need to test the exit of multiple change points and estimate the number and date of breaks.This paper studies the problem of mean multiple change-point test of stock data.The limiting distribution of the test under null hypothesis is present.In addition,we derive analytical expression for asymptotic critical value.The estimator of the date and number of breaks is instantaneously obtained via the test procedure.Real data analysis supports the validity of our test.