基于中国股票市场的长记忆模型应用研究

Application of Long Memory Model Based on Chinese Stock Market

针对金融时间序列具有长记忆性这一特征,采用广义双曲线分布族下的ARFIMA模型估计上证指数收益率的长记忆强度,并对比不同时间划分下时间序列长记忆效应的效果;实证结果显示,上证指数收益率不仅存在长记忆效应,而且时间和事件对长记忆性的效果有显著影响.

According to the characteristic of financial time series with long memory,this paper uses ARFIMA Model under generalized hyperbolic distribution cluster to estimate the long memory strength of the return rate of Shanghai Stock Index,compares the values of long memory effect of time series under different time segmentations, and empirical results show that the return rate of Shanghai Stock Index not only has long memory effect but also its time and events have significant impact on the effect of long memory.